AMD, Intel and Nvidia driver issues and last recommended version

[ryrynz]

What a mess.

https://www.windowscentral.com/how-disable-automatic-driver-updates-windows-10

[oldpainlesskodi]

So, still no fix in sight for the AMD Navi cards not sending the BT2020 flag.....

[Fandangos]

Hi everyone,

I'm the developer of the Kodi HDR Edition and I'm trying for quite sometime to fix the Ryzen bug that does not send the BT2020 flag.
I have a RX 550 here that I use for testing and with this gpu everything is correct.

The user _DMU_ has been of great help and tested several builds without any changes.
The entire metadata is passed perfectly, Red X and Y, Blue X and Y, Green X and Y, White X and Y, min Master, max Master, max FALL, max CLL, the HDR flag is triggered and it stays at RGB and not RGB BT.2020.

So I was looking at my TV EDID and it is a Sony X905E, corresponding model in the US is X900E I believe.
Which is HDMI 2.0 level B.

For everyone suffering with this problem, can you post a picture of the EDID capture of the colorimetry and if HDMI 2.0 support shows up in your EDID when capturing data with Custom Resolution Utility?

Here's the colorimetry of my TV for reference.
https://ibb.co/dGvtKk4

It shows 600mhz and HDMI 2.0 support.

I want to exclude that this is not a display problem because on the software side, there's nothing much else I can do using AMD AGS API.

[oldpainlesskodi]

Mine shows HDMI 2 support, and, same here, HDR is triggered, but the TV reports RGB Colour space, but when using the Win HDR Toggle, the set reports the colour space as BT2020.

https://ibb.co/DYxvCDB

[Fandangos]

Thanks for trying.

I've compiled the AGS Sample that will output some information gathered by the AMD driver, if you can give this exe a go:
https://mega.nz/#!LUsk0KzK!hVcfg1ki4Qu_XrLVNHPIm9EBDhmv_6zmz9orPkajYIA

Source code is here:
https://github.com/GPUOpen-LibrariesAndSDKs/AGS_SDK/tree/master/ags_sample

As I was explaining in the Kodi HDR thread, I could add an automatic way to trigger the Windows HDR switch to ON when playback starts and off when it ends but it comes with a caveat.
When we do this it will revert movie playback from 23hz to 60hz.
The only other way around this for my fork is adding a shortcut that the user would have to press prior to starting the video to toggle this switch on, this would cause correct colors as BT.2020 would be sent.

Still I would like to find the cause of this bug to avoid such solution.

[oldpainlesskodi]

Does the program create a log for me to share? or an image? if so, let me know where - happy to assist.

Screenshot of results:

https://ibb.co/3v1xR0M

[Fandangos]

Does the program create a log for me to share? or an image? if so, let me know where - happy to assist.

Screenshot of results:

https://ibb.co/3v1xR0M

Just run it, click on the left corner of the command pront, edit, select all.
Again into edit and copy, paste it here.

EDIT: Can you run it with windows HDR enabled and disabled?

[oldpainlesskodi]

Display Device: 0: AMD Radeon RX 5700 XT, \\.\DISPLAY1 (primary)
Display Device: 1: AMD Radeon RX 5700 XT, \\.\DISPLAY2 [disabled]
Display Device: 2: AMD Radeon RX 5700 XT, \\.\DISPLAY3 [disabled]
Display Device: 3: AMD Radeon RX 5700 XT, \\.\DISPLAY4 [disabled]
Display Device: 4: AMD Radeon RX 5700 XT, \\.\DISPLAY5 [disabled]
Display Device: 5: AMD Radeon RX 5700 XT, \\.\DISPLAY6 [disabled]
Display Device: 6: AMD Radeon RX 5700 XT, \\.\DISPLAY7 [disabled]
Display Device: 7: AMD Radeon RX 5700 XT, \\.\DISPLAY8 [disabled]

Driver check could not determine the driver version for 18.8.randombetadriver
Driver check shows the installed 18.8.123randomdriver driver is newer or the same as the 18.8.2 required version
Driver check shows the installed 18.9.randomdriver driver is newer or the same as the 18.8.2 required version
Driver check shows the installed 18.8.2 driver is newer or the same as the 18.8.2 required version
Driver check shows the installed 18.8.2 driver is newer or the same as the 18.8.1 required version
Driver check shows the installed 18.8.2 driver is older than the 18.8.3 required version


AGS Library initialized: v5.4.0
Is *not* WACK compliant for use in UWP apps
-----------------------------------------------------------------
Radeon Software Version: 20.1.3
Driver Version: 19.50.15.01-200117a1-351045E-RadeonSoftwareAdrenalin2020
-----------------------------------------------------------------

---------- Device 0 [primary], AMD Radeon RX 5700 XT
Vendor id: 0x1002 (AMD)
Device id: 0x731F
Revision id: 0x00C1

Architecture: RDNA, 40 CUs, 20 WGPs, 64 ROPs
core clock 1755 MHz, memory clock 1750 MHz
9.0 Tflops
local memory: 7805 MBs (437.5 GB/s), shared memory: 8150 MBs


Eyefinity not enabled on this device

---------- Display 0 [primary]----------------------------------------
device name: \\.\DISPLAY1
monitor name: AV Receiver

Max resolution: 3840 x 2160, 60.0 Hz
Current resolution: 3840 x 2160, Offset (0, 0), 60.0 Hz
Visible resolution: 3840 x 2160, Offset (0, 0)

chromaticity red: 0.065600, 0.034000
chromaticity green: 0.032000, 0.062400
chromaticity blue: 0.015600, 0.005600
chromaticity white point: 0.000000, 0.000000

luminance: [min, max, avg] 0.000000, 0.000000, 0.000000
screen reflectance diffuse 0.000000
screen reflectance specular 0.000000

HDR10 supported
Dolby Vision supported

logical display index: 12
ADL adapter index: 0


-----------------------------------------------------------------
Primary Display detected
Display name: x
Device name:
DisplayFlags: 0

done

[oldpainlesskodi]

Win HDR toggle enabled:

Display Device: 0: AMD Radeon RX 5700 XT, \\.\DISPLAY10 (primary)
Display Device: 1: AMD Radeon RX 5700 XT, \\.\DISPLAY11 [disabled]
Display Device: 2: AMD Radeon RX 5700 XT, \\.\DISPLAY12 [disabled]
Display Device: 3: AMD Radeon RX 5700 XT, \\.\DISPLAY13 [disabled]
Display Device: 4: AMD Radeon RX 5700 XT, \\.\DISPLAY14 [disabled]
Display Device: 5: AMD Radeon RX 5700 XT, \\.\DISPLAY15 [disabled]
Display Device: 6: AMD Radeon RX 5700 XT, \\.\DISPLAY16 [disabled]
Display Device: 7: AMD Radeon RX 5700 XT, \\.\DISPLAY17 [disabled]

Driver check could not determine the driver version for 18.8.randombetadriver
Driver check shows the installed 18.8.123randomdriver driver is newer or the same as the 18.8.2 required version
Driver check shows the installed 18.9.randomdriver driver is newer or the same as the 18.8.2 required version
Driver check shows the installed 18.8.2 driver is newer or the same as the 18.8.2 required version
Driver check shows the installed 18.8.2 driver is newer or the same as the 18.8.1 required version
Driver check shows the installed 18.8.2 driver is older than the 18.8.3 required version


AGS Library initialized: v5.4.0
Is *not* WACK compliant for use in UWP apps
-----------------------------------------------------------------
Radeon Software Version: 20.1.4
Driver Version: 19.50.17.01-200124a-351282E-RadeonSoftwareAdrenalin2020
-----------------------------------------------------------------

---------- Device 0 [primary], AMD Radeon RX 5700 XT
Vendor id: 0x1002 (AMD)
Device id: 0x731F
Revision id: 0x00C1

Architecture: RDNA, 40 CUs, 20 WGPs, 64 ROPs
core clock 1755 MHz, memory clock 1750 MHz
9.0 Tflops
local memory: 7805 MBs (437.5 GB/s), shared memory: 8150 MBs


Eyefinity not enabled on this device

---------- Display 0 [primary]----------------------------------------
device name: \\.\DISPLAY10
monitor name: AV Receiver

Max resolution: 3840 x 2160, 60.0 Hz
Current resolution: 3840 x 2160, Offset (0, 0), 60.0 Hz
Visible resolution: 3840 x 2160, Offset (0, 0)

chromaticity red: 0.065600, 0.034000
chromaticity green: 0.032000, 0.062400
chromaticity blue: 0.015600, 0.005600
chromaticity white point: 0.000000, 0.000000

luminance: [min, max, avg] 0.000000, 0.000000, 0.000000
screen reflectance diffuse 0.000000
screen reflectance specular 0.000000

HDR10 supported
Dolby Vision supported

logical display index: 12
ADL adapter index: 0


-----------------------------------------------------------------
Primary Display detected
Display name: x
Device name:
DisplayFlags: 0

done

[Fandangos]

Yeah.........

It's not the first time I try to fix this and once again this is a dead end.
I can't find anything that might result in bt.2020 not being sent to the display.

[oldpainlesskodi]

Well, at least thanks for looking.

[DMU]

Your post (https://forum.kodi.tv/showthread.php?tid=345566&pid=2919981#pid2919981) on the Kodi forum:

the workaround is turning windows hdr toggle to on.
You can do this before opening kodi, just going into windows display settings, turning it on and colors will be correct.
This is not a good solution for the end user so i can make it automatic but.... And a huge but..
The code we have to turn hdr on will default the refresh rate back to 60hz.
So this will kill 23hz playback.

Solve this problem so that the frequency does not switch. Do not need anything else.
I don’t understand why the frequency is switched? This code simply activates the Windows HDR mode, the frequency does not change in the desktop!

[Fandangos]

Your post (https://forum.kodi.tv/showthread.php?tid=345566&pid=2919981#pid2919981) on the Kodi forum:


Solve this problem so that the frequency does not switch. Do not need anything else.
I don’t understand why the frequency is switched? This code simply activates the Windows HDR mode, the frequency does not change in the desktop!

Me neither. I have no idea why it does, but it does.

I tried with both the code you provided and the refactored code made by Jogal both resulted in 60hz being activated again.

Tried with the code inside Kodi, tried with external .exe.

Tried switching at the start of video playback, tried switching a few seconds after playback starts.
Every scenario results in 23hz reverting to 60hz.

This is true for both AMD and Nvidia.

If I switch the refresh using Windows display settings and run the code it will mantain 23hz.
I even tried triggering kodi's updateresolution call that would set 23hz again, nothing.

And I can't activate it automatically prior to playback start because we have no means to know if the file played has HDR metadata or it's primaries are bt.2020 before playback starts.

So the only solution I see to not break refresh rate switching is to turn it ON before the playback starts manually.

Maybe some other dev could point a solution, I honestly have no idea how to overcome this limitation.

EDIT:
Opened another issue at the AMD AGS:
https://github.com/GPUOpen-LibrariesAndSDKs/AGS_SDK/issues/35

[DMU]

Tried switching at the start of video playback, tried switching a few seconds after playback starts.
Every scenario results in 23hz reverting to 60hz.
Is this the reason why Jogal does not want to automatically activate the HDR mode?


Opened another issue at the AMD AGS:
https://github.com/GPUOpen-LibrariesAndSDKs/AGS_SDK/issues/35
Have you read my (https://github.com/GPUOpen-LibrariesAndSDKs/AGS_SDK/issues/33)?

[Fandangos]

Is this the reason why Jogal does not want to automatically activate the HDR mode?


Have you read my (https://github.com/GPUOpen-LibrariesAndSDKs/AGS_SDK/issues/33)?

I don't know.
He is a far better dev than I am so I don't know if he faced another issue besides that one.

And yes I've seen yours, it's very informative, sadly I don't think anyone looks at the ags github anymore.

[huhn]

i just say it again they have more important things to do for now.

the 5600 is a very important product this one just has to work.

[oldpainlesskodi]

Huhn - Not sure where you are coming from/what you mean.

[Asmodian]

The drivers have issues that are a lot more serious than this bug with the HDR API.

[oldpainlesskodi]

Oh well. I don't game, so, here's to the next driver........

[mclingo]

oddly, gaming on my 5700 is flawless :)

[huhn]

if you think so...

better don't read the driver change log you may knock yourself out from hitting the table.

you just have to disable a couple of things to avoid the bigger issues.
freesync
hardware decoding

just make sure you don't have more then 2 screen active at a time.
reduce your ram clock.
don't use PCIe 4.0
down clock/volt your GPU.
don't use more then 60 hz.

and you are most likely fine.
except for the games/software that generally doesn't work but "details".

look who is back:
Some Radeon RX 5700 series graphics users may intermittently experience a black screen while gaming or on desktop. A potential temporary workaround is disabling hardware acceleration in applications running in the background such as web browsers or Discord.

ohh wait is was never gone...

so let me guess fixing a meta data issue or driver crashes from watching youtube videos what could be more important right now?

releasing a new buggy GUI didn't help here at all. the user base is "in a foul mood" now.

[chros]

Oh well. I don't game, so, here's to the next driver........
Then why don't you buy a cheap nvidia Pascal card (1070, 1080), use it with the good old driver and get rid of this crap? :)

[oldpainlesskodi]

Ha...good one lol

[Fandangos]

Since this works with older cards without any problem.

Have you guys tried the earliest drivers for your gpu (ryzen or 5700) to see if this fixes the problem?

[huhn]

i had the card when the first driver was released so no thank you. with disabled hardware acceleration i'm pretty much "fine" now. so the driver is not constantly crashing for me.

btw. ryzen has nothing to do with this AMD CPU are doing just great they are pretty much a general recommendation.

[mclingo]

i've tried all the old drivers myself too, the only reason i'm still with red is 3D and that I do a little gaming now i've upgraded my PC to 5700 / Ryzen 3600.

To be honest I dont mind the HDR toggle now, i have it mapped to a key on my remote so if I forget all I need to do is stop the move, toggle and restart it, I dont even have to close KODI DS.

What would be nice however... Fandangos... :):):):):) - if you could make a HDR toggle version of KODI DS :):):):):):):):) - just based on the current version - that would make a lot of AMD NAVI users very happy indeed - :):):):):):):):

but that might be difficult as there would have to be some sort of delay before the movie started in MADVR so this toggled before MADVR started...

[Fandangos]

Just a remote function inside kodi dsplayer to toggle windows hdr on and off?

I never compiled dsplayer but it seems identical.
If that is the case totally possible.

What the end user would have to do is press it before the movie starts and again when the movie ends.

Would that be of any help?

[oldpainlesskodi]

Short answer, yes.

Long answer, I have managed a dirty hack by using advanced launcher tied to HDR toggle.exe and then added it my my favorites. Dsplayer normal, just press play, HDR movie = go into favs, toggle Win HDR Toggle, plays HDR movies, back to favs, toggle off.

Like I said, it's a hack, but beggars cant be choosers.

[mclingo]

Fandangos - if you cant get it to turn on and off automatically when a HDR movies starts I honestly dont think its worth your time mate, it would be nice to have it built in sure but its not really going to any different to our current method really.

but thanks for thinking of us AMD users :)

[Fandangos]

I can make it automatic the problem is that when it is triggered it reverts to 60hz.

So this kills 23hz refresh rate change.

[oldpainlesskodi]

Fandangos - if you cant get it to turn on and off automatically when a HDR movies starts I honestly dont think its worth your time mate, it would be nice to have it built in sure but its not really going to any different to our current method really.

but thanks for thinking of us AMD users :)

I agree - god bless.

[mclingo]

[QUOTE=Fandangos;[/QUOTE]

No worries I wouldnt bother then, even if you could get it to change before the movie starts I dont think it would be stable, we're just going to have to hope AMD fix this soon.

thanks.

[Fandangos]

Could someone with the Ryzen or the RX 5700 bt.2020 problem test my latest release?

You can grab it here:
https://github.com/fandangos/Kodi-HDR-Edition/releases

Install, open it up, press CTRL+H while still in the Kodi interface, play a HDR movie and tell me if the colors are ok.

This can be made automatic for HDR content by adding autohdr true to advancedsettings.xml but this will disable refresh rate change so the movie will be at 60hz all the time.

[mclingo]

hi, i've used your previous versions on occasion just to test and provide you feedback but I need MADVR for tonemapping lovelyness. Your version has always worked fine for me, colours are fine as it uses WINDOWS HDR API not AMD private does it not, or did you switch back?

[Fandangos]

hi, i've used your previous versions on occasion just to test and provide you feedback but I need MADVR for tonemapping lovelyness. Your version has always worked fine for me, colours are fine as it uses WINDOWS HDR API not AMD private does it not, or did you switch back?

Deleted the last build there is something wrong with it.

It uses the AMD private api but uses Windows to force the bt.2020 flag.

Maybe you are confusing my build with somebody else's.

EDIT:

Now everything is fixed:
https://github.com/fandangos/Kodi-HDR-Edition/releases/tag/jan-2020

[mclingo]

I just got confused, there was a lot of talk of using WINDOWS API on your thread.

The build i currently have installed is Kodi-HDR-Edition-4.0-rc3-Python2.exe, if this is also using AMD private API then this works fine for me, no saturation issues. I will however install your new build over the top and try it again when I get home tonight.

[Fandangos]

Yes that's mine but I don't use windows api.
Jogal the one that works on the other fork does use windows api.

So you mind sharing what gpu you have and driver version?

[mclingo]

Hi, AMD RX 5700 20.1.3 - win10 1809 - I didnt try your latest as the other one works and is python2

[mclingo]

ok this is odd, the version I had installed was working fine when I first installed it, now the colours look horrific, massively saturated and contrasty and blown out and HDR 200%, if that makes senese. So I tried your new version, now that is back to the usualy problem with the API, unsaturated colours.

So it triggers the AMD HDR API ok but colours but it doesnt force BT2020, at least two of your versions I tried worked fine initially.

[Asmodian]

No driver updates between then and now?

[mclingo]

totally lost track of my drivers installs, done a lot of jumping around over the last few weeks, couldnt tell you which version I was on when it was working i'm afraid.

[mclingo]

Hi, I can actually get your version to work but only under certain circumstances. Windowed mode has to be applied and I also have to run my desktop in 4k, otherwise it plays the movie massively zoomed. I'm salso getting black screens all ovcer the place when switching two and from 4k to 1080p so i'd have to run my desktop in 4k all the time.

The reason I was running my desktop in 1080p was mainly that 3D movies worked better when KODI is running in 1080p and 4k movies still played fine. However I havent tried running my desktop in 4k for ages, since I was using FSE so i'll give it another go as this new version seems to be quite stable as long as everything is 4k.


edit: tried 4k desktop with KODI DS and 3D, even worse if anything, black screens on stop, and only playing top left hand side of screen, with FSE or windowed mode, its been like this since early versions of windows 10, if I run my desktop in 1080p its all fine, i logged this as a bug with MADVR but madshi insisted it was a driver issue but I dont think it is, MADVR stumbles when changing to 1080p from 4k and back again for 3D, always has for me anyway with about 5 different RX cards and multiple new win10 installs over the years.

so those running everything in 4k will probably beneift from this version but those who want tonemapping will still have to use the manual HDR win toggle for now.

[Fandangos]

ok this is odd, the version I had installed was working fine when I first installed it, now the colours look horrific, massively saturated and contrasty and blown out and HDR 200%, if that makes senese. So I tried your new version, now that is back to the usualy problem with the API, unsaturated colours.

So it triggers the AMD HDR API ok but colours but it doesnt force BT2020, at least two of your versions I tried worked fine initially.

Can you press CTRL+H before opening the movie and let me know if the colors are fixed?

Or just enable autohdr on advancedsettings.xml, will do almost the same if you don't care about 23hz playback.

[mclingo]

23hz is essential to me i.m afriad but colours are fixed.

[Fandangos]

23hz will only be broken with autohdr turned on.

If you press ctrl+h before opening the video file it will fix the colors and 23hz will work.

[mclingo]

great work my friend, just wish i had a use for it, madvr playback is so much better than stock KODI i couldnt even consider using their player unless KODI DS gets broken.

just wish MADSHI would get onto this bug, all he needs to do is add a switch in MADVR to force BT2020, then we could write that switch into a HDR profile.

[oldpainlesskodi]

@Fandangos - One of the biggest downsides to stock Kodi is its scalers and lack of picture tweaking options.

However, you have done some great work, which I am sure the Kodi community greatly appreciate.

[DMU]

Has anyone installed the latest AMD 20.2.1 drivers? Checked HDR?

[oldpainlesskodi]

No, its now even more broken - any copyback over blows the colours & contrast and darkens the image to the point you cant really see whats going on, and still outputs RGB 709.....oh well, back to the previous less broken driver.

[huhn]

VP9 hardware decoding was to much to ask for yes so gone?
p010 d3d11 processing is a new low for them... i have to test copyback now that sounds amazing.

[oldpainlesskodi]

Yep, wonder if we should offer to code for them...

[chros]

If you are bored (~30 mins): How AMD Radeon Sabotages Itself & Its Partners: Development Timelines & Failures (https://www.youtube.com/watch?v=sDyP9nDw_q0)

[oldpainlesskodi]

And I though Nvidia drivers were poorly managed (change control, fix 1 thing and break 5), guess I'll just wait until they release one with the features I need, then lock the driver down.

[chros]

And I though Nvidia drivers were poorly managed (change control, fix 1 thing and break 5), guess I'll just wait until they release one with the features I need, then lock the driver down.
:) Exactly. Unfortunately, it applies to everything "today", not just drivers, and not just GPUs: mobiles, TVs, cars, etc ... you name it :)

[oldpainlesskodi]

Agreed.

[mclingo]

have we all tried the driver and are we all having the same issues, i'll try it when I get home just in case, as it has two issues, HDR & raised blacks with EVR renderer, perhaps it fixes the EVR issue?

[huhn]

nope DXVA2 handling is still broken they just broke d3d11 processing too just for fun i guess.

[oldpainlesskodi]

Heaven knows who they outsourced the dev work to. Am sure they will get there.....

[mclingo]

feck, I wont bother then thanks Huhn :)

[chros]

nope DXVA2 handling is still broken they just broke d3d11 processing too just for fun i guess.
:D It's a good first step, now they can calmly start from scratch :D

[SamuriHL]

You AMD guys are making me SO glad I opted for an RTX 2080 in my new laptop. LMAO Can't wait to play with that toy next week. But uhhhh wow I thought nVidia had some issues with drivers. The last few pages of this thread, though, really have not convinced me to go back to AMD any time soon.

[mclingo]

to be honest AMD are usually pretty good, i've used them for years and only had minor issues, this is the first time they've really broken HTPC stuff for a long time, at least they've kept 3D - for now.

[SamuriHL]

I've used amd in the past so I know how good they can be. Nvidia is no saint when it comes to features and drivers, however, lately it seems amd is making some pretty big missteps. Do these latest cards have hardware encoding and decoding support for h265? Driver wise they'll hopefully eventually get the issues worked out but I'm curious if they've come up to parity on features yet. Yes 3d is nice that they kept it but isn't enough to sell me on one.

Sent from my SM-G975U using Tapatalk

[NoTechi]

I recently switched from a 1080 to a 5700. Main reason was the 3D support without having to keep an outdated driver.

For my use case I have no disadvantage at all, as I use a JVC7900 where madvr HDR-SDR conversion is superior to output in HDR.

When I look back at the difficulties I had to get my setup to run stable with nvidia for my use case, I feel like I'm in heaven now and don't regret the change at all.

NoTechi

[WuNgUn]

I've used amd in the past so I know how good they can be. Nvidia is no saint when it comes to features and drivers, however, lately it seems amd is making some pretty big missteps. Do these latest cards have hardware encoding and decoding support for h265? Driver wise they'll hopefully eventually get the issues worked out but I'm curious if they've come up to parity on features yet. Yes 3d is nice that they kept it but isn't enough to sell me on one.

Sent from my SM-G975U using Tapatalk

I know my 960 has hardware h265 decoding...the reason I bought it so long ago :)

[huhn]

even the polaris card had 10 bit HEVC support.
it no secret that amd was ahead with the media capabilities.
nvidia didn't even support multichannel PCM over HDMI for a long long time and audio bit streaming wasn't possible too the first nvidia card to support dolby and dts was the 460 where this was a standard feature on AMD but that's been a lot of years and amd is not even competing anymore...

the 960 had VP9 support and guess what is still not working 100 % with amd up to this day VP9 they totally lost there edge and are now behind far behind.

the hardware encoder is a bad joke too compared to nvidia.

most of this is totally irrelevant for me but that doesn't change the fact i mean i couldn't care less about the hardware encoder to me even nvidia encoder is not great so it doesn't matter that the AMD one is awful i wouldn't use them anyway.

[SamuriHL]

h.265 encoding is very important to the things I do on an almost daily basis. The nVidia cards do a good enough job in that category to satisfy my requirement. (It's basically for archiving whole seasons of TV shows recorded off my TiVo and I appreciate speed over insane quality for that given the source material). I've no issue with AMD at all. Like I said, I used to use them almost exclusively for many years. This driver situation with both nVidia and AMD, however, pretty much sucks. Right now nVidia is doing a fairly decent job not breaking stuff completely. Speaking of nVidia....did anyone else notice that they seemed to drop the standard driver vs the dch driver in this last update? At least, I didn't see it listed anywhere. I just thought it strange that's all.

[huhn]

if you care about speed why would you use h265 instead of h264?

[SamuriHL]

Because for tv compression h265 is considerably smaller.

[huhn]

it's only significantly better in bit rate starving and in term of speed/quality not at all for CPU and even for hardware it's questionable. up to this day x265 isn't generally better then x264.

nvenc is not really able to beat h264 faster preset in term of quality and i wouldn't be shocked if nvenc h264 is still better then nvenc h265.

so if you use a "high" bit rate you may just not doing yourself a favour by using h265 even with hardware.

testing of this is usually a waste of time by filtering with different denoising filter for example but for pascal i have no doubt at all that nvenc h265 doesn't stand a chance against nvenc h264.

[WuNgUn]

it's only significantly better in bit rate starving and in term of speed/quality not at all for CPU and even for hardware it's questionable. up to this day x265 isn't generally better then x264.



nvenc is not really able to beat h264 faster preset in term of quality and i wouldn't be shocked if nvenc h264 is still better then nvenc h265.



so if you use a "high" bit rate you may just not doing yourself a favour by using h265 even with hardware.



testing of this is usually a waste of time by filtering with different denoising filter for example but for pascal i have no doubt at all that nvenc h265 doesn't stand a chance against nvenc h264.I'm more about image quality than arguing the merits of calculating encoding processing power. In my opinion, 265 is much better. The only stuff I see that's 264 is crap that's meant to be mass streamed illegally or downloaded illegally. Lol

[SamuriHL]

I've done a LOT of my TiVo stuff in h264 in the past so I'm well aware of what it can do. However, for the content I'm dealing with, I can say assuredly at the LOW bitrates I'm using (~3kbps) h265 is far higher quality and about half the space of h264. And while there is certainly a speed difference, for the size files I'm talking about the difference between h264 and h265 using nvenc is fairly minimal. A 45 minute show (with commercials cut) @1080 takes roughly 4 to 6 minutes on my RTX 2070 to encode. We'll find out how the new laptop will do with its RTX 2080 next week for comparison but I expect it to be almost identical. Those are speeds I can most definitely live with for what I'm trying to accomplish. Now if I were trying to compress lossless UHD streams and maintain quality, that'd be an entirely different story, of course. (Note, my use of the term lossless means unmessed with off a UHD disc, not that UHD is inherently lossless).

[mclingo]

Hi, just posting here too for max coverage, a bit of welcomed KODI DS dev, . I see you're already there OPK :)

https://forum.kodi.tv/showthread.php?tid=351534

[oldpainlesskodi]

Yep....

[foxyshadis]

the LOW bitrates I'm using (~3kbps)

Holy cow, that's "Shrek on a floppy" level. I wonder how VVC does with that kind of stress test, since we already know AV1 is just barely watchable.

(I know, you means 3mbps, but testing the limits is fun!)

[SamuriHL]

Roflmao wow that was a fantastic typo on my part lol! Yes if course I meant 3mbps but that was hilarious! [emoji23]

Sent from my SM-G975U using Tapatalk

[WuNgUn]

Just installed the latest nVidia driver...
Now in madVR, I can't play ANY NGU chroma upscaling without a ton of dropped frames...
May NOT be driver that did this, but reverting back to my last driver to be sure...

[WuNgUn]

....or, maybe because I installed the latest MadVR test build...
Ugh

[SamuriHL]

No I think it's this driver as that's the only thing that changed for my setup and I'm seeing a similar problem. I'll be downgrading later to test it.

Sent from my SM-G975U using Tapatalk

[WuNgUn]

No I think it's this driver as that's the only thing that changed for my setup and I'm seeing a similar problem. I'll be downgrading later to test it.

Sent from my SM-G975U using TapatalkI went back to the driver I was using...and went back to stable MadVR...
Still having drop frame issues now...
Even 1080p to 4k upscaling gives me issues if I use anything other than Jinc.
If I try any chroma upsampling in NGU, it's drop frame city.

Wondering if the test MadVR settings (even though I can't access them now) were retained when I over wrote the files and installed the stable version.

[ryrynz]

Now in madVR, I can't play ANY NGU chroma upscaling without a ton of dropped frames...

Check Nvidia power profile.

[VBB]

No difference for me with the latest driver. As ryrynz said, always check that power is set to "adaptive" instead of the default "optimal" after install.

[SamuriHL]

Not sure what's up with my system. I did a clean downgrade install to the previous driver and things are back to normal for me. I'll re-upgrade again tomorrow and see if it reoccurs.

[huhn]

d3d11?
i see these even with an HD 4000.

edit like this: https://abload.de/img/curroptiony4ksh.png

[DMU]

Do all Navi owners have such problems with HDR on the latest driver version as mine?
Video. (https://drive.google.com/open?id=18kLHSAYc15AgDwQK3NOrdAHkXQp6ZXbI)
Picture. (https://drive.google.com/open?id=1E_uOjnrY02XJa9dL61zLLalse9jTi_cY)

[oldpainlesskodi]

I do, as the AMD API is not sending the BT2020 flag, which is why I have to use the Win 10 HDR Toggle, which forces the BT2020 flag.

[DMU]

I meant without the Win 10 HDR Toggle (AMD HDR mode). Previously, the image was a bit unsaturated, now super-saturated.

[oldpainlesskodi]

Ah right, you mean driver 20.2.1 - yes, I had to roll back because of that...I did post here about it.

[huhn]

so do i get insane or what are there more and more post missing i answer questions of?

[Grimsdyke]

So, is it currently not (really) recommended to switch to an AMD setup ??

[nevcairiel]

The only reason to even consider that in my book is when you are absolutely set on wanting 3D, can't live with SBS/OU modes to get that, and can't use an older NVIDIA driver either.

[mclingo]

the other advantage is native 23,976 support but most people now have had a lot of success setting customs res in MADVR I guess, its nice to have though.

But yeah I guess nobody in their right mind would go AMD right now if they are not using 3D+gaming as we still have no BT2020 HDR flag fix and we've still got a problem with colour space in EVR renderer too, all have workaronds however so i'm still getting plenty of use out of my RX5700.

[aufkrawall]

There is still no VP9 support via general D3D11VA/DXVA2 API with Navi cards. I reported this issue to them, nothing happened.
It works with Nvidia since GTX 960 of January 2015, just go team green...
(or use Linux with Navi and live with the restrictions vs. Windows)

[huhn]

nev pretty much found the issue. is just special...

[nevcairiel]

Yeah AMD violates the DXVA API. Only a little bit, but enough for software that actually checks (like LAV) to not accept it. I'm still pondering if I should try a workaround or just let them suffer.

[aufkrawall]

I suppose it doesn't work with plain ffplay either, since it doesn't work with mpv. So it seems there was absolutely no testing with upstream ffmpeg by AMD. Just another indicator in what a lousy state their Windows driver team must be.

[mclingo]

AMD users - have any of you tried signing up to the vanguard beta program, I thought I might give it a try in the hope I could open up a channel for feedback on the HDR issues as they should also be affecting gaming, however I made it quite clear I was mainly a HTPC user but suggested that maybe some feedback from this angle might be useful to them. I'm not expecting an acceptance but I thought i'd give it a go anyway as I'm getting pretty cheesed off with AMD right now.

[huhn]

this looks like fun.

just look how serious amd is taking this: primary GPU series no option for NAVI.
you can be sure that you are not going to be ignored.

[foxyshadis]

I hope it works, mclingo; for years Intel has been the only one that actually tries to minimize video playback regressions, and even their record is quite spotty depending on who got the ticket. (Which is still something; before then, no one bothered testing video at all beyond "it plays, ship it".)

[xabregas]

So what is the recommended driver for nvidia anyways?

[Klaus1189]

It depends on your model and what you need, for example do you need 3D support? RTX or GTX model?

[xabregas]

GTX 1660 SUPER, no 3d support but i have 4k HDR screen even knowing i dont play a lot of HDR videos. But usually just watching bluray 1080p videos upscaled to 4k and im very confused about this studio vs gaming driver. What is the best for madvr? studio or gaming? I know my gpu is very new and only the last 2 drivers work.

[Klaus1189]

Use the latest GRD or SD, it shouldn't matter, as long as the version number is the same. GRD is updated more often, SD rarely.

[SamuriHL]

The latest nVidia drivers, knock on wood, have been pretty stable lately when it comes to video playback. No major issues with them on my RTX cards thankfully!

[Klaus1189]

I signed up at vanguard beta program, but didn't received an email :(

[mclingo]

i didnt either mate, guess thats a no then :)

[mclingo]

i think they've changed something in 20.2.2 AMD, colours look right but image doesnt have the right brightness, can somsone else install to see if you are seeing what i'm seeing.

[mclingo]

this might look fixed for people with low nit displays or projectors but the picture is very dim compared to Windows HDR turn on, this suggests its with still not outputting Bt2020 or its somehow sending the wrong meta data or something, this is really odd.

[SamuriHL]

Maybe someone with an hd fury can check what's being output for you.

[mclingo]

huhn is running a navi card but not sure what display, might be hard to see what i'm seeing without a bright display, ideally someone with a NAVI card and an OLED.

[oldpainlesskodi]

i think they've changed something in 20.2.2 AMD, colours look right but image doesnt have the right brightness, can somsone else install to see if you are seeing what i'm seeing.

It's busted, so don't bother - switch to Win HDR, press play...all fine...fast forward/skip, and it seems to flip to limited range (at least on my rig).


Back to the last lest busted version.......

Do they hire monkeys?

EDIT - and no, the AMD API still does not switch to BT2020, outputs 709....but they also manged to break the Win HDR toggle when you fast forward/skip, as explained above.

[Klaus1189]

the AMD API still does not switch to BT2020, outputs 709

How can I check that?

[mclingo]

I think you may have some issues with your rig mate, i'm seeing no issues at all apart from the image being a little bit dimmer than windows HDR. Toggle is rock solid here, tested KODI DS and MPC-HC, skips also fine in both.

if you have FSE turn off your movie might just be popping in and out of HDR mode if you move your mouse, put FSE on and try again, i see what you are seeing in MPC-HC with FSE turned off when i move my mouse, this is normal.

[mclingo]

lets see what DMU and Huhn report back, sure this is BT2020 now, just too dim.

[Klaus1189]

How can I check that?

I ask because the colors do look right to me, but that doesn't mean there could be still something wrong. I also think it is a bit dim too, but usable nevertheless. Fast forward does not produce any problem here.
Did you wipe the old driver? I used the AMD Cleanup Utility, linked in the first post.

But it looks like things are getting better now ;-)

[oldpainlesskodi]

Ok, I'll try again to rule out FSE. But my display was still reporting 709 with the AMD toggle, and 2020 with the Win HDR toggle, but i'll double check and report back.

[oldpainlesskodi]

Yep, rolling back to 20.1.4 - have double checked the variables I can think of, so, at the moment, my original post on this driver stands as is on my set up.....if it's working for you guys, would be good to understand the difference.

[mclingo]

you should not roll back just jet, you need to find out whats going on. its not really clear whats happening for you, is this all happening in KODI DS?, is windowed mode on or off in KODI, how are you determining what output is being sent to your TV?

[oldpainlesskodi]

1 - Kodi yes.
2 - Windowed in Kodi - no, always keep fullscreen, but tried FSE and windowed in Madvr.
3 - TV reports 709 in the latest AMD driver with HDR playback, and using the WinHDR toggle, like I said, it reports bt2020, and looks fine until in Kodi you press next chapter, and then it lose's all of the dark/lower detail, like its gone to limited range. No such issue with this on 20.1.4 - I have triple checked.

Like you said, lets see what others find.

[mclingo]

hi, your TV actually reports what colour space its receiving? and if you turn on windows HDR it reports bt2020 and with it off bt709?

[oldpainlesskodi]

Yep......

[mclingo]

there is an ongoing conversation about this issue here, this is how I found out there was a new driver, if you're sure its still not outputting BT2020 then I need to report this, I could do with final confirmation from someone else though as colours look right to me but I have no way of knowing what i'm sending to my TV.

https://github.com/GPUOpen-LibrariesAndSDKs/AGS_SDK/issues/33

[DMU]

lets see what DMU and Huhn report back, sure this is BT2020 now, just too dim.

I don’t understand what they did, but the colors began to look right! TV reports that the color space is BT.709, and HDFury says there is no flag 2020.

[huhn]

nothing good to say here i never tested HDR i don't have a HDR screen and i'm not going to add a HDR screen to the PC the card is in.

they "fixed" d3d11 scaling a bit but ruined d3d11 processing even more in general chroma is still NN.
so they touched it but made it worse...

[mclingo]

DMU, whats you guess as to why the colours now look right, this makes no sense if no BT2020 flag is being sent as you have now confirmed to AMD.

[DMU]

I don’t know why, instead of fixing errors, they complicate things even more. Now users will think that the AGS HDR (AMD Private API) is working correctly.

[mclingo]

I feels like the fudged it hoping that nobody will notice, or they just dont know what they are doing at all, either way this doesnt bode well for future HTPC use, just dont understand how they can get something so fundamental to this card wrong.

However MADSHI did suggest he might be able to put windows HDR toggle back into MADVR, when I asked he said "maybe", a private vs windows HDR switch in MADVR, wouldnt be hard to do I dont think.

[oldpainlesskodi]

From what it looks like to me, the latest driver over compensates by overblowing colours and contrast in HDR mode via the AMD API to try and compensate for the fact its not sending the BT2020 flag and it's still in BT 709 mode - total screw up in my book.

One day......

[clsid]

This Reddit topic (https://old.reddit.com/r/Amd/comments/favtfv/radeon_software_adrenalin_2020_edition_2022/) is a good place for bug reports. There is AMD support staff active there.

[mclingo]

thanks i've added a post to that, would be helpful all of us with AMD cards can contribute to this.

[DMU]

@mclingo
By the way, could you check the new driver with Polaris? It is believed that there will be oversaturated colors.

[mclingo]

hi, sorry i sold my RX580, just running a 5700 now.

[oldpainlesskodi]

Can you share link to your post in the reddit forum?

[DMU]

One more thing, with the new AMD driver the "Windows HDR switch" bug fix no longer works. He switches to the HDR mode, but does not set metadata; standard Windows metadata is used. So it makes no sense to ask madshi to implement the Windows HDR switch in madVR. The only way for AMD owners is to use WinAPI, as in Kodi (https://forum.doom9.org/showthread.php?t=177375).
Madshi planned to implement HDR support for Intel. This will be a solution for AMD owners as well.

[oldpainlesskodi]

That would explain what I saw when I mentioned it looked like the driver had flipped to limited, as it lost shadow and highlight detail.

So, sticking with 20.1.4 for now, as at least win hdr toggle seems to be working.

[mclingo]

i've been tinkering with this for the last couple of hours, i've been swapping between driver 20.1.3 and 20.2.2 and comparing LG TV tone mapping to MADVR tone mapping with windows toggle on and off. I'm going back to what i first said, everything is right but its just too dim, I'm even sure meta data is being pass as when I tone map using MADVR i can see that highlight recovery is definitely working.

The dimmed HDR output is also affecting my standard LG tone mapping, normally if I send meta data direct to my TV the faces in harry potter the goblet of fire about 2:09 are usually almost completely clipped, there is virtually no detail at all in the bright highlights of the ghost faces, with MADVR highlight recovery all the detail comes back, this is why I use it.

Whats different with 20.2.2 is that my standard LG tone mapped image now sits somewhere in between the two, there is a lot less clipping, i've no idea why this happening, this is odd.

i'm staying on 20.2.2 for now as i'm not seeing any other issues with it to be honest, i canty explain oldpainless's issue at all though, that only happens to me when i move my mouse when playing a HDR movie in MPC-HC and FSE is off.

[oldpainlesskodi]

On your system does the Win HDR toggle put madvr in windowed fullscreen mode even though FSE is enabled?

[mclingo]

Hi, whatever settings you or I are using it must be going into some sort of full screen mode as HDR wont trigger otherwise.

All I can tell you is how i've got mine setup. In MADVR windowed mode and FSE are unticked. In KODI DS i'm using Full SCREEN and "use windowed mode" is SYSTEM is turned off.

Does that help?

[oldpainlesskodi]

Sorry if this is a daft question, but how can windowed mode AND FSE be both unticked in MadVR?

[Klaus1189]

Sorry for interrupting, what issue is fixed now, or only half? Like HDR looks right, but not outputted as BT.2020 becuause philips TV reports video signal as BT.709
How are Hardware decoders now? ...


I want to collect infos, at least AMD is doing anything right now, so I hope I could point them to the right direction anyhow.

[oldpainlesskodi]

It's still BT 709....nothing has changed.

[mclingo]

something has changed, colours are now correct, picture is just dimmer compared to windows HDR which will be something to do with it still sending BT709.

[mclingo]

Sorry if this is a daft question, but how can windowed mode AND FSE be both unticked in MadVR?


Hi, i meant window overlay

[el Filou]

How can colours be 'correct' if the card is still flagging the output as BT.709? That would mean the driver is downconverting to BT.709 behind madVR's back, which is bad as you lose wide gamut colours.

[mclingo]

i think they've fudged it somehow yeah, the colours look right, the only issue is the picture looks dimmer, hard to explain, its quite subtle, so much so joe public wont even notice, they'll just think its fixed, and i cant capture it on my phone.

using windows HDR in conjuntion with MADVR still works as well which is odd, i've no idea whats going on really, DMU has a better handle on things than I do.

https://1drv.ms/u/s!AgvFafeelEBij9hlrbMjllvxvSibPA?e=FSDSab

[DMU]

I also think that AMD are cheating - сustomize the results for initially incorrect image output. Sounds like tone mapping. Some images show strong clipping in the highlights.
Pic. (https://drive.google.com/open?id=1gE5E65EOi0vlu3NvPEBFr4Wq5qwZy_4P)

But in general it looks VERY similar to the correct HDR.

[mclingo]

exactly !, i'm not losing my mind afterall :) OPK, you really need to have a look at your setup mate, something really odd going on there, you should be seeing what myself and DMU are, I think the drop to limited is something to do with your setup and its not a driver issue, although its really odd that going back to .4 solves that?

[oldpainlesskodi]

Out of curiosity, what version of madvr are you on? Using DVXA or DX11?

[mclingo]

Hi, its in my sig, latest version 92.17 + 113 beta copied over the top.

[oldpainlesskodi]

Ok, figured it ok. It's MadVR's tone mapping with HDR output kicking in. Strange it only happens on the one driver.

Anyways, all good.

[mclingo]

I still dont understand whats happening mate, what win version are you on, i've stayed put on 1809 for now

[oldpainlesskodi]

On the latest fastring (need to update my sig)....

I've tried older drivers, and newer drivers....with Madvr HDR passthrough (no tone mapping), makes no difference regardless of driver. With Madr HDR passthrough AND tonemapping, apart from driver 20.1.4, all other drivers results in fine black levels, until you press pause or fast forward/back, and then it looks like it blacks levels become grey like/washed out...

Hope that makes sense. Maybe others can try and confirm (I dont think in it's Windows version issue - just a gut feeling).

K

[huhn]

amd doesn't even have a driver for the new WDDM so just layback.

[oldpainlesskodi]

Could be that hunn. Just odd that 20.1.4 is less broken lol.

[mclingo]

I also think that AMD are cheating -[/URL]

.


out of interest, as you have older NAVI hardware than my relatively recent 5700 card, is there a driver which does work properly with HDR that you can use , or has this never worked for you either, since day 1 on your hardware?

[huhn]

5700 and 5700 XT are the first navi based GPU.
the generations before that are vega and polaris.

[mclingo]

sorry yes i meant vega, I was just wondering if we could work out the point vega stopped working properly we could give AMD something to work with.

[DMU]

Starting with this version
• 19.5.1 2019-05-13 3840x2160 59/60Hz 4:2:0 10 bit doesn't work, but up to 50 Hz it works
there was no BT.2020 flag.
What happened to this version, I do not know, since I did not have HDFury.
Install version 19.4.3 does not work.

[Manni]

New nVidia driver V442.59 released today.

[mclingo]

[QUOTE=oldpainlesskodi;[/QUOTE]

hi, i put win 1909 on this weekend, no issues found if you are holding off for any reason.

[Manni]

hi, i put win 1909 on this weekend, no issues found if you are holding off for any reason.

Not with AMD, maybe, but with nVidia it breaks the NV control panel. It takes minutes to access the panel and to change some of the options. It's a real pain. I had no such issue with 1809. All I know if had I known, I'd have stayed with 1809.

It also breaks E-GPU compatibility for those with a macbook-pro and an E-GPU.

So if you're still on 1809 and have no reason to upgrade, I advise to stick with it.

[el Filou]

No problem opening NV control panel (441.12) on 1909 here.
Are you using the DCH driver by any chance? If yes then an idea is to try the standard one, or vice versa if you're using the standard.

[Manni]

No problem opening NV control panel (441.12) on 1909 here.
Are you using the DCH driver by any chance? If yes then an idea is to try the standard one, or vice versa if you're using the standard.

Thanks, I tried both when I started having the issue following the windows upgrade to 1909, it didn't make any difference.

I know that not everyone has this issue, but if you do, it's a real pain.

No amount of reinstalling, clean install, etc makes any difference. Even going to much older drivers (I have a 1080ti) doesn't help.

[huhn]

you have not by any chance installed the 3d fix?

[Manni]

you have not by any chance installed the 3d fix?

I might have at some point. Could that be the reason? Can I get rid of it?

[huhn]

it's a known issue with DCH driver disabling 3D in the driver lower the problem drastically.

uninstalling 3D vision should fix this to but installing it again is not fun... and madshi seems interested.

[Manni]

it's a known issue with DCH driver disabling 3D in the driver lower the problem drastically.

uninstalling 3D vision should fix this to but installing it again is not fun... and madshi seems interested.

I have done many clean install so 3D Vision hasn’t been installed for many moons. That’s not the issue.

[GoodGravyMan]

Is there a known issue with the RX580 with MadVR and outputting 10 bit color? If i enable 10 bit output in the AMD settings, every time i seem to go into D3D11 fullscreen window mode (10bit) my black levels get washed out. It seems fine in windowed mode (8bit).

I've tried a few versions of the driver and 20.1.4, 20.2.2, 19.7.5, 19.9.2 and they all seem to do the same thing.

If i set the output to 8bit in the AMD settings it seems like everything works.

you can see it clearly in the black clipping test here:

https://imgur.com/a/ES9Xtia

[oldpainlesskodi]

hi, i put win 1909 on this weekend, no issues found if you are holding off for any reason.

I'm on 19582.1001, but good to know you have no issues.

[NoTechi]

AMD Adrenalin 20.3.1 released (https://www.amd.com/en/support/kb/release-notes/rn-rad-win-20-3-1)

NoTechi

[DMU]

Congratulations to all AMD owners, finally, with the correct HDR BT.2020 flag.

[mclingo]

something aint right, I run my desktop in 1080p and i cant get full RGB, it keeps defaulting back to 444 ybcr

[mclingo]

yes HDR is fully working but I defo cant get full RGB for 1080p at all, anyone else?

[mclingo]

back to 2.2 for me with broken HDR, I cant achive 3d in KODI DS at all unless I drop my desktop to 1080p first which then washes out the deskop and kodi interface into a grey mess as its now ybcr444, it plays find after that, this is much more hassle for me than just turning on windows HDR before starting a movie.

The reason I run my desktop in 1080p is two fold, firstly playing a 3d movie in KODI DS and MADVR has always been a bit buggy both playing and stopping the movie, madvr makes a mess of dropping back from 1080p to 4k desktop which often results in loss of HDMI, running your desktop in 1080p is more stable.

the other reason is I connect to my PC from work via 1080p screen, if I run my desktop in 4k at home I have to log off and on again as windows messes with the DPI, its just easier all round to run a 1080p desktop with a 4k rig I find as movies and games will still load and play in 4k or whatever you set them to run in.

i'm now pining for my old RX580 which just... worked...

[oldpainlesskodi]

Congratulations to all AMD owners, finally, with the correct HDR BT.2020 flag.

Yep, all working great here...so far......

UPDATE - correct color flag, but the image is way undersaturated. I can compensate through adjustments, but...

Anyone else have this issue?

UPDATE 2 - scrub that, 2 reboots later and all good.

[Klaus1189]

Congratulations to all AMD owners, finally, with the correct HDR BT.2020 flag.

How do you test this? What do I need for it? special brand TV? ...


something aint right, I run my desktop in 1080p and i cant get full RGB, it keeps defaulting back to 444 ybcr

I use an UHD screen and switched to 1920x1080 -> restart -> switched from RGB full to YCbCr 4:4:4 and back to RGB full -> I can not confirm.
But I did wipe the old driver with AMD Cleanup Utility (https://www.amd.com/en/support/kb/faq/gpu-601), also listed in first post.


Yep, all working great here...so far......

UPDATE - correct color flag, but the image is way undersaturated. I can compensate through adjustments, but...

Anyone else have this issue?

UPDATE 2 - scrub that, 2 reboots later and all good.

So color flag is correct now and also saturation and all other stuff is fine? My "wooden eyes" can not see any issue ;-)

How about Vega cards? Everything fixed now? Was it already before?


I ask, because I want to summarize things in 1st post ;)

[oldpainlesskodi]

Yes, it now sends BT2020 flag in the AMD HDR API, so thats fixed with a NAVI card.

[DMU]

How do you test this? What do I need for it? special brand TV? ...
Samsung TV can show it. Picture -> Expert Settings-> Colour Space Settings.
And i have HDFury.


How about Vega cards? Everything fixed now? Was it already before?
Vega also had an issue like Navi cards. See my signature.

[Klaus1189]

Sorry, but can you please create a snapshot for me, where I can read the info about BT.2020 in Samsung OSD?

In the signature I see that there is Vega 8 but I need to know was it fixed already in the past or also with this new release, I remember a blown red or so with Vega?

[DMU]

Sorry, but can you please create a snapshot for me, where I can read the info about BT.2020 in Samsung OSD?
Pic. (https://drive.google.com/open?id=1WCioSIElyKfg51sn5zFc-oVLwUUA2VPm)

In the signature I see that there is Vega 8 but I need to know was it fixed already in the past or also with this new release, I remember a blown red or so with Vega?
This problem has been identified here a long time ago. NAVI cards were not there then.

[mclingo]

hi, please can someone with a 4k TV switch to 1080p 60hz and see if they can select full 4:4:4 FULL RGB please, dropping back to an old driver fixes this for me so cant just be me surely.

i've tried a full driver wipe with DDU no joy.

EDIT, I think this driver doesn't like my LG OLED, it works fine with my 4k samsung N8000, well cheesed off.

[DMU]

@mclingo
All OK here...

[mclingo]

Thanks DMU.

If I could get 3d to work with KODI DS i wouldnt mind but when my desktop and kodi gui is in 4k 3D just wont trigger at all. If I put KODI DS is 1080p it works but as gui is now in ybcr its all washed out and nasty, plus when I come out of KODI is more than often doesnt drop back to 4k so i'm forever swithing back to 4k, just a total nightmare.

3d plays fine in MPC from a 4k desktop so that bug is just with KODI DS which again will never get fixed unfortunately, i cant win, there is something wrong with every F**king driver, its infuriating.

[Klaus1189]

@DMU: I can only find an option to switch between Auto and Native. What am I missing? Please help me :)

[DMU]

@Klaus1189: You need to be in not the PC mode and choose Custom.

[Klaus1189]

Thank you

[webbo]

Any feedback, please for the latest nVidia driver, regarding HDR passthrough?
Highly appreciated
Especially regarding madVR passthrough, that's what i'm interested in

And how i can check myself within the CTRL+J display info OSD from madVR if the metadata is bogus or correct

Thanks a lot

[Klaus1189]

The latest Nvidia drivers are fine as far as I know.

In Display Stats you can‘t check it. You need to measure it. See the post I linked to in the first post with the broken Nvidia drivers.

[SamuriHL]

The latest nVidia drivers work fine for HDR, yes.

[Magik Mark]

Problem with the GForce v445.75

When running HDR videos, it switches back to SDR when UI of the player is activated.

Sometimes, it can no longer switch back to the right dynamic range

[VBB]

Confirmed. Seems to be a bug in the latest Nvidia drivers.

[VBB]

The issue is with the bit depth. Anything higher than 8-bit causes the display to drop out of HDR when a UI is triggered.

[tyguy]

Thank You, guys, CRU 1.4.1 with DisplayID also solved my above mentioned issue. That's what I did:
- wrote down the madvr timings, removed the madvr created custom resolution, then restart
- in CRU 1.4.1
-- add DisplayID to Extension blocks and set it as the 1st entry
-- set madvr timings and set Pixel clock and Native as well

It's a big help to get rid of a major annoyance, now I can set 23p as the default refresh rate of the TV :)

Couple of notes:

The difference between CRU and nvidia created (by madvr) custom timing is:
- nvidia adds a new one to the already existing ones: that can confuse applications which one to use (e.g. madvr)
- CRU replaces an existing one with the custom one on the OS level, so there won't be any confusion

When you create custom timings with madvr (http://madvr.com/crt/CustomResTutorial.html):
- always start from the default one (EDID) and pay special attention to the "sync" attributes (e.g. + , +):
-- don't select a mode that is different!
- don't deviate a lot from the default EDID values
-- remember, this approach is modifying "back porch" (horizontal and vertical values) only!
- the above 2 wrong settings can easily modify the response of the panel! (e.g. modifying gamma curve, causing black crush, etc.)
- if you get >= 10 hours after the first optimisation then it's already really good, you don't need to do more runs

Examples, original EDID timing:

EDID/CTA:
frontsywi backpixels sync
hor1276 882961660 34805500+
ver 8 10 72 90 21602250+
picl 296.70 mhz
23.9757575757576
OSD result: 23.97791Hz (~4.41 minutes)


Result after the first optimisation run:

Custom:
frontsywi backpixels sync
hor1276 88 3561720 3480 5560+
ver 8 10 72 90 21602250+
picl 299.94 mhz
23.9760191846523
OSD result: 23.97537Hz (~22 hours - 1 day)



About nvidia driver versions:
- some of them can block custom timings completely (doesn't matter how they were created)
- so always check in madvr OSD whether your timing is applied if you install a new driver

Btw, I still use the Display Changer II (https://forum.doom9.org/showthread.php?p=1863306#post1863306) util to easily manage 2 displays.
I was able to create a custom resolution but nvidia displays it as 24 hz and I have to manually activate it for it to work in mad vr. Madvr won’t automatically switch to this exact resolution. Is there a way to do that?

Edit: It looks like you use the hack direct 3D for 24 and 60 hz option.

[SamuriHL]

The issue is with the bit depth. Anything higher than 8-bit causes the display to drop out of HDR when a UI is triggered.

Oh that's going to irritate the hell out of me. I've not watched a movie since I upgraded to the latest driver and I've got the Star Wars UHD collection coming tomorrow (I hope). Guess I'll be rolling back cause um NO.

[VBB]

You'll be fine. Just don't trigger any overlay ;)

[SamuriHL]

Meh no thanks. :D I reverted the driver. I've been gaming on the new laptop anyway so I don't (currently) need the latest and greatest driver. I'm assuming at some point nVidia will fix it in an upcoming release. Not that my movies shipped on time anyway. Tracking numbers with no actual shipment should be illegal. LOL

[Cowboydude99]

Hi all,
I just found this thread...

I'm having issues with my NVIDIA GTX 1080 and JVC PJ.

I use MADVR and JRIVER.

What are the best drivers to use? I do like to game, but it might be a while before I get the chance to game anyway...

Any drivers kind of recent that would allow me to use 444:12bit?

I have a ton of UHD disks.... but can't get HDR working nicely.

Color space shows BT.709 (Best Guess) in MADVR display...

Very frustrating. I just updated to 445.75, then literally found this list/post.

[huhn]

best guess means the file doesn't have meta data so you have to fix this by hand.

that pretty much means the file is broken or the source filter/decoder can't read it doesn't. this is usually the file.

[NikosD]

Hello.

I'm using Win 10 x64 1909 (18363.752) with nVidia 1660 (442.74) connected on a UHD TV using HDMI 2.0.

When native 4K@60Hz resolution/ refresh rate is selected along with HDR from Windows settings, I get only 8bit option for color depth.
My other options selected are RGB for color format and Full RGB.

The diagnostics confirm the settings reporting "8bit with dithering"

If I turn HDR off from Windows settings, I get 8bit & 12bit color depth options.

a) Where is 10bit option, why is missing and I get 12bit with HDR off ?

b) How can I get 10bit using 4K@60Hz with HDR on ?

I mean using 4:2:2 or 4:2:0 options, because I think 4:4:4 is impossible using HDMI 2.0

[huhn]

for some reason nvidia doesn't like 10 bit as a format it does it but if 12 is possible it is doing 12.

you talk about full range RGB and that you know you need 4:2:0 or 4:2:2 to get 10 or 12 bit at 60 which is YCbCr.
just stick to 8 bit RGB it's "better" anyway.

[NikosD]

for some reason nvidia doesn't like 10 bit as a format it does it but if 12 is possible it is doing 12. I think nVidia likes 10bit for DisplayPort only.
Which is definitely weird and it's not common for SmartTVs


You talk about full range RGB and that you know you need 4:2:0 or 4:2:2 to get 10 or 12 bit at 60 which is YCbCr. Last time I checked there was no YCbCr option.
The control panel offers RGB full or RGB limited.
I'll check it again.

[huhn]

just stick to RGB.

and well displayport devices often doesn't support 12 bit and that's pretty much the only reason.

[NikosD]

12 bit color depth support is simply non existent natively in almost all HDTVs and extremely rare in all types of monitors in general.

Also, 12 bit color needs more bandwidth.

The decision of nVidia is weird and there is no reason to defend it.

[huhn]

i'm not defending it...

and 12 bit is pretty much mandatory for TV if the TV would only support 10 bit or if 10 bit would be possible with the bandwidth but not 12 nvidia can do 10 that's why UHD 60 HZ 4.2:2 10 bit exists 12 is not possible with the pathetic HDMI spec.

is it dumb to remove the option to send less sure...
12 bit would be better in theory and mathematically but not in practise.

[NikosD]

It is surely weird and possibly wrong to support 8bit, to skip 10bit and to support 12bit again.

Weird for sure.

HDMI v2.1 supports everything with huge bandwidth but it's still so rare.

[Cowboydude99]

Am I the only person that has really terrible 23fps/24fps with NVIDIA cards?

I'm on GTX 1080. When playing any 1080p/24fps video in any player, it is like a fast slideshow. It's stop-go-stop-go. Very annoying.

I'm using a projector, so no motion interpolation from the display device.

All drivers seem to be the same.

If I play a 4k/60 file (Billy Lynn's Long whatever) it plays just fine and smooth.

[huhn]

screen of the OSD ?

you had ffdshow installed so you doing quite some harmful things.

[Cowboydude99]

Hi again. Thanks for the response. I'm not sure what is going on, even if I just use KODI 's defaults, it's still choppy.

Here's the screenshot of my madvr OSD:
https://drive.google.com/open?id=1jDuWhzBVSBk5vjWvQlS5fB0FkQbaL0nW

[huhn]

i don't know why you are using DXVA2 but there seems to be an issue with it on your system.

[NikosD]

I think nVidia likes 10bit for DisplayPort only.
Which is definitely weird and it's not common for SmartTVs

Last time I checked there was no YCbCr option.
The control panel offers RGB full or RGB limited.
I'll check it again. Using a custom resolution 1080p@144 Hz created in nVidia's Control Panel, not exposed initially by the display, I get very limited options for configuration.

Just RGB - 8bit color - Full (0-255) or Limited (16-235) dynamic range - HDR on/off

But using a native resolution like 3840x2160 I get:

60Hz - RGB - 8bit (only) - Full or Limited - HDR on/off

60Hz - YCbCr420 - 8bit (only) - Limited (only) - HDR off (only)

60Hz - YCbCr422 - 8bit - Limited (only) - HDR off (only)
60Hz - YCbCr422 - 10bit/12bit - Limited (only) - HDR on/off

60Hz - YCbCr444 - 8bit - Limited (only) - HDR on/off
30Hz - YCbCr444 - 12bit - Limited (only) - HDR on/off

So the only configuration for 4K@60Hz using HDR and 10bit color is YCbCr422, with only a minor limitation of Limited dynamic range.

@huhn
What would you prefer/ is "better" ?

A) 60Hz - RGB - 8bit - Full - HDR on
B) 60Hz - YCbCr422 - 10bit - Limited - HDR on
C) 60Hz - YCbCr444 - 8bit - Limited - HDR on

A or B or C ?

[WuNgUn]

This is a GOOD question... I've wondered this as well.

Sent from my SM-G965W using Tapatalk

[Asmodian]

A) 60Hz - RGB - 8bit - Full - HDR on

[el Filou]

Here's the screenshot of my madvr OSD:
https://drive.google.com/open?id=1jDuWhzBVSBk5vjWvQlS5fB0FkQbaL0nWYour decoding queue can't keep up and it looks like it's impacting rendering queue.
Are you using DXVA copyback? If you're not using black bar detection and never watch interlaced material, you can try setting the decoder mode to D3D11 Native and see if it helps.

Also, you're playing a 23,976 fps file with a 24,000 Hz display refresh, so you're gonna have a dropped frame every 30 seconds or so at best even if everything else is perfect.

Edit: and I don't see the point of using madVR on a GTX 1080 if you're gonna use DXVA processing instead of some higher quality algos, tbh.

[NikosD]

A) 60Hz - RGB - 8bit - Full - HDR on One interesting comment I read in one from the infinite discussions around this matter:

Only use RGB 8-bit for everything on a PC, including HDR games and movies, even when connected to a HDR TV over HDMI.
The GPU does dithering for 10-bit content and there will be no banding.

A 10-bit signal to the display is only required when the source doesn't perform dithering (PS4 , Blu-ray player, etc.).
If the PS4 did dithering, it could run RGB 8-bit 60 Hz instead of subsampling at YCbCr420 10-bit 60 Hz because there isn't enough bandwidth for RGB 10-bit 60 Hz over HDMI 2.0.

If the display is 8-bit + FRC, the 10-bit signal is dithered internally by the display anyway.
A true 10-bit panel is pointless since the quantization noise on 8-bit + dithering is invisible.

On Windows, HDR apps render to a 10-bit surface and the GPU does dithering automatically if the signal is 8-bit.
So you should just use 8-bit RGB for maximum quality.

You can verify this with the banding test in the VESA DisplayHDR app in the Windows Store.For 8bit panels it is obvious that Full RGB 8bit 4K@60Hz is the best option.

The only debate is around HDR content (games, videos) using a true 10bit HDR panel at 4K@60Hz on Windows PC.

[huhn]

if you know how a computer works it's still C.

[NikosD]

Asmodian said A) you said C) and knowing how computers work I would say B)

The whole thing is a matter of compromises and implementations.

Windows HDR model has gone through a lot of changes lately, who knows the exact situation of this.

OEM drivers of GPUs are not the best components to trust, they have so many bugs and break things occasionally that used to work.

I choose 10bit native color even with 4:2:2 over any other solution.

Although in reality, I'm not sure that the difference is easy to tell.

[huhn]

ohh yeah sure A.

subsampling is very easy to tell.

[Asmodian]

I really don't understand this preference for subsampling, subsampling to 4:2:2 is an ancient compression scheme that drops 1/3 of the video data in a fairly inefficient way in order to lower bandwidth requirements. madVr dithers quite well so 8 bit and 10 bit end up visually identical while 4:2:2 is visually worse.

If you know how computers work and what the various formats are why would you choose subsampling?! If you don't trust your GPU to handle bit depths properly I would suspect the 10 bit paths a lot more than the 8 bit ones. Also, why do you trust the drivers to resample both chroma planes to 50% horizontal resolution while keeping proper 10 bit?

A and C are very similar, while B is obviously worse. A is the only option that does not rely on the GPU driver doing something correctly.

Edit: All the HDR Games I have played dithered as well, 8 bit at 30 Hz is indistinguishable from 10 bit at 30 Hz (30 Hz only for the comparison). HDR with low latency and 444 chroma is poor quality on my TV though so I don't do much HDR gaming. If your display is subsampling internally B is fine, but not actually better.

[tyguy]

I really don't understand this preference for subsampling, subsampling to 4:2:2 is an ancient compression scheme that drops 1/3 of the video data in a fairly inefficient way in order to lower bandwidth requirements. madVr dithers quite well so 8 bit and 10 bit end up visually identical while 4:2:2 is visually worse.

If you know how computers work and what the various formats are why would you choose subsampling?! If you don't trust your GPU to handle bit depths properly I would suspect the 10 bit paths a lot more than the 8 bit ones. Also, why do you trust the drivers to resample both chroma planes to 50% horizontal resolution while keeping proper 10 bit?

A and C are very similar, while B is obviously worse. A is the only option that does not rely on the GPU driver doing something correctly.
Are you using pc input label for everything sdr but switch to a different label like generic hdmi or game console for hdr content with your c9? Id love to get full chroma 100 percent of the time, but hdr is just horrible in pc mode. Grays are banded, clouds look like oil floating in water in dark scenes. Also, I thought you used ycbcr 444 with your setup?

[Asmodian]

Well, this was a response to a general "what is better" question. If asked what is better on a LG C9 in HDR mode it changes due to faults of the C9.

I use C) YCbCr 444 limited, 8 bit, 60 Hz, Nvidia HDR On with this TV. I do change the label of the HDMI port to Home Theater when I watch HDR content but I keep the same output format. This of course subsamples in the TV but it seems to do a decent job. Sending it 10 bit 422 does not look any better, in fact it seems to be slightly worse with more banding because this TV doesn't handle 10 bit (or RGB) input very well. Others disagree, and I only notice subtle differences between all input formats in Home Theater mode, but in PC mode 10 bit and/or RGB is obviously worse. :mad:

This all gets pretty complicated when we bring the failings of individual displays into it. :(

[Klaus1189]

Sorry for interrupting you guys, I created a dedicated thread for issues or failings of individual displays here:
https://forum.doom9.org/showthread.php?t=180421

If you want to post some info there, that would be very nice.

I want to collect as much info as possible, so it is easier for us all to find better tvs in the future

I summarize the info in the first post. It is almost ridiculous of how many issues modern TVs can have.

Please go on, very informative here ;)

[tyguy]

Well, this was a response to a general "what is better" question. If asked what is better on a LG C9 in HDR mode it changes due to faults of the C9.

I use C) YCbCr 444 limited, 8 bit, 60 Hz, Nvidia HDR On with this TV. I do change the label of the HDMI port to Home Theater when I watch HDR content but I keep the same output format. This of course subsamples in the TV but it seems to do a decent job. Sending it 10 bit 422 does not look any better, in fact it seems to be slightly worse with more banding because this TV doesn't handle 10 bit (or RGB) input very well. Others disagree, and I only notice subtle differences between all input formats in Home Theater mode, but in PC mode 10 bit and/or RGB is obviously worse. :mad:

This all gets pretty complicated when we bring the failings of individual displays into it. :(
Do you use pc mode for sdr pc games and sdr movies?

[Asmodian]

Absolutely, I really hate subsampling.

[tyguy]

Absolutely, I really hate subsampling.
I'm wondering if the harmony remote can be programmed to change input labels. that way we only have to press 1 button instead of long pressing input button, click on edit, then click on input, scroll down to pc, apply it, and then click in the window.

[NikosD]

@moderators

Maybe the last few posts, from my question and later, could be moved to the dedicated thread regarding TVs.

And this one too.

Regarding the issue of 4K@60Hz using 10bit with HDR on, it's a corner case that happens more frequently nowadays due to the capability of many UHD TVs , implementing that configuration but on HDMI which is incapable of transferring the data.

But the issue can be more frequent in cases that RGB is not offered in all its glory for some reason (full, 10bit, HDR, high resolution, high refresh rate) and you have to compromise between RGB and some form of subsampling.

In any case I have said in the past that I don't have such good eyes distinguishing slight differences, so I can be happy in almost all cases.

But the debate is useful to know which is better in theory and in practice to check what is better in your specific configuration.

[huhn]

you wouldn't believe how much this topic has been tested and discussed on this forum.

welcome to the world of TV insanity have a small taste:
https://forum.doom9.org/showthread.php?t=180421&page=2

[NikosD]

Yes, that thread I meant of Klaus1189 if it wasn't clear in my previous post.

[x7007]

Which latest nv driver works with HDR MadVR because 445.75 doesn't work. the last one worked was 442.5x or something but it has other weird issues.

[Klaus1189]

Try latest Studio driver 442.19

[VBB]

445.75 HDR works just fine. The only issue is that it exits HDR mode when you trigger a player UI, but it goes right back after. Note that this happens only if you use 10-bit or higher output in the Nvidia control panel. I will try 445.87 tonight, but I doubt this is fixed.

[VBB]

Just installed the latest Nvidia 445.87 driver. As predicted, the UI issue has not been fixed. It still switches out of 10-bit/12-bit full-screen windowed and into 8-bit windowed mode in both HDR and SDR once you pull up a player menu or status bar. This is obvious in HDR, because it also switches the display out of HDR at the same time. Again, this doesn't really break any functionality, and once the UI disappears, the player goes right back to the original mode.

[NikosD]

@VBB
Yes, I have exactly the same problem and I also realized it better when switching from SDR to HDR and then back to SDR.

[lazostat]

So which is so far the most stable nvidia driver?

[ryrynz]

They're pretty much all stable. Just use the latest one.

[Cowboydude99]

Hi all.
What is the recommended NVIDIA driver for Windows 10 with 10b HDR, running 4K/60? I am also a JVC projector owner...

Lots of weird issues with this set up.
Sometimes after stopping HDR, the screen just stays black, other times it is a green tint with tons of random lines/noise running in it. Happens on multiple HDMI cables.

Is the studio driver the way to go? What about if you also game sometimes?

[huhn]

just try the current one with 8 bit RGB

[Klaus1189]

What driver are you using at the moment?
With Nvidia you won‘t get 10 bit HDR, but that is really no dealbreaker, just use RGB 8 bit 3840x2160 and you should be fine, at least if there is no issue with this JVC projector.

You won‘t recognize the difference between 8 bit and 10 bit.

It depends on the games, if you want to game a very new one and the driver crashes or has issues with it you will have to use the newest which is designed to fix the issue with that specific game. Older games are normally fine.

Try latest studio driver from yesterday, do you have any issue with it? It is the best tested driver branch for now.

[Manni]

Hi all.
What is the recommended NVIDIA driver for Windows 10 with 10b HDR, running 4K/60? I am also a JVC projector owner...

Lots of weird issues with this set up.
Sometimes after stopping HDR, the screen just stays black, other times it is a green tint with tons of random lines/noise running in it. Happens on multiple HDMI cables.

Is the studio driver the way to go? What about if you also game sometimes?

If you're a JVC owner, don't use 10bits, especially if you do games or desktop work. If you send anything above 8bits in 4K 4:4:4 (whether RGB or YCC), JVC does an internal downconversion to YCC 4:2:2 behind madVR's back that damages chroma. Sending 12bits 422 is a workaround that prevents you from making the most of madVR's chroma upscaling, so leads to a worse picture than 8bits 4:4:4, especially when upscaling 1080p and lower res to 4K.

This is on all models, including the latest 4K (I have an RS2000). When I reported this to JVC, instead of fixing it they made all options in the colorspace menu (including RGB and YCC444) send YCC422. That way it's not obvious anymore (before it was breaking the picture), but there is probably an internal limitation that prevents them from accepting 4K 12bits 4:4:4 even when there it fits the 18Gb/s bandwidth limitation, for example at 30p and below.

Use 8bits, it works at all resolutions / frame rates and it doesn't cause any banding (provided madVR dithering isn't disabled) even with 10bits HDR.

There are too many issues in 12bits with the nVidia drivers: black levels mismatch, colorspace/calibration issues, just forget about it. I use 442.74 in 4K 8bits RGB Full, MadVR set to 16-235 and the JVC set to HDMI standard, it works fine. If you need gaming/desktop, you can set madVR to 0-255 and the JVC to HDMI Enhanced. I use a CRU custom resolution and I get no frame drops for a day or more.

You won't be able to see any difference PQ-wise if you use 8bits. You might see a very slight increase in noise level near black if you stick your nose to the screen, but no added banding. It will just work. :)

[Cowboydude99]

I have sorted the issue.

I had ticked ‘Report BT.2020’ under my calibration settings in Mad VR for BT.709. Once I unticked this, the magenta issue was gone.

Your post from 1 year ago really saved me from going down a new projector rabbit hole (I was already half way in).

Unticking this box, "Report BT2020 (NVIDIA ONLY)" in MADVR solved 2 issues:

1. Green scrambled screen.
2. Magenta colored display.

I no longer have to change the hz/resolution/color after stopping/starting video.

As of now I'm getting 4K/60/422 12 bit from my GPU to my PJ.

Smooth motion is helping a lot with the judder!

Thank you all for your help!

[NoTechi]

If you're a JVC owner, don't use 10bits, especially if you do games or desktop work.
.....
You won't be able to see any difference PQ-wise if you use 8bits. You might see a very slight increase in noise level near black if you stick your nose to the screen, but no added banding. It will just work. :)

Thanks Manni for the good explanation, which finally convinced me to use 8bit. :D
At some point I switched from Epson to JVC because I had strong banding with Epson especcialy in games and couldn't use 10/12 bit to try to mitigate it. That made me hesitate to switch to 8bit voluntarily until now since the JVC being "capable" to use 10/12 bit was one major selling point to me.

Just to be on the safe side, I have to select 8bit within the gpu driver and madvr display properties, right?

NoTechi

[huhn]

yes in the driver and madVR.

banding in computer games is pretty normal it just depends on the game.
use a banding test with madVR to make sure it's not the device.

[claw]

Thanks Manni for the good explanation, which finally convinced me to use 8bit. :D
At some point I switched from Epson to JVC because I had strong banding with Epson especcialy in games and couldn't use 10/12 bit to try to mitigate it. That made me hesitate to switch to 8bit voluntarily until now since the JVC being "capable" to use 10/12 bit was one major selling point to me.

Just to be on the safe side, I have to select 8bit within the gpu driver and madvr display properties, right?

NoTechi

I recently switched from RGB 12-bit to RGB 8-bit as well with a JVC RS500.

I had to disable Report BT2020 in MadVR to avoid the magenta bug in the JVC.

[Manni]

Just to be on the safe side, I have to select 8bit within the gpu driver and madvr display properties, right?

Correct, 8bits in both. Never use Auto for bit depth in madVR, it’s not meant to detect the bitdepth.

[brazen1]

You guys are aware that nVidia drivers are back to retaining 12bit after a reboot right?

[sat4all]

Yeah noticed that too, but forgot to alert the folks over here. :p

[Manni]

You guys are aware that nVidia drivers are back to retaining 12bit after a reboot right?

Unfortunately there is still a mismatch in black levels between sdr and hdr unless you set madVR to send different levels (16-235 in SDR, 0-255 in HDR), which isn’t ideal if the pj is set to video levels.

If anyone knows a correct combo that works with the JVCs, please post.

Of course there is the forced chroma downconversion to YCC 4:2:2 behind madVR’s back in 12bits with the JVCs (and I suspect most displays/projectors), but let’s put that aside.

[mzso]

Hi!

Probably a dumb question, but how do I get my nvidia card (ENGTS450) to produce 4:4:4 with a TV? I see no settings or even what subsampling is used shown anywhere.
But the test image (https://www.rtings.com/images/test-materials/2017/chroma-444.png) I just found looks like crap in the blue rows and the last two, I guess the video signal is sub-sampled.

[huhn]

the issue is usually the display not the graphics card setting.
becaues for so reason it's cool to subsample incoming signal.
PC mode is usually needed and there are a lot of older TV that can't do 4:4:4 at all.

[mzso]

the issue is usually the display not the graphics card setting.
becaues for so reason it's cool to subsample incoming signal.
PC mode is usually needed and there are a lot of older TV that can't do 4:4:4 at all.

I see. Thanks for the tip.
What I found is that if I enable the "1080p Piexl by 4pixels" option I get correct text. Though I'm nut sure why. Up-scaling to UHD can't be this dreadful... Maybe some sort of subsampling is active until I enable this for some reason.

The "1080p Pure Direct" option also fixes the appearance. And looking at it now closely it looks like I get proper up-scaling. (if I enable the former option atop of this I get the more aliased text of native 1080p.)

Anyway. I still have more (possibly) on-topic issues.
Such as I can't figure out why resolutions above FHD appear in the driver's or Windows' settings. The TV is UHD and the card goes up to 1600p.
Hmm. As I'm writing this I notice that only this is specified by the manufacturer: "D-Sub Max Resolution : 2048x1536, DVI Max Resolution : 2560x1600"
Could it be that the HDMI port only goes up to 1080p, and I have to use the DVI with and adapter?

[huhn]

just to tell you in the first place pure direct mode breaks 23/24p output is a panasonic special they do this intentionally with pride.

it's HDMI 1.3a so try a lower refreshrate and it can't do 3d and neither offically UHD (but the bandwidth is the same as 1.4).
scroll down to PC resolutions is may be listed there or try a custom resolution.
DVI -> HDMI shouldn't work on passive it's just sending the same HDMI signal and a active cable would cost more then your GPU.
just get a "new" GPU.

[NoTechi]

New AMD Adrenalin 20.4.2 driver (https://www.amd.com/en/support/kb/release-notes/rn-rad-win-20-4-2?utm_source=pardot&utm_content=&utm_campaign=2020-04-23_Driver_Alert-Radeon_Software-Release20.4.2-apr-de-email&utm_medium=email&utm_term=) available

[mzso]

just to tell you in the first place pure direct mode breaks 23/24p output is a panasonic special they do this intentionally with pride.
In what way is it broken?
I could create 23/24 custom res for 1440p but not for 2048p.




it's HDMI 1.3a so try a lower refreshrate and it can't do 3d and neither offically UHD (but the bandwidth is the same as 1.4).
scroll down to PC resolutions is may be listed there or try a custom resolution.
The port? How can you tell? I can't see it on the manufacturer's specification's list. I managed to get 1440p@60. So that's good at least.
(Though some desktop elements seem to have gotten magnified for some strange reason.) On the other end refresh rate seems to be firmly capped at 60Hz at whatever resolution.
I guess my GPU is more outdated than I thought.


DVI -> HDMI shouldn't work on passive it's just sending the same HDMI signal and a active cable would cost more then your GPU.
just get a "new" GPU.
I was thinking that they're different ports. Maybe the HDMI is restricted to 1080p. Rather than the 2560*1600 specified for DVI.

PS:
Since You seem to be familiar with Panasonic, can you tell me what "241p Smooth Film" does? Is it just interpolation?
And what's panasonic's fantasy name for black frame insertion?

[Klaus1189]

https://forum.doom9.org/showthread.php?t=180421&page=2

Panasonic :eek:

[huhn]

smooth film is interpolation if the TV get 23p and if it get's 60 and has a 120 Hz panel it has intelligent frame creation which is the same again both are known to not work with pure direct mode.

my TV is still a firefox OS things may have changed.

[QUOTE]HDMI 1.3a support: Fully integrated support for HDMI 1.3a
including xvYCC, deep color, and 7.1 digital surround sound
including Dolby TrueHD and DTS-HD. Upgradeable to HDMI
1.4 support for 3D formats with upcoming NVIDIA 3DTV Play
software. Refer to http://www.nvidia.com/3dtv for more details./QUOTE]
source: https://www.nvidia.com/content/geforce-gtx/GTS_450_User_Guide.pdf

it was 1.3a on release but whatever you can't do 120 hz (i have my doubt the TV can do it anyway) because you need HDMI 2.0 for 1080p120 there is no technically reason for that the specs wants it.

windows 10 doesn't like a screens which are bigger and is apply scaling automatic if it detect it you can change this in the display settings.

if you are enabling pure direct mode to get full chroma resolution and you are sending a 23/24p signal it will go and do a 3:2 pulldown (because we are on doom9 i should say it's progressive pulldown) to 60p even on a 120 HZ screen so a 6:4 pulldown (someone shoot me for saying this please). it's totally intentional and it's beyond me. panasonic doesn't know what blackframe interpolation is they may use it for frame interpolation but you can't control it individual like on a samsung or even better on sony which can do that without subsampling. there blackframe interpolation is not known to be good and more used to get some "math" behind there frame rate claim.
i mean 120 frames a sec with a 10 step blackframe interpolation is clearly 1200 HZ isn't it?

[Charky]

Quick question for those who kept track of this stuff.

I've used old 385.28 nvidia drivers for ages. Windows 10 forced an update and i'm pissed.

Before I downgrade to my old drivers, can anyone confirm that they are still the last that allow custom timings + 12 bits sticking on reboot ?

Has this changed ?

EDIT : well, apparently not.

Latest version didn't work. Couldn't create a custom res for my LG C8, even with CRU.

Reverted back to 385.28, works like a charm.

Go figure...

[WuNgUn]

The the latest nVidia drivers claim a fix for HDR... but in gaming...?
Would this also translate to video playback?

Sent from my SM-G988W using Tapatalk

[VBB]

I will check in a few to see if this latest hotfix reverts the current HDR bug. But as I've said before, HDR is fully usable with any of the recent Nvidia drivers, as long as a player UI isn't triggered for more than a second or so.

[WuNgUn]

I'm not seeing it on their driver page.... WHQL nor beta....?

Sent from my SM-G988W using Tapatalk

[VBB]

https://nvidia.custhelp.com/app/answers/detail/a_id/5022

[WuNgUn]

So I was just playing around...
4:4:4 at 12 bit and the video image is pink-washed...
4:4;2 @ 12 bit and just a black screen...
432.00

Sent from my SM-G988W using Tapatalk

[VBB]

Please remind us what your display is...

[WuNgUn]

Please remind us what your display is...75" Samsung UN75RU8000FXZC
Managed to get an image, but playback is choppy AF with that colour setting...
Testing RGB again to verify.

Edit: yup, for some reason it doesn't like playback in 4:4:2. Never seen that problem before now....


Sent from my SM-G988W using Tapatalk

[VBB]

OK, to follow up on my post above: The latest Nvidia hotfix driver 445.98, released yesterday, does NOT fix the current bug with 10-bit full-screen windowed kicking back to 8-bit windowed when pulling up a player UI.

[WuNgUn]

Well, dammit... I have dropped frames like crazy now, no matter what I do!
Ugh!

Edit: okay... Smooth playback now. Not seeing HDR kick on when toggling HDR metadata off and on in MadVR...

[VBB]

All I can tell you is, it works here. Same as every driver before it, except for that little bug that's been around for a few weeks.

[WuNgUn]

Still no HDR.... toggle metadata off and on does squat. Ugh!
And I can only get workable 4K playback with RGB full. Anything else and it drops frames.

[VBB]

That doesn't make sense. The color space has no influence on the frames. You sure you're not accidentally comparing 23Hz with 60Hz or something? And when you say HDR metadata toggle, you mean the pass-through option in madVR, or the HDR switch in Windows?

[WuNgUn]

Something is causing bad frame drop... It's like the video card is doing some extra processing to the signal, getting bogged down... Not just the video image, but mouse/laggy, etc...

Full screen I'm at 23Hz, windowed, at 60Hz, which plays smoothly. Is this what you mean?

Pass thru in MadVR, yes
I never touch Windows HDR setting

[VBB]

Did you make sure to set the Nvidia power profile from OPTIMAL to ADAPTIVE? That resets itself after every driver install, certainly if you do a clean install. Oh, and if the mouse is laggy, then you're not in 60Hz at that time.

[WuNgUn]

Did you make sure to set the Nvidia power profile from OPTIMAL to ADAPTIVE? That resets itself after every driver install, certainly if you do a clean install. Oh, and if the mouse is laggy, then you're not in 60Hz at that time.Ummm... I've never touched that before.
I don't see it in nvidia CP

[VBB]

You never changed the power profile? You can find that in the 3D settings. What card do you have?

[WuNgUn]

K... It was on Optimal... Set to Prefer max performance.
No change in FPS rendering with 4:2:2 color. Wtf!

[WuNgUn]

And HDR pass thru in MadVR messed still...
:(

[VBB]

Hmmm. I don't have too much experience with Samsung's, but it doesn't sound like the TV is at fault here. Just to make sure, you currently have the video card set to YCbCr 4:2:2 10-bit at 60Hz? If so, do you have smooth motion enabled in madVR? There would be stutter if you didn't. If your output is set to 23Hz, then the problem is elsewhere.

EDIT: I just re-read your respone above. Your desktop is at 60Hz, but madVR switches to FS 23Hz. So, there is no problem with RGB, only with YCbCr? Why not just use RGB Full 8-bit and call it a day? That might very well be the best output with the Samsung anyway.

[WuNgUn]

Hmmm. I don't have too much experience with Samsung's, but it doesn't sound like the TV is at fault here. Just to make sure, you currently have the video card set to YCbCr 4:2:2 10-bit at 60Hz? If so, do you have smooth motion enabled in madVR? There would be stutter if you didn't. If your output is set to 23Hz, then the problem is elsewhere.

No, I wouldn't say it's the TV either...
4:2:2 10 bit.... @60HZ, Correct, in desktop mode. 23Hz with video playback. MadVR is set to change the display mode with playback.
Smooth motion OFF

[WuNgUn]

With 4:2:2 colour, I have to disable "enable windows overlay mode" or all I see is a black screen...
In RGB if I don't enable the overlay, the frame drops. With it enabled, playback is smooth.

[VBB]

I'm still not sure what card you have. You should not have to use overlay at all. Have you tried D3D11? What is it set to in LAV?

[WuNgUn]

I'm still not sure what card you have. You should not have to use overlay at all. Have you tried D3D11? What is it set to in LAV?Sorry, yeah... The card is a EVGA GTX960
Ahhh sh*t... LAV was set to nVidia CUVID
DXVA copy back I believe works...
Never tried D3D11

Testing...

[VBB]

Great, same card as mine! Set it to D3D11 native in LAV, uncheck overlay in madVR, and hopefully you'll be good to go.

[WuNgUn]

Great, same card as mine! Set it to D3D11 native in LAV, uncheck overlay in madVR, and hopefully you'll be good to go.Cool! And you run at 4:2:2?

I might have to adjust buffers. Cooyback works well after 10 seconds of stuttering.

[WuNgUn]

4:2:2 still stutters like mad

[VBB]

I only run 4:2:2 10-bit because of the quirkiness of the LG OLED I use (see the other thread LOL). Any other display, and I'd very likely run RGB full 8-bit. With the 960, I use default buffers in madVR. That card, old as it is, still works great for 4K upscaling and everything else, except tone mapping. I'm waiting for the 30xx series to finally upgrade. My whole rig is super old. I built it in 2008. Runs like a champ, though.

[WuNgUn]

Windows overlay mode is much better at smooth playback. With it off, it takes a long time for the playback to smooth, and even then, it gets the occasional drop out.
I dunno...
Tone mapping? HDR?
It has no effect on my rendering other than picture quality over here... No performance hit.
But seeing as metadata pass thru is broken :(

[VBB]

Very strange, since we have the same card. I installed the latest driver today, played something in HDR. No problem. Wish I could help you more. And you did not have an issue before today?

[WuNgUn]

The hotfix driver?
That is weird.

[VBB]

Yup, running 445.98 right now.

[WuNgUn]

Yup, running 445.98 right now.Usually in windowed mode, when I toggle the pass thru off and on, I can see the display change (in windows) to that 'over saturated' look when HDR is on...
Nothing with 445.98

[VBB]

And full-screen windowed also does not trigger HDR on the TV? The bug in the current driver would cause what you're seeing in windowed mode, but FS should work.

[WuNgUn]

And full-screen windowed also does not trigger HDR on the TV? The bug in the current driver would cause what you're seeing in windowed mode, but FS should work.Not sure what you mean by "full screen windowed"...? You mean if my video player (JRiver) is in full screen but the video isn't, compared to full screen video?
Neither makes any difference

[VBB]

OK, if your player doesn't trigger HDR when in any full-screen mode, then I'm stumped. Have you tried a different player at all?

[WuNgUn]

OK, if your player doesn't trigger HDR when in any full-screen mode, then I'm stumped. Have you tried a different player at all?Well, VLC or WMP... but I have no idea how to see if it's on or off with those, assuming they even support HDR.
Windows Movie & TV wants $1 for decoder. Pfff....

[VBB]

I would try going back to the last driver that worked for you, clean install. I'll be back here tomorrow, if you need any more help. Good luck!

[VBB]

Quick update: Tonight, I set madVR output to FS Exclusive after having used FS Windowed for what seems forever now. From prior experience, I expected nothing to happen differently, mostly because more recent Windows 10 installations supposedly don't allow for exclusive mode any longer (as witnessed by myself). But lo and behold, after a reboot and not touching anything else, I start a video and it goes straight into FS Exclusive (says so in the madVR OSD). You can tell it's different from FS Windowed, because for a split second there is a sort of discoloration flash on the screen, which I had seen before when FS Exclusive still worked normally. Anyway, the reason I wanted to test this to begin with was the bit-depth bug in recent Nvidia drivers. Turns out it doesn't do it in FS Exclusive, but there's a reason for it: It doesn't pull up the normal player UI, but rather the madVR search bar (if checked in madVR). Long story short, this is useful right now, because it doesn't kick the video out of 10-bit/HDR mode. Caveat: Apparently, it's tricky to get FS Exclusive mode to stick. I was able to play three episodes of a show consecutively, but when I played a test video right after, I wasn't able to trigger it anymore, until a reboot...

[huhn]

it doesn't stick is pretty much part of the issue that FSE stop working correctly and it's fake.

if you are really into testing more FSE you can do this at your own risk:
go to the mpc-hc64.exe right click compatibility and check disable fullscreen optimisation. this is supposed to make FSE again a real FSE like the old one and it may create some serious issues like crashing your system.

[aufkrawall]

go to the mpc-hc64.exe right click compatibility and check disable fullscreen optimisation. this is supposed to make FSE again a real FSE like the old one and it may create some serious issues like crashing your system.
Is this specific to madVR HDR etc.? It works without issues in games with SDR output.

[huhn]

no but depends on the game.

games have problems with FSE to. the point of FSE is the exclusive part that means nothing else can be shown over it and that's simply not the case anymore.

[VBB]

I've had full-screen optimization for MPC-HC disabled for years now, so that's probably why I saw what I was seeing last night. But like I said, the brief flash before kicking in is an indicator that it somehow still works how it used to work, if it does work. Make sense?

[aufkrawall]

no but depends on the game.

games have problems with FSE to. the point of FSE is the exclusive part that means nothing else can be shown over it and that's simply not the case anymore.
I've never seen turning off FSO causing issues in any game with neither AMD nor Nvidia.
There btw. are no FSO for Vulkan fullscreen exclusive mode, seems graphics vendors hate this crap.
It's just D3D12 where there is only DWM direct presentation instead of FSE.

[SamuriHL]

Yea, really nice to see them separate out the driver update vs the customizations. That was starting to irritate me on my new Alienware. Good job, Intel!

[foxyshadis]

https://downloadcenter.intel.com/download/29530/Intel-Graphics-Windows-10-DCH-Drivers



Intel 26.20.100.8141 driver released.

Well that's certainly a lot nicer than having to uninstall, delete via pnputil, & reboot first, recently, or a few years ago install a modded driver package.

It's not often that customers hating on OEM lock-in overcomes Intel's willingness to bow to them, but I'm glad they gave us some freedom again.

[SamuriHL]

Interesting. Dolby vision support eh?

Sent from my SM-G975U using Tapatalk

[nsnhd]

Should it be Core-i or a Pentium with HD610 is sufficient ? I think of a G4600 with HD630.

[littleD]

Not sure what support they claim because such file
Format : HEVC
Format/Info : High Efficiency Video Coding
Format profile : Main 10@L5.1@Main
HDR format : Dolby Vision, Version 1.0,

Plays as green and pink. (Dolby Vision Profile 5 )

Edit.
Ctually
Format : HEVC
Format/Info : High Efficiency Video Coding
Format profile : Main 10@L5.1@Main
HDR format : Dolby Vision, Version 1.0, dvhe.08.09, BL+RPU, HDR10 compatible / SMPTE ST 2086, HDR10 compatible

Plays ok. Its is described as Dolby Vision Profile 8.1

[el Filou]

Are you sure the Dolby Vision support isn't just for games?
I suppose support for media playback would need some proprietary decoding filter?

[amayra]

is there a different between intel windows driver and i965 Linux driver for 7gen igpu?

[Blue_MiSfit]

Dolby Vision profile 8.1 is an HDR10 base layer, so you're likely just seeing that. There's also an enhancement RPU (reference picture unit) that lets a compliant Dolby Vision decoder get more quality and get dynamic tone mapping metadata.

Profile 5 is non-backwards compatible, so you're seeing the IPT representation of the image interpreted as YCbCr by your player that doesn't know any better :) You need a real Dolby Vision decoder to play this.

Basically it looks like maybe this driver can output Dolby Vision for games over HDMI. Maaaaybe a custom player could output Dolby Vision video as well, but not sure about that. Lots of special Dolby stuff has to be implemented :)

[aufkrawall]

is there a different between intel windows driver and i965 Linux driver for 7gen igpu?
i965 is only an OpenGL driver in Mesa and is replaced by Iris. The kernel driver module is still named i915.
The kernel drivers of Intel and AMD support a lot of what is supported on Windows these days, though graphical user space in Linux world can't catch up on HDCP and HDR support due to lack of consensus, interest or developer resources.

[VBB]

I'll test in a bit to see if the bit-depth/GUI issue is fixed.

[SamuriHL]

Would be nice. I didn't get a chance to even look at it today.

[VBB]

I did, and it's not fixed...

But in the process, I found out something I did not know: Nvidia drivers dither by default. See my post over at AVS https://www.avsforum.com/forum/40-oled-technology-flat-panels-general/3138274-2020-lg-cxa-gx-dedicated-gaming-thread-consoles-pc-16.html#post59718362

[bleakassassin]

I feel like this is gonna be one of those things that doesn't get fixed for months upon months until it just suddenly gets fixed one odd update awhile from now. Unless it's somehow something that can happen in games, too. Nvidia isn't catering to madVR and HDR video users, after all.

Out of curiosity, is there a reason this isn't mentioned in the first post? I feel like it's something that people using one of the last few drivers should know.

[VBB]

Klaus, could you add the bit-depth/GUI issue to the last few Nvidia drivers in the first post, please? "Playback at greater than 8-bit drops back to 8-bit (and out of HDR) in Windowed Full-Screen mode when player GUI is triggered, regardless of bit-depth of the file. Can be circumvented by using Full-Screen Exclusive mode or by setting madVR to 8-bit to begin with."

Danke aus Long Beach ;)

[NoTechi]

New AMD driver Adrenalin 20.5.1 (https://www.amd.com/en/support/kb/release-notes/rn-rad-win-20-5-1) released.

NoTechi

[Klaus1189]

@VBB: What was the first version affected?

Greetings from Bavaria ;)

[brazen1]

Klaus, you might also want to add that every nVidia driver after 388.59 has a faulty updated HD Audio driver in the newer Windows driver installer packages. Turn on AVR. Right click Audio Icon in tray. Hover over 'speaker setup'. It should show 5.1, 7.1, (whatever you selected running the audio setup). Turn off AVR. It should show 'stereo'. Toggling the AVR on/off should automatically toggle Windows audio properties. This doesn't occur with newer drivers though. It sticks on whatever you last selected and isn't automated. You have to manually change it. This affects flash video in Windows and programs like Kodi player for certain audio like AAC.

After testing new drivers to see if the bug remains, I always unpack new driver packages, delete the included HD Audio driver and replace it with 388.59 HD Audio driver which brings back normalcy. Don't be fooled thinking Windows audio properties have nothing to do with passing audio duties to AVR via HDMI. This only affects folks who use systems with AVR in both on and off states and not just one or the other. Just sayin'...

[el Filou]

I have filed a feature suggestion to MS Feedback Hub for adding a 'audio-only display output' feature to Windows (could be useful for HTPCs). Here's the link if anyone is interested: https://aka.ms/AA5emb3Well guess what? Looks like MS *did* implement it: "specialized displays" in Windows 20H1 https://www.tenforums.com/tutorials/154973-how-remove-display-desktop-windows-10-a.html
From the description it sounds very much like what I asked i.e. displays that can be connected and used by specific apps but don't interact with the desktop. However as this is MS we're talking about they decided to restrict the feature to Windows 10 for Workstations... :( so I can't even test if it would work for HDMI audio to an AVR then. Shame.

[VBB]

@VBB: What was the first version affected?

Greetings from Bavaria ;)

Let's go with the first 445 release, 445.75. Fairly certain that's when it started.

[huhn]

But in the process, I found out something I did not know: Nvidia drivers dither by default. See my post over at AVS https://www.avsforum.com/forum/40-oled-technology-flat-panels-general/3138274-2020-lg-cxa-gx-dedicated-gaming-thread-consoles-pc-16.html#post59718362

it's well known. AMD does it too.
it's so well known that they both dither that madshi added a workaround for a bug in the nvidia driver where it would not dither in the past. i'm pretty sure this workaround is still in madVR. in this case madVR doesn't allow sending 10 bit.

[VBB]

I had a feeling huhn would show up eventually ;) Anyway, it was news to me that Nvidia dithered by default. I knew about AMD.

So, this is probably a dumb question, but here goes: Do we want the driver to dither in addition to madVR? If so, why? What happens if only madVR dithers?

[huhn]

that's a confusing question.
you can't dithering 8 bit to 8 bit or 10 bit to 10 bit.
so dithering is applied when bit deep is "lowered" if the GPU has to do that for what ever reason.
yes we do want it.

[VBB]

I see you're up late. OK, so what happens if we disable dithering at the driver level? I noticed it removed some of the smoothing from my gradient tests, but it didn't introduce any new banding. Playback of normal content was noticeably sharper.

[huhn]

i tried the tool at 6 bit without dithering on the desktop it's "clearly" not applied on everything. so i don't know where it put's it self. does it work similar to reshade and only effact d3d programs?

nvidia was now tested to be bit perfect by madshi for envy development using capture cards.
or in short nothing should happen if no dithering is needs and banding should be the result if it is needed.

[VBB]

Right, so you would think, but in an all 10-bit chain, there is a very noticeable difference between Nvidia dithering on or off. This applies to playing back both 8- and 10-bit content. That's why I'm asking, because I was always under the impression that the video card doesn't dither if no dithering is needed. I noticed the difference in both RGB and YCbCr output, btw.

[huhn]

if you output YCbCr you need dithering.
if it is full range RGB you may not need it else you pretty much always need it.

and source bit deep is irrelevant.

[chros]

VBB uses ycbcr 4:2:2 10 bit output.
madvr 10 bit RGB -> ycbcr 4:2:2 10 bit : dithering is needed after every conversion step (in this case at nvidia driver)

@VBB: how did you test nvidia dithering? sample file?

[Klaus1189]

@VBB: added info to 445.xx ;)

@brazen1: added in Notes to Nvidia section, because too many versions affected ;)

I had cut shorten a bit, the post is to long, in the future I will have to split AMD and Nvidia into the first two posts, we'll see :)

[mclingo]

20.5.1 is looking good for me, i can now get FULL RGB with 1080p with HDR, for some reason everythign after 20.3.1 gave me raised blacks, this is the first fully working driver i've had since I bought my RX5700 !!! yippeee....

not tested anything else yet though ;)

[VBB]

Thanks, Klaus!

@chros - Yeah, after my little discussion here with huhn and also reading the first few introductory posts of the madVR thread again, I now understand better how it all works in theory. Having said that, without any driver dithering, what exactly should I be seeing? Increased banding due to reduced noise?

I tested this once again with my usual banding tests from both the AVS as well as the S&M discs.

[huhn]

if the tool is working in that eco system you should get 12 bit banding with 12 bit(should be close to invisible) or 8 bit with 8 bit output as an example.
it should be pretty much the same as madVR with disabled dithering.

with full range RGB nothing should change.

[VBB]

Thanks, huhn. I'll do some tests with full range RGB later. That tool makes it so easy to switch. The only thing missing is refresh rate.

[kostik]

Thanks, huhn. I'll do some tests with full range RGB later. That tool makes it so easy to switch. The only thing missing is refresh rate.

You can combine it with Display Changer (DC not DC2) which will change refresh rate and create one batch file for each mode

[chros]

DC2 is also fine, that's what I use to change resolution and refresh rate.

[VBB]

Some observations on Nvidia dithering after using the tool for a while now:

- The default Nvidia setting for dithering seems to be "auto".
- By default, there is no driver dithering with RGB, as expected, but it can be enabled.
- By default, there is less than 10-bit driver dithering with YUV, even when the output is set to greater than 8-bit.
- Dithering applies to everything, including the desktop. This is most obvious when forcing 6-bit dithering, as the noise then becomes highly visible.

If you enable dithering for RGB, you can get rid of any banding in the LG OLEDs' PC mode, even in HDR. The downside is an increase in noise level, which might not be tolerable for everyone. Certainly not for me, but it's fun to test, and it might be perfectly acceptable for gamers. After all this testing, I am back to my old YUV422, 10-bit, normal HDMI. It's the "Triple Dither™" that makes it look great after all (driver+madVR+display processing).

[huhn]

what did you do to get 6 bit working it's clearly not working here.

[VBB]

That tool I linked to earlier allows you to do that. I used nvcolorcontrol.exe 8 (this is the driver bit-depth) RGB (color space) 1 (this enables dithering) 0 (this is 6-bit) 4 (for temporal, but you can use whatever you like) 0 (for SDR), so nvcolorcontrol 8 RGB 1 0 4 0

[huhn]

ok it doesn't work with dithering disabled or mode 0.

temporal is horrendously bad.

[VBB]

Well, yeah, disabled turns dithering off altogether. Mode 0 is auto, which you can't be sure what it does exactly. 6-bit just proves that you can make even the worst banding smooth, and it'll make all the old plasma guys feel nostalgic ;)

What do you consider the best type of dithering, since you dislike temporal? With 10-bit and even 8-bit, I can't see the noise, and I sit less than 5 feet from my 65".

[huhn]

didn't check the rest in detail if they are not random dithering they should be pretty transparent i need to use dithering test images for this but it's hard for me to care if i can just avoid it.

i don't even know why they would add temporal if you can't use it for what it is made for.
i mean nvidia get's 60 hz in and outputs 60 hz so temporal can not be used as it should be...

maybe it's made for half frame rate vsync games.

[SamuriHL]

Ya know....I really wish Windows would stop F***ING with my drivers! I was content and happy on 442.74. Apparently Windows decided to upgrade my drivers for me. Holy God how can any of you stand the latest nVidia driver?!?! Seriously it broke everything for me. HDR was a complete mess while triggering HDR it was not outputting in the correct color space (even though it claimed to be bt.2020). Even SDR was a complete mess. No one else has seen this? I went back to 442.74 and I'm back to video bliss.

[huhn]

sorry didn't update in a long time and could be that the windows update driver is broken so everyone that upgrades normally doesn't notice.

[SamuriHL]

Yea I hadn't updated in a while either for the obvious reasons. I suppose it's entirely possible that WU broke the driver install. It was bad though. Freaked me right out. LOL Downgraded back to my normal driver and all is well so no big deal I guess. Just apparently no one else has run into it.

[VBB]

I haven't watched anything HDR in a while, but as you know, I've been using the latest drivers. No issues here other than the now well known one mentioned in the first post. What did you notice wrong with SDR? How did you figure HDR wasn't being output in the correct color space?

[RXP]

Does 442.74 passthrough meta data oK? I don't notice that mentioned on the first page anymore.

[VBB]

Yes, meta data has been passed through properly for a long time now.

[SamuriHL]

Yes 442.74 is the driver I'm sticking with until they fix the stupid issues introduced in the later drivers. It works perfectly for me.

[webbo]

Yes 442.74 is the driver I'm sticking with until they fix the stupid issues introduced in the later drivers. It works perfectly for me.

Hello SamuriHL

But which would be those issues - for example, tested 446.14 myself for a few movies / TV Shows, 4K HDR ones, no issues whatsoever

Bugs in colorspace, you mean?

Thanks for your input

[SamuriHL]

The issues that have been reported about if you're outputting 10 bit or higher, it'll pull you out of HDR if you bring up an overlay (like your player). It's just distracting. Not that I tend to bring up the overlay very often when I'm watching a movie, but, I have no reason to upgrade that machine unless I have a driver with no compromises to go to. For others, especially those using 8 bit output, it's a non-issue.

[Calvi]

I am running 442.92 Studio drivers and it seems as good as 442.74

[SamuriHL]

Maybe I'll give those a try.

Sent from my SM-G975U using Tapatalk

[nsnhd]

The issues that have been reported about if you're outputting 10 bit or higher, it'll pull you out of HDR if you bring up an overlay (like your player).
So with the 442.74, outputting 10-bit FSW stays when bringing up overlay ? I'm using the very old 385.28 for sticking to 10-bit in all modes but when overlay (seek bar, madVR controls...)comes up it changes to 8-bit shortly.

[SamuriHL]

Yup works just fine.

Sent from my SM-G975U using Tapatalk

[VBB]

Updated Windows 10 to v2004. Kept all drivers and software settings intact, but reset display resolution, gamut, and bit depth to Nvidia defaults. Haven't run into any issues, other than the ones associated with the latest Nvidia drivers (see 1st post in this thread).

[SamuriHL]

I've been running 2004 for a few weeks now and I'm using it with the latest studio driver with no issues.

[littleD]

https://www.nvidia.com/en-us/geforce/news/directx-12-ultimate-game-ready-driver/

Nvidia GeForce 451.48 WHQL WDDM 2.7 driver, new driver branch so will need testing.Finally new features:

Supports DirectX 12 Ultimate
Supports Vulkan 1.2
Supports hardware-accelerated GPU Scheduling.
Supports CUDA 11.0.

[SamuriHL]

Ah damn! I'll see if I can test those out tonight if no one gets to it first. Those look fun!

[littleD]

Ah damn! I'll see if I can test those out tonight if no one gets to it first. Those look fun!Be careful i ve lost "add run with graphic processor" to context menu. I still look for fix.

Microsoft's shit again https://nvidia.custhelp.com/app/answers/detail/a_id/5035

[kostik]

Nvidia GeForce 451.48 WHQL again the same issues I've seen with the last drivers... wrong colorspace when launching HDR with madVR ... to fix this, after launching the movie first time, in the same mpc-hc window , I click open file and pick the same movie file and it switches to NV HDR and to the correct colorspace.

[VBB]

@kostik Could you give us a little more detail? What color space does it launch in? What does it say instead of NV HDR when it first does this? Are you using 10-bit?

[kostik]

@kostik Could you give us a little more detail? What color space does it launch in? What does it say instead of NV HDR when it first does this? Are you using 10-bit?

I am using 12bit RGB 24Hz. . . If you disable "send hdr metadeta to the display" in madVR settings and turn on HDR in windows display settings , play an HDR movie, you will see the exact phenomenon I get launching hdr movie first time with this driver (and some of other 45X drivers) .
It uses OS HDR. While the movie plays, if you open it again (file->open) , it switches to the right mode.

WEIRD ... 446.14 works fine, also the studio drivers work good. Maybe I will try the Studio version of this 451.48 driver but I guess its an issue with this branch and will have to wait for a fix

[Klaus1189]

Maybe I will try the Studio version of this 451.48 driver

Is there difference between Studio Driver and Game Ready Driver, when the number is identical? I thought it's the same, but I was already mind blown by minor things ;)

[kostik]

Is there difference between Studio Driver and Game Ready Driver, when the number is identical? I thought it's the same, but I was already mind blown by minor things ;)

I also think its basically the same, maybe they just add some profiles for pro editing software and that's it. Well, will try anyway . I think I will stay with these drivers... I don't mind opening movie twice, it takes a second and works fine afterwards.

[chros]

How is the new GPU scheduler? Is there any performance impact compared to the previous drivers with the same settings? :)

[nevcairiel]

Nvidia GeForce 451.48 WHQL again the same issues I've seen with the last drivers... wrong colorspace when launching HDR with madVR ... to fix this, after launching the movie first time, in the same mpc-hc window , I click open file and pick the same movie file and it switches to NV HDR and to the correct colorspace.

To be honest madVR is likely just a bit out of date when it comes to HDR handling. Multiple years of driver and OS HDR development - an area that was never fully fleshed out before, and no madVR updates to speak of.

[kostik]

To be honest madVR is likely just a bit out of date when it comes to HDR handling. Multiple years of driver and OS HDR development - an area that was never fully fleshed out before, and no madVR updates to speak of.

Would you suggest using OS HDR?

[clsid]

to fix this, after launching the movie first time, in the same mpc-hc window , I click open file and pick the same movie file and it switches to NV HDR and to the correct colorspace.Tip: you can press Ctrl+E to quickly reload the current file in MPC-HC.

[VBB]

I am using 12bit RGB 24Hz. . . If you disable "send hdr metadeta to the display" in madVR settings and turn on HDR in windows display settings , play an HDR movie, you will see the exact phenomenon I get launching hdr movie first time with this driver (and some of other 45X drivers) .
It uses OS HDR. While the movie plays, if you open it again (file->open) , it switches to the right mode.

WEIRD ... 446.14 works fine, also the studio drivers work good. Maybe I will try the Studio version of this 451.48 driver but I guess its an issue with this branch and will have to wait for a fix

I guess your reply is confusing me a little. If you disable "Send HDR metadata" in madVR, and then manually flip the HDR switch in Windows, doesn't madVR normally say OS HDR in that case? Wouldn't it only say NV HDR if madVR was actually sending the metadata?

[SamuriHL]

Studio driver is no different. Remember a few weeks ago when I was complaining about the Windows supplied driver and asking why no one saw the issue I was seeing? Yea, the current drivers are exactly what I was seeing but I didn't know at the time that you could just restart the video and it'd kick it into the proper HDR mode. That works for me for now. I guess. Not seeing any adverse effects from turning the new option on in Graphics Settings for windows. The reduce latency and improve performance option.

[kostik]

I guess your reply is confusing me a little. If you disable "Send HDR metadata" in madVR, and then manually flip the HDR switch in Windows, doesn't madVR normally say OS HDR in that case? Wouldn't it only say NV HDR if madVR was actually sending the metadata?

It says OS HDR when launching the file for the first time, when reloading the movie again, it switches to NV HDR. My comment was only to describe how it looks when playing HDR movie and not my real setup . If you do what I wrote , you will see how the movie looks with the new driver on a normal setup

[VBB]

So, here's my experience with this latest driver:

First off, only a DCH version is available at this time. Perhaps that's how it's gonna be from now on. That means the control panel (NVCP) component of the driver will be installed (and updated) from the Microsoft store. I am not a fan of that, but oh well. Did a clean install, reboot, then configured as usual. The only change I made, other than color space and bit depth, is the switch from "Optimal" to "Adaptive" power. Played an HDR test clip and noticed that on first try, it says NV HDR in the OSD, and everything looks washed out. On subsequent tries, OSD shows OS HDR, and everything looks as it should. That alone was enough for me to not use this driver, but it turns out that, at least with my 960, performance was also almost cut in half. madVR seemed to add around 20ms on average, regardless of the decoder in use. Tried FS exclusive, thinking that might make a difference, but nope. That caused such bad flickering, it was unusable.

After wasting almost two hours, I did a clean install of the previous driver. All good.

[VBB]

That is good to know. For some reason those do not come up in my results, like they normally do.

[VBB]

Anyone else try this latest Nvidia driver? I'd love to know why my card's performance was affected like that. Made it entirely unusable for me.

[SamuriHL]

I have it installed and do not see any performance issues. I did enable the new windows option in the graphics settings. I also changed power to adaptive as we all do.

[nsnhd]

The reason why you update to the latest driver is that you're playing games ? I don't see other ones because I believe that newer versions couldn't improve madVR performance, is that right ?

[VBB]

@SamuriHL That new option seems to be for newer cards only. It did not show up for me with 451.48/W10 v2004. Did you install the standard or DCH version?

[VBB]

@nsnhd Being a seasoned gamer/admin/geek, I tend to always use the latest version of everything, knowing the risks, of course. I don't recall the last time a new driver actually improved madVR performance, but I like to test things. This latest Nvidia package also included a newer audio driver, for example. Unfortunately, Nvidia has been catering almost exclusively to gamers for as long as I can think back (Bought my first RIVA in the late 90s).

[SamuriHL]

@SamuriHL That new option seems to be for newer cards only. It did not show up for me with 451.48/W10 v2004. Did you install the standard or DCH version?Dch. Of the studio driver not that it makes a difference in this case. And yeah I've got my 2070 founders edition in that machine. I enabled it on my 2080 on my laptop as well using the gaming driver but I've not tried madvr yet. The laptop tone maps to sdr though.

Sent from my SM-G975U using Tapatalk

[el Filou]

@SamuriHL That new option seems to be for newer cards only. It did not show up for me with 451.48/W10 v2004.Yes, HAGS is for Pascal+.

[SamuriHL]

I had to go back to a working driver. It just simply stopped outputting HDR for me. My HD Fury was showing no HDR metadata, but, the C8 went into HDR mode. The usual trick of stopping it and reloading the file no longer worked for me. This is a mess. Went back to the previous studio driver for now.

[ShadowVlican]

Can confirm that Nvidia 451.48 breaks MadVR HDR passthrough... Many thanks to this thread! I thought I had screwed up something then I googled the problem and realized it was Nvidia's fault!

Downgrading back to an older driver because I had passthrough working just fine before...

[VBB]

Gonna wait for someone else to test this time around ;)

[kostik]

Gonna wait for someone else to test this time around ;)

Same as previous 451 drivers

[tyguy]

Is there a bug in latest drivers regarding hdr playback? NV HDR is all grey and washed out, but my tv says HDR is on. Using OS HDR looks proper. I’m using madvr tonemapping with pixel shaders for what it’s worth.

[VBB]

Yes, the last two do not work properly with madVR's HDR implementation. This is due to changes in the way Nvidia handles HDR, so not really a bug in the driver, but something that needs to be fixed in madVR.

[tyguy]

Yes, the last two do not work properly with madVR's HDR implementation. This is due to changes in the way Nvidia handles HDR, so not really a bug in the driver, but something that needs to be fixed in madVR.
Okay thanks for letting me know. Im assuming it will be awhile before this is fixed in madvr since development seems to be on hold. Having to toggle HDR isn't the end of the world.

[huhn]

a new test build is expected every day now. there is a good chance it will be complain with the API changes.

[VBB]

As much as I'd like to see one, I wouldn't get my hopes up. Mathias is simply too busy with finalizing Envy. Once it ships, any further madVR development will also benefit Envy, so that should be good for us here. What I'd really like to see in addition to bug fixes is support for Dolby Vision, especially now that single layer rips without compromising audio are possible.

[huhn]

i have bad news for you.
this may never ever happen without papers.
you need a decoder first to even output it to madVR.

[SamuriHL]

There are decoders on the PC if you can figure out how to use them. And Dolby Vision has been added to the MKV standard in the past couple weeks. Things have been progressing quite nicely outside of the PC world for DV.

[VBB]

Well, hopefully Nev will find it worthy of his time, but I know we've discussed this before. The problem with DV is that it's closed pretty tightly. It will happen eventually.

[Klaus1189]

@brazen1: I think I read somewhere that there is a new Audio driver version in recent Nvidia drivers, does that fix this issue?
https://forum.doom9.org/showpost.php?p=1913991&postcount=1319

[SamuriHL]

Well, hopefully Nev will find it worthy of his time, but I know we've discussed this before. The problem with DV is that it's closed pretty tightly. It will happen eventually.Some of the docs have been released. Still can't properly encode but the reason so much progress was made on mkv is because they got access to the profile docs and had the mp4 muxer Dolby released. They were able to piece everything together from that. All of my dv movies are sitting on my nas in ts format as profile 04.06 for playback on my shield. Those Intel drivers that were released in May have a decoder. And Nvidia drivers have had a beta decoder for a while... I'm guessing for games. So it's not impossible anymore. Just time consuming and difficult.

Sent from my SM-G975U using Tapatalk

[fairchild]

Any testing done on 20.7.1 drivers, how can i help to test compatibility? I'm running them currently on my RX 56 Vega.

[Klaus1189]

...

I am glad you want to help, but sadly I don't understand what you mean with testing compatibility?

[fairchild]

I am glad you want to help, but sadly I don't understand what you mean with testing compatibility?

Well since that driver hasn't been verified, I was willing to run whatever test you do to check if the passthrough for HDR is working. I have an HDR MKV file I use.

LG Colors of Journey HDR UHD 4K Demo.mkv

Does show HDR passthrough working as my TV switches to HDR mode and appropriate Mode on my TCL Roku 4k TV with latest MadVR and 20.7.1 drivers

[Klaus1189]

Something like this would be really nice, since I don't have equipment or knowledge to do tests like this:
https://forum.doom9.org/showthread.php?p=1865444#post1865444

I use my RX 5700 XT to check if the passthrough works, but please keep on checking with your Vega (other generation Navi) if the HDR switch still works and if there is anything please report here.

[solidservo]

so what is the recommended nvidia driver for madVR with HDR content?
442.74?

[Klaus1189]

For now with madVR I recommend one of the following:
Last working GRD: 446.14
Last working SD: 442.92
Please keep in mind the issue introduced in 445.xx drivers:
This version and all 445.XX and higher: Playback at greater than 8-bit drops back to 8-bit (and out of HDR) in FS windowed mode when player GUI is triggered, regardless of bit-depth of the file. Can be circumvented by using FSE mode or by setting madVR to 8-bit.

[solidservo]

For now with madVR I recommend one of the following:
Last working GRD: 446.14
Last working SD: 442.92
Please keep in mind the issue introduced in 445.xx drivers:
This version and all 445.XX and higher: Playback at greater than 8-bit drops back to 8-bit (and out of HDR) in FS windowed mode when player GUI is triggered, regardless of bit-depth of the file. Can be circumvented by using FSE mode or by setting madVR to 8-bit.

shouldn't I just used GRD 442.74 because of this? Or is there a good reason to go with 446.14 instead?

[Klaus1189]

I would go with SD 442.92 over GRD 442.74, but if you really want to go with GRD only and need/want prevent the 445.xx issue mentioned in my former post yes. Because there are workarounds mentioned but with drawbacks, like no player gui in exclusive mode.

[Filmgeek47]

I’ve got a frustrating issue that I’m hoping someone can help me with. I’m running a GTX 1660ti and I’m trying to get working 3D and 4k. I’m really close with the 425.31 driver. The only issue I’ve got is that every time I reboot, my bit depth settings reset back to the default of 8 bit, which on my projector makes everything go magenta. If I set the bit depth to 12 bit at 24hz, it’ll switch back and forth correctly between 24/60hz and 8/12bit, but on reboot I have to reset those bit depths again. Is there a solution to this? Was this a common issue with earlier drivers? The bit depth is persistent on the newer drivers, but then 3d doesn’t work.

[onekmilesbehind]

@Filmgeek47 I have run into that issue on numerous driver builds with my 2070. Bit-depth settings seem to break, get fixed, then break again at various times. I don't have any specific record of what worked when on my end, but do know that the current 451.27 build preserves bit-depth settings across reboots.

Hopefully somebody running a build in that 420 series range can chime in if bit-depth is preserved across reboots. It's one of those pesky low-priority bugs that the HTPC crowd seems to have to deal with.

[Klaus1189]

Need info of new Nvidia SD driver of today regarding HDR passthough ... Thanks :)

@Filmgeek47 and @onekmilesbehind: If you found a driver that gets working 3D and 4K and doesn't loose bitdepth after restart, please let me know, I can pin that info. I personally do not use CRU, but isn't that a possible way to get/hack 12 bit output which survives restart?

Just a thought, but might a AMD card a choice for 3D and 4K?

[kostik]

NVIDIA Studio Driver Version 451.77 WHQL - same HDR behavior as other 451.X . Hope madshi's beta driver will fix this but I grew to like OS HDR but I am not sure what value I should use for Brightness sliders in windows HDR settings for my 65CX oled.

[Filmgeek47]

Need info of new Nvidia SD driver of today regarding HDR passthough ... Thanks :)

@Filmgeek47 and @onekmilesbehind: If you found a driver that gets working 3D and 4K and doesn't loose bitdepth after restart, please let me know, I can pin that info. I personally do not use CRU, but isn't that a possible way to get/hack 12 bit output which survives restart?

Just a thought, but might a AMD card a choice for 3D and 4K?

I was hoping for a solution relating to (or similar to) CRU, but so far I haven't been able to see a method for specifying bit depth, only display timings/resolution.

It sounds like my best bet is to test each of the drivers post 1660Ti launch and find one that does 3D and doesn't lose bit-depth. I'll post back if I'm successful.

@onekmilesbehind Thanks for the heads up.

I already have the Nvidia card so I'm stuck with it, but in any case I opted against AMD as I need the option to flag output as Bt2020 (it tells my HDfury device that I'm outputting 4k/HDR). I don't believe that check box works with AMD.

[Filmgeek47]

So I tested 425.31, 419.67, 419.35, and 419.17. All of those drivers fail to keep bit-depth settings on reboot. According to Toasty over at CRU, that program can't specify bit-depth. Sounds like this is an issue we're stuck with, unless you happen to have an older Nvidia GPU that can run even earlier drivers, where the issue may not occur?

The only other idea I had was if someone who knows how to script might be able to write something that manually sets the bit depth info in the driver, which could be assigned to run on startup. Sadly I'm not really code savvy, but if anyone following this is, this would be a really useful workaround if it's even possible.

[VBB]

I know you're also on AVS, so here's a link to a tool that allows you to do what you're looking for:

https://www.avsforum.com/forum/40-oled-technology-flat-panels-general/3138274-2020-lg-cxa-gx-dedicated-gaming-thread-consoles-pc-14.html#post59699820

You could create a batch file with the settings you need and put it in the startup folder.

[mzso]

Hello!

How do I get full range videos like this to appear normally in the browser (Firefox) with an AMD GPU? It appears washed out...
With my old Nvidia GPU I had a setting that I could set to full range (0-255) and the videos appeared correctly after.

[Klaus1189]

Sample?

[huhn]

if the browser is using d3d11 or DXVA processing then you are doomed level (and lot's of other stuff) is broken for over a year now.

[Klaus1189]

@huhn: Could you make a full list of the broken things? To have a todo list for AMD and have a known issues list for users. I pin it at first post.

[huhn]

i stopped following them and the decoding bug got a workaround.

but here are some small one i just tested:
deint is generally bad not as bad as at the beginning(i guess) it's still very low quality it sometimes just stops working too.

wrong levels AMD is outputting limited range doesn't matter for madVR madVR ask for limited.

d3d11 processing is utterly broken 10 bit or more input/processing in mpcVR it produces something "special" with soem upscal magic (but levels are correct with 10 bit processing).

in general it's just GPU video processing. in the past the image was totally crushed with d3d11 processing and 10 bit.

there is no over write for video settings anymore i never needed it with nvidia and browser are not forced to use DXVA or d3d11 processing but i didn't really checked them.
maybe you can still find video overwrites in the registry custom currently has "brightness".

[DMU]

deint is generally bad not as bad as at the beginning(i guess) it's still very low quality it sometimes just stops working too.

Do you mean DXVA deinterlacing?
Take a look at the face and neck (https://drive.google.com/file/d/1aOdmvI9-DUtptfWDInAjzpD6AGT1tJ1U/view?usp=sharing).

[huhn]

trust me it was worse. this is a massive improvement: https://abload.de/img/deint3hjhm.png
and it's kinda slow i'm pretty sure you can ignore the DXVA rendertimes completely but still.

[DMU]

Fixing the issue using software deinterlacing in LAV?

[huhn]

the problem is the deint from amd, intel and nvidia is on a different level then yadif well it was on a different level on amd.

[mzso]

Sample?

https://keepvid.works/?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3Dg9ak928cKfE

[mzso]

the problem is the deint from amd, intel and nvidia is on a different level then yadif well it was on a different level on amd.

For me madVR deinterlancing is not engaged, even though the OSD shows enabled when I use SmartDVB. Any way I can fix this?

[Klaus1189]

For me madVR deinterlancing is not engaged, even though the OSD shows enabled when I use SmartDVB. Any way I can fix this?
Do a test: Don't use deinterlacing in madVR, instead use DXVA copyback in LAV Filters and select Yadif (50/60p of course) then deinterlacing is an a very good level, isn't it?

[mzso]

Do a test: Don't use deinterlacing in madVR, instead use DXVA copyback in LAV Filters and select Yadif (50/60p of course) then deinterlacing is an a very good level, isn't it?

LAV wouldn't even let me to enable HW deinterlacing with either of the HW decoder options...

[Klaus1189]

Not HW deinterlacing, we use copy back, so we can use software deinterlacing, see here what I am talking about:

https://drive.google.com/file/d/1jFZR2UIyMlyuTRRBzywUVGwYNR00eOe0/view?usp=sharing

[mzso]

Not HW deinterlacing, we use copy back, so we can use software deinterlacing, see here what I am talking about:

https://drive.google.com/file/d/1jFZR2UIyMlyuTRRBzywUVGwYNR00eOe0/view?usp=sharing

I don't want SW deinterlacing, that's not what my question was about.

[Klaus1189]

Why not? YADIF is far superior to any hw deinterlacer and you can still use hw acceleration for videodecoding.

[mzso]

Why not? YADIF is far superior to any hw deinterlacer and you can still use hw acceleration for videodecoding.

Not my experience. (https://drive.google.com/file/d/1JNTd0AwlY4xEMRSQS124vAtMbfazOjWe/view)

I also don't want HW decoding. It's just slow and troublesome.

[Klaus1189]

I don't want SW deinterlacing.
You don't want software deinterlacing.
-> Ok, I got that.

I also don't want HW decoding.
You don't want HW accelerated video decoding.
-> Ok, I got that, too.
-> So you want software video decoding instead.

Wait... You don't want software deinterlacing when using Software decoding?
Do you want to watch interlaced video without a deinterlacer, or am I missing something?

:confused::confused::confused:

[Klaus1189]

https://drive.google.com/file/d/1JNTd0AwlY4xEMRSQS124vAtMbfazOjWe/view

Can I have the original file please? I know it will have some gigs.

[mzso]

You don't want software deinterlacing.
-> Ok, I got that.


You don't want HW accelerated video decoding.
-> Ok, I got that, too.
-> So you want software video decoding instead.

Wait... You don't want software deinterlacing when using Software decoding?
Do you want to watch interlaced video without a deinterlacer, or am I missing something?

:confused::confused::confused:

"or am I missing something"

Apparently. I don't want to use HW decoding because its inferior. As well as SW yadif, it seems. So HW deinterlacing and SW decoding.

[mzso]

Can I have the original file please? I know it will have some gigs.

Not much actually:

https://drive.google.com/file/d/12wO8NJbIrfa2MD4yVaA38RrBBdziCda8/view?usp=sharing

[Klaus1189]

please cut off "https://drive.google.com/file/d" from the link

[Klaus1189]

This file is MBAFF according to Medianfo. I did also some tests with my NLE. MBAFF means that some parts of the image is partly progressive and partly interlaced. In this case the moving parts are interlaced to preserve the motion of 50 "fps" but the deinterlacer deinterlaces also the progressive parts like the info bar which results in flickering.

Can I post this sample somewhere else for help of guys I know?

I played it with one of my VU+ses and it plays it perfectly.

[huhn]

yadif is not frame adaptive it's "inferior" to intel and nvidia deint and AMD was frame adaptive at one point too but up to this day the AMD driver claims to be frame adaptive while doing a simple bob or no deint at all

of cause you will find files where yadif will win but generally a frame adaptive deint algo is superior.

MBAFF is just an encoding trick you can ignore it it doesn't tell you if a file is interlaced PsF hybrid or what ever.

and well you can't use lav filter hardware deint with AMD. quicksnyc is intel only and cuvid is nvidia only. you don't need them anyway because the video renderer can access the hardware deint.

no clue why a hardware decoder should be "slow" that happens from time to time but generally they are in there only league but they have nothing todo with hardware deint so what ever.

[Klaus1189]

Hmm ... psf = progressive segmented frames. Isn’t that something different, since psf is true progressive but stored in fields? In this case it would be real 25p but the file would be encoded as interlaced, but to play it properly you only need weaving which only presents both fields at the same time. For true interlaced content you don‘t want that since the fields should be presented alternating. I can not describe it better. Bear me me ;)

[huhn]

a decoder outputs a interlaced 30i stream as 30"p" images if it is PsF, interlaced or hybrid doesn't matter. it's the job of the deint algo to turn this into 60 fields and then to 60 frames.

or with other words a decoder isn't outputting fields. while this is technically not impossible it's simply not done practically. well there was/is even a "bug" with HEVC interlace while the file was encoded in fields and ffmpeg didn't output the file as full weaved frames.

there is also soft telecine but i'm not going into that.

[Klaus1189]

I found the wiki page and it is decribed exactly like I thought it is.
Although it is german:

Progressive segmented Frame (PsF) ist ein Videoformat, das Vollbildmaterial als Halbbilder speichert.

Jedes Vollbild wird dabei in zwei Halbbilder ohne Phasenverschub der einzelnen Zeilen innerhalb eines Halbbildes geteilt (segmentiert). So lassen sich aus den im Zeilensprungverfahren gesendeten Halbbildern durch Weaving die ursprünglichen Vollbilder vollständig und verlustfrei rekonstruieren, ohne dass eine Interpolation vorgenommen zu werden braucht oder der für das Zeilensprungverfahren typische Kammeffekt auftritt.
Source: https://de.m.wikipedia.org/wiki/Progressive_Segmented_Frame

[mzso]

Can I post this sample somewhere else for help of guys I know?

I played it with one of my VU+ses and it plays it perfectly.

Sure.

What's VU+ses?

[huhn]

that's not the point the deint algo doesn't know that.

and the encoder doesn't have to know that too.

[mzso]

yadif is not frame adaptive it's "inferior" to intel and nvidia deint and AMD was frame adaptive at one point too but up to this day the AMD driver claims to be frame adaptive while doing a simple bob or no deint at all


Up to which point did the AMD drivers had a decent deinterlacer?
So I guess this means that MadVR enables deinterlacing but it does nothing?

and well you can't use lav filter hardware deint with AMD. quicksnyc is intel only and cuvid is nvidia only. you don't need them anyway because the video renderer can access the hardware deint.
So I guess it was disabled at one point?

Any good directshow deinterlacing filters around? Either SW or HW?

[Klaus1189]

Thanks :)

defective plural of "VU+" :D

It is a manufacturer of Enigma2-Receivers .
I've set it to 1080p50/2160p50, so the TV only receives a progressive signal and all deinterlacing is done in VU+ box.

[Klaus1189]

Here is a 25 psF file straight from a Camcorder which I remember to be psf:
http://www.aktiv-verlag.com/~webvideos/testvideos/2010/sony/Sony_HDR_AX2000.MTS

MediaInfo will tell you Interlaced TFF, but it is 25p stored as 50i or 25i, whatever you like to call it.

[huhn]

@mzso
they break it with a driver update and then forget about it.
i know the "last" CCC version has frame adaptive deint. and i know the first new driver "experiences" crimson or what it's name was doesn't have it working. there are some crimson driver where it works too.

they break this stuff all the time and now is been broken for a long long time.

so yes you could get combing and yes you could get wobbling with AMD.

i'm not aware of a frame adaptive deint algo that matches hardware deint but this is considered as a good deint algo: https://forum.doom9.org/showthread.php?t=174544&highlight=qtgmc

[Ejziponken]

So assuming madVR won't get an update anytime soon to fix the latest Nvidia drivers and the new HDR API changes, what's the best option?
Because I play games on the same PC as Im watching movies, and I want to run the latest GRD drivers while still being able to watch 4k HDR movies.

Currently Im using MPC-BE + madVR beta + LAV

[mzso]

@mzso
they break it with a driver update and then forget about it.
i know the "last" CCC version has frame adaptive deint. and i know the first new driver "experiences" crimson or what it's name was doesn't have it working. there are some crimson driver where it works too.

they break this stuff all the time and now is been broken for a long long time.

so yes you could get combing and yes you could get wobbling with AMD.

i'm not aware of a frame adaptive deint algo that matches hardware deint but this is considered as a good deint algo: https://forum.doom9.org/showthread.php?t=174544&highlight=qtgmc

It would be nice is if I could get the latest that had decent deinterlacing (Maybe it would be worth a try.), but it would require a great amount of testing it seems.

[huhn]

i can't i don't have a polaris card i'm using terrible navi.

AMD has a better out of the box sync then nvidia and better control over the output signal then nvidia the "rest" like hardware decoder and processing are buggy slower or simply years behind nvidia.

i can only say avoid. i have a navi card for a year now and it was a not been funny i have seen to much... i mean my 240 hz freesync screen didn't work with it correct (still has issue) but it's flaw less with nvidia. nvidia has better freesync support then AMD.

[pistacho]

So assuming madVR won't get an update anytime soon to fix the latest Nvidia drivers and the new HDR API changes, what's the best option?
Because I play games on the same PC as Im watching movies, and I want to run the latest GRD drivers while still being able to watch 4k HDR movies.


You may be interested in trying Kodi v19 HDR

https://forum.kodi.tv/showthread.php?tid=355800

It works fine with the latest NVIDIA 451.77 studio drivers or GRD 451.xx (with default / clean install settings).

[littleD]

You may be interested in trying Kodi v19 HDR

https://forum.kodi.tv/showthread.php?tid=355800

It works fine with the latest NVIDIA 451.77 studio drivers or GRD 451.xx (with default / clean install settings).
Thanks for the info.
Player plays all demanding stuff on intel, so dxva is quite well optimized.
[PS. Nah, found some particular problems]

However it treats brightness differently to other players. Some samples looks like not so colourful when played in other players. On the other hand HDR videos are not completely pale, i think player tries to preserve most black/white tonal range upon saturated colours.
But im on internal laptop screen so, maybe will check more options available.

[Ejziponken]

You may be interested in trying Kodi v19 HDR

https://forum.kodi.tv/showthread.php?tid=355800

It works fine with the latest NVIDIA 451.77 studio drivers or GRD 451.xx (with default / clean install settings).

I really don't like the whole library thing kodi has going on. Its actually driving me nuts with all that crap built in. You have any other suggestions? I rather just have a stand alone media player like mpc or something..

[Klaus1189]

So assuming madVR won't get an update anytime soon to fix the latest Nvidia drivers and the new HDR API changes, what's the best option?
Because I play games on the same PC as Im watching movies, and I want to run the latest GRD drivers while still being able to watch 4k HDR movies.

Currently Im using MPC-BE + madVR beta + LAV

Set up a dedicated HTPC which uses last recommended driver regarding madVR. Don‘t need to be highend hardware. Just take a look at 3300X. Not expensive and should be enough.

Use a dedicated gaming PC for gaming and always use latest GRD.

[Magik Mark]

You may be interested in trying Kodi v19 HDR



https://forum.kodi.tv/showthread.php?tid=355800



It works fine with the latest NVIDIA 451.77 studio drivers or GRD 451.xx (with default / clean install settings).How's the video quality in kodi 19 compared to dsplayer with madvr?

Sent from my SM-G988B using Tapatalk

[pistacho]

For HDR passthrough it should be the SAME because a player should not alter the video data (in passthrough mode).

[el Filou]

I've seen strange interactions between video decoders and hardware deinterlacing. For example, for a long time after I bought my NVIDIA I thought the deinterlacing was inferior to my old Radeon (HD 4650) because it had really obvious artifacts with 576i, but I discovered by chance that it only happened with DXVA copyback. Using DXVA2 native, the artifacts are gone and deinterlacing is same quality as my previous very old Radeon.
So now I have to switch between DXVA2 native (when I want best deinterlacing quality), DXVA2 copyback (only for black bars detection), and D3D11 native (for best performance).
It's a bit of a pain really. Tahnk god I don't have an HDR TV yet or I would be pulling my hair out.So I guess this means that MadVR enables deinterlacing but it does nothing?At long time ago a SmartDVB user reported a bug with madVR where it didn't enable deinterlacing even though it was enabled in settings, even forced it didn't work (I tested and confirmed). Maybe it was you? Anyway maybe that bug still hasn't been fixed.For HDR passthrough it should be the SAME because a player should not alter the video data (in passthrough mode).It's only the HDR metadata that is passed through, the other parameters affecting picture quality (chroma upscaling, debanding and other image enhancements, internal processing and dithering, etc.) are still dependent on the renderer.

[chros]

So now I have to switch between DXVA2 native (when I want best deinterlacing quality), DXVA2 copyback (only for black bars detection), and D3D11 native (for best performance).
It's a bit of a pain really. Tahnk god I don't have an HDR TV yet

:D You don't need black bar detection with a TV if you can use d3d11 native. And stop watchong old DVDs and problem will be solved :D

[el Filou]

I know, but I like variety of content more than perfect quality ;) and some TV channels are often showing old or rare movies that are nowhere to be found on streaming services or Blu-ray.
I use black bar detection only to crop TV logos/age warnings on 2.39:1 content, not for rendering performance. I wish it would work with D3D11 native.

[chros]

Oh, I see, makes sense now.

[mzso]

At long time ago a SmartDVB user reported a bug with madVR where it didn't enable deinterlacing even though it was enabled in settings, even forced it didn't work (I tested and confirmed). Maybe it was you? Anyway maybe that bug still hasn't been fixed.
Yes, it was me, probably. Anyway, I forgot how, but it was fixed on either SmartDVB or madVR's end.
But now I got a newer GPU and and upgraded to Win10, because it's not supported on win 8.1...

Anyway, one big difference is that back then it worked without a problem in media players like Potplayer, MPC, now it doesn't work anywhere.

[claw]

I’ve got a frustrating issue that I’m hoping someone can help me with. I’m running a GTX 1660ti and I’m trying to get working 3D and 4k. I’m really close with the 425.31 driver. The only issue I’ve got is that every time I reboot, my bit depth settings reset back to the default of 8 bit, which on my projector makes everything go magenta. If I set the bit depth to 12 bit at 24hz, it’ll switch back and forth correctly between 24/60hz and 8/12bit, but on reboot I have to reset those bit depths again. Is there a solution to this? Was this a common issue with earlier drivers? The bit depth is persistent on the newer drivers, but then 3d doesn’t work.

I use driver 436.48 as I have had no reason to update it.

I have a JVC projector and initially set the 2080Ti to output RGB 12-bit due to the Magenta bug.

I only had issues with retaining 12-bit on restart if I set the desktop to 4K59. I changed it to 4K23 and no longer had any issue with 12-bit reverting to 8-bit.

I also use a CRU custom resolution but I doubt that had anything to do with it.

I recently switched to RGB 8-bit and disabled Report BT2020 to display in MadVR. Reporting BT2020 also includes HDR metadata and that is what triggers the Magenta bug.

So I manually change picture mode in my JVC to one that is set up for SDR BT2020 for MadVR output.

I don't do 3D.

[angmav]

Hi Does anyone know how to delay the refresh rate change time for nvidia as I don’t have enough time for me to pick yes from the time it takes for my projector to handshake screen very annoying I run a team view screen and accept change that way there must be a setting I can change to delay timeout change many assistance would be greatly appreciated


Sent from my iPhone using Tapatalk

[huhn]

you can use a different program to change the refreshrate with madVR it's just doen't with no questions asked.

out of memory you could just use arrow key left enter.

[RXP]

Set up a dedicated HTPC which uses last recommended driver regarding madVR. Don‘t need to be highend hardware. Just take a look at 3300X. Not expensive and should be enough.

Use a dedicated gaming PC for gaming and always use latest GRD.

I dual boot my projection setup.

I need to keep the last 3D drivers and everything stable for the projector. So it has. Then my other boot is the latest drivers.

[angmav]

you can use a different program to change the refreshrate with madVR it's just doen't with no questions asked.

out of memory you could just use arrow key left enter.


Thanks what program do you recommend ?


Sent from my iPhone using Tapatalk

[huhn]

well madVR.

just set the device to 60 HZ and when madVR plays a file it will change the refreshrate and when it stops it's back to 60 hz.

if you need thsi for something not related to movie playback there are other tools but i never used one of these.

[Manni]

Thanks what program do you recommend ?


Sent from my iPhone using Tapatalk

If you have AnyDVD, you can also use setdisplayfrequency.exe from a batch file to change the refresh rate if you want to do it for non-video playback related reasons, or want to force a refresh rate manually.

For example, to change the refresh rate to 23p, here is the line to add in a batch file:

"c:\program files (x86)\Redfox\AnyDVD\setdisplayfrequency.exe" 23

There is a registry setting to extend the delay for the prompt to confirm the display mode change in Windows, but I can't remember what it is. I've extended it because my JVC takes a long time to sync, and I don't have this issue anymore, but can't recall how I did it.

[angmav]

If you have AnyDVD, you can also use setdisplayfrequency.exe from a batch file to change the refresh rate if you want to do it for non-video playback related reasons, or want to force a refresh rate manually.

For example, to change the refresh rate to 23p, here is the line to add in a batch file:

"c:\program files (x86)\Redfox\AnyDVD\setdisplayfrequency.exe" 23

There is a registry setting to extend the delay for the prompt to confirm the display mode change in Windows, but I can't remember what it is. I've extended it because my JVC takes a long time to sync, and I don't have this issue anymore, but can't recall how I did it.


Thank you for all your assistance


Sent from my iPhone using Tapatalk

[biship]

To be honest madVR is likely just a bit out of date when it comes to HDR handling. Multiple years of driver and OS HDR development - an area that was never fully fleshed out before, and no madVR updates to speak of.

Yes, it's very sad that he has another project that consumes all his time. I got to think a few days of Visual Studio and he'll fix a couple of the breaking bugs.

[biship]

HDR passthrough with MadVR is still broken in 451.77

[VBB]

And it will keep being broken with every new Nvidia driver, because Nvidia chose to change their API. It works just fine with drivers before 45x.xx

[SamuriHL]

I wouldn't hold out hope for a fix in madvr any time soon. madshi is focused on other things. At MOST we can expect to see some improvements to HDR tone mapping at some point this year. If we're very lucky. He didn't seem interested in this issue at all when it was brought on int he HDR thread. Even though it does impact HDR testing.

[*MidnightWatcher*]

For those you doing HDR to SDR tone mapping @ 8bit with madVR -- v. 451.67 is providing a pretty good 23.976 performance. At 4K, I'm now only seeing at most one dropped frame per movie. This is much better than the dropped/repeated frame every 4 or 12 minutes from the the previous driver I was using.

[Manni]

For those you doing HDR to SDR tone mapping @ 8bit with madVR -- v. 451.67 is providing a pretty good 23.976 performance. At 4K, I'm now only seeing at most one dropped frame per movie. This is much better than the dropped/repeated frame every 4 or 12 minutes from the the previous driver I was using.

Yes but HDR passthrough is still broken. :(

You can get no dropped frames with CRU and HDR passthrough with an older driver. I use 442.92 to get 12bits support as well (I need it occasionally for testing).

[*MidnightWatcher*]

Yes but HDR passthrough is still broken. :(

You can get no dropped frames with CRU and HDR passthrough with an older driver. I use 442.92 to get 12bits support as well (I need it occasionally for testing).
I was using that one as well and decided to stop using 12bit.

[Manni]

I was using that one as well and decided to stop using 12bit.

Sure, if you don't need HDR passthrough either the latest are fine.

[SamuriHL]

Manni, I don't know if you saw my post about it but these latest drivers aren't passing metadata through even with HDR turned on in Windows. This seems completely broken to me. I currently have the driver installed but will be going back to the latest working studio driver. It's completely frustrating as we're so close to having an almost perfect driver.

[Manni]

Manni, I don't know if you saw my post about it but these latest drivers aren't passing metadata through even with HDR turned on in Windows. This seems completely broken to me. I currently have the driver installed but will be going back to the latest working studio driver. It's completely frustrating as we're so close to having an almost perfect driver.

I have but I wasn't saying anything different in my previous post(s). :)

The latest drivers are broken for HDR passthrough and 12bits, but if you don't need either of these features, I guess nothing stops the OP from using the latest.

I personally prefer to use the older studio drivers that handle both 12bits and HDR passthrough, as I get perfect refresh rate with CRU with these.

Maybe you misread my post, or it wasn't clear?

[SamuriHL]

It wasn't fully clear what "broken" meant. :) Mainly because you can force the driver to display in HDR, but, without it passing the metadata (which people without an HD Fury may not know about), yes, it's still "broken". :D I'll be going back to the previous studio driver, as well. I keep hoping they get one of these newer drivers to work with ALL the fun features working. Yes, I could mess with CRU to get better timing, but, who's got time for that?! LMAO

(P.S. I realize it's not clear why I keep trying when I have a working driver. I'm concerned that when I get a 3080TI I'll be forced into a newer driver that supports it...that's my real concern here.)

[Asmodian]

I keep hoping they get one of these newer drivers to work with ALL the fun features working.

Are you thinking Nvidia should change back to the HDR technique madVR currently uses? I think we are waiting for a madVR update to fix HDR, not Nvidia.

[SamuriHL]

You're missing the larger point. It doesn't (currently) matter what madvr does or doesn't do. The latest nVidia driver branch is not sending the HDR metadata at all. So even if you enable it using the Windows HDR setting, it doesn't help because the metadata is missing. IOW, this issue is bigger than hoping madshi changes madvr to enable HDR properly.

[SamuriHL]

Just to do some proof of what I'm saying, I enabled Windows HDR and loaded up an HDR mkv in PowerDVD 20 Ultra. As mentioned, no HDR metadata. Dropped back to 442.92 studio driver, all metadata restored, both in madvr hdr output and in PowerDVD + Windows HDR. nVidia broke it. They need to fix it.

[Asmodian]

When was the last update for PowerDVD?

Edit: Hmm, April 2020 was really recent, I wonder if they don't have the clout to get Nvidia to fix things either. :(

[VBB]

Nvidia didn't break it from their perspective. They simply changed the API, and now software manufacturers have to play catch-up. Kinda messed up to do this without announcing such a change way ahead of time.

[Asmodian]

So we are waiting for the software to update and PowerDVD is simply not up to date either?

That does seem very possible. Hoping for Nvidia to fix it seems like a false hope.

[SamuriHL]

I have my doubts on that theory. Not sending the HDR metadata is a bug we've dealt with before, and it wasn't every program under the sun that had to change. It was nVidia fixing the damn thing. And it took them a LONG time to do it, too. IOW, we've been through this nonsense before with them. So no, I don't buy an "API change that everyone else has to deal with". Enabling HDR in Windows isn't using the nVidia API. And I'm telling you, enabling HDR using the Windows display option does NOT send HDR metadata. What is it that these programs are supposed to do? PowerDVD doesn't use nVidia or AMD's HDR api...they force you to enable it in Windows if you want to use it.

[VBB]

Don't get me wrong. I agree with you that something needs fixing, even if the API change was intentional. But I fear that's something for Mathias to communicate to his Nvidia contacts, once he can figure out what madVR needs to do to comply. I have a feeling Nvidia wants us to use the OS HDR trigger. Probably a push from Microsoft, too.

[SamuriHL]

Which you can do manually. I've tried it and while it did trigger hdr output without the metadata the display is likely aggressively tone mapping it. And in fact if you do the trick where you start the video, stop it, and play again to get it to trigger hdr, you'll see that the windows hdr setting in display properties is enabled.

Sent from my SM-G975U using Tapatalk

[brazen1]

@brazen1: I think I read somewhere that there is a new Audio driver version in recent Nvidia drivers, does that fix this issue?
https://forum.doom9.org/showpost.php?p=1913991&postcount=1319

451.67 another complete waste of time. 12bit still doesn't retain after a reboot. Windows + nVidia audio coordination still broken.

Despite reports of fewer frame drops, nothing has changed for me. Same ol', same ol' drop every 4.5 mins.

NV HDR does activate (as usual) at first attempt despite reports claiming to have to restart the video.

I don't know who these driver 'improvements' are aimed at but it must be a very narrow specific group because all they do for me is make things worse and waste my time.

Back to 445.87 using 388.59 audio driver.

[Klaus1189]

Despite reports of fewer frame drops, nothing has changed for me. Same ol', same ol' drop every 4.5 mins.

yes madshi got a bug report through with this and they fixed it.

the refreshrate is supposed to be closer now with 1080p23 (for a while now) and now with 2160 23 now too with this driver or up coming ones.

looks like they have to fix this with every signle refreshrate and connector type...

just took them like a decade and it doesn't fix the core issue anyway.

edit: i just test my 1060 i have no frame repeat/drop expected ever with analog audio output so very lucky.
can a couple you guys test digital hdmi/DP over the same card that the only space where nvidia has any way to affect the audio clock?
i tested 451.77 studio.

edit2: can you plz check 119? if you have the time for 1440p and 2160p DP/HDMI if available.
i will try to check 1080p119 my TV doesn't support 1440 for some reason.
edit3: 1080p119 is working too.


Can you two please test it again, just to be sure that no CRU settings can create false results. HDR stuff beside. How is the 23 Hz refresh rate out of the box?
I am really interested in that. Thanks in advance ;)

[huhn]

with overlay rendering it'S 23.7617.

just as before not that important the important part comes now. i currently get a repeated frame every 10- unlimted hours room is at 34 degrees currently i'm using analog audio output using a realtek onboard audio card. if i get one of my "good" sounds card back in the system and i have the "newest" soundblaster and a asus essence STX here(and to many more rubbish cards i'm not willing to list here...) i would be shocked if i'm back to something like 20 mins or so.

i never really used CRU i had a panasonic TV there was no point in using it... my TV is still new so every custom refreshrate would be deleted anyway.

i will later try HDMI output with my sony it may have a properly working analog output unlike my panasonic which again is by design not working "properly" and everything can beat a PC DAC too so possible win win.

edit: HDMI digital is 1 h for me absolutely useless

[brazen1]

I wasn't and never have used CRU. I also always use digital HDMI provided by nVidia drivers and not analog DIN output with add-on audio cards. I tested a few HDR titles using passthrough matching refresh rates @23Hz. I didn't even bother to test SDR and I'm not going to. I didn't take note of the exact refresh rate before I reinstalled working drivers.

Using the combo I posted, for SDR and HDR at 23Hz through 60Hz... I have zero problems except a frame drop every 4.5 mins. All audio types work like they're supposed to and audio is produced automatically from display speakers and/or AVR speakers per ON/OFF state of AVR. 12bit retains after reboot. I get no banding introduced using my Q90R and I (think) colors pop more. NV HDR is used when it should be and O/S HDR is used when it should be (automatically). The only reason I even attempt new drivers is to simplify the audio workaround and this time was only because folks are jumping for joy about no more frame drops... which might be great for analog DAC users but 99% of us carry audio over HDMI digitally.

[huhn]

theses fixes are mostly for digital user.
the clock generator for audio and video could and even should be the same if only 1 HDMI out is used.

[brazen1]

theses fixes are mostly for digital user.


Not this one.

3D was fixed alright.
Auto switching between display vs AVR audio was fixed wonderfully too.
Offering 10bit RGB output is another ignore and sweep under the carpet adventure.
A frame drop every 4.5 mins remains. This is the latest fix.
Retaining 12bit so we could dither to 10bit after a reboot was a beauty of a fix also.
I can't even remember all the other fixes.

Hey we did get an API for software devs to utilize though! And after months of trial and error it finally switched HDR on and off. That was a nice bone tossed to us. Other than that, I can't think of a single fix that did me any good tbh. Gamers might be jumping for joy though... except I don't play games and never will.

[huhn]

you are missing the point. because it works for me it doesn't mean it will do anything for anyone else. i wanted to make that as clear as possible by stating i'm using analog and that's i'm lucky and the digital HDMI user should check this better out in detail.

nvidia has no clue and no affect on the clock of any sound card that is not there own digital output. so the only generally thing user could test on the fast is if this change did anything is the shown refresh rate using studio driver 451.77.

if you really care about motion this is the biggest upgrade you could get (you should use CRU if you really care about motion but what ever).

if it doesn't work for you you should say so because madshi is still in contact about this topic.

[brazen1]

I am a digital user and I have checked it out and I am passing on my observations. I have no motion problems that needs CRU to cure because I'm not using an OLED that has motion problems to begin with. Just one single dropped frame every 4.5 mins which I don't consider a motion problem beyond the 5ms it lasts. I don't think I'm missing any point? Fwiw, I'm not referring to your posts about your observations specifically. I'm actually referring to another person on another forum stating the latest driver 'fixed' the frame drop problem. It didn't for me. I don't know what that person could be doing different. You stated you are using analog with good results. That's a difference. Unfortunately, I could care less about analog improvements and as you and I both stated, HDMI might be a completely different outcome and my quick testing resulted in exactly that.

[huhn]

well you have a major motion issue every 4.5 mins. if you don't notice well fine but that doesn't mean it's not there.

1080p is been fixed for a long long time now this should count for 3D too.

i have no clue why you would take OLED or what ever tech into this dicussion why have nothing todo with this all technologies suffer the same under this because they get a flawed signal that can not be fixed.


why would a gamer or a game ever care they work totally different this change is only for video playback 23p is only for video playback.

ok i will report to madshi that nvidia fix doesn't work. neither the reported frame rate nor the dropped repeated frames changed what so ever correct?

edit: because you edit it is 42 ms just saying and no frame interpolation doesn't change that 42 MS of information are missing by a drop and 42 ms of information are repeated for a repeat.

[brazen1]

Oh, I notice it but it doesn't bother me enough to strive for a cure. If nVidia provides one great otherwise I'm not messing with CRU, custom timings, or anything else. It was fun at one point but I've lost interest. I didn't test 1080p only 2160p. I don't use 3D anymore so I will not be testing anything to do with it. I brought up gaming fixes because that is the only thing ever listed in detail about driver updates. Mostly the actual names of specific games and the actual improvements for them. I don't see too much of anything listed as addressing HTPC users though. However, I am glad madshi has been and continues to be a voice for the HTPC community and gets results. That said, before you report my findings to him... take account of others too. I only tested enough to know I was never going to keep and use it. I never even thought about repeated frames being interpolated or not. Interesting.

[Klaus1189]

Very informative. I am interested in audio over HDMI to AVR, so 23 Hz will get about 10 hours+ with no framedrop/repeat out of the box.
@huhn: please let madshi know that it doesn't work now with HDMI. Nvidia has to do some more tweaking.

[nevcairiel]

@huhn: please let madshi know that it doesn't work now with HDMI. Nvidia has to do some more tweaking.

Did you actually confirm this yourself, or just repeating what others said? Because only one of those is useful. :)

[Klaus1189]

I can not test this, because I don't have a Nvidia card here at the moment and both huhn and brazen1 are very reliable, they are not average users.

Perhaps you can test this and contribute to this issue?

[huhn]

so we have 1 test with numbers under terrible conditions (which is on par with my AMD results i'm used to) and one report of doesn't work with no numbers.

i don't know what to say anymore.

[Asmodian]

Would a test with HDMI and an on-motherboard digital or analog audio out be useful?

[huhn]

audio output with the same card is interesting the rest is all gambling because there is just a different clock generator.

[Klaus1189]

Then I must apologize. I think I missunderstood the hole thing. Please use easy words for a dummy like me :o
I want to buy an ampere card soon and when I read through the threads here I came across the news of better 23 Hz timings out of the box for new drivers and madshi is still in contact with a guy of Nvidia.

I have at the moment no Nvidia card to test, so I have to rely on user feedback here.
Please answer the following question, so a silly guy like me can understand that, too:
1.) What driver version (exact version) is supposed to have better out of the boxed 23 Hz timings?
2.) Are there more than one now?
3.) To what audio outputs does that apply?
4.) I am only interested in HDMI audio. Does that have better dropped/repeated frames?

I kindly ask you guys who have a Nvidia card: Please post a screenshot of madVR stats with the new driver, so I can clearly see the better timings out of the box without CRU and possible other stuff.

[Klaus1189]

HDMI digital is 1 h for me absolutely useless

if it doesn't work for you you should say so because madshi is still in contact about this topic.

These are some of the parts that led me to ask this. So 1 hour is definitely to short in my opinion, but hey what do yI know. Perhaps I misread it also. But as always I want to learn and improve my knowledge.

[huhn]

because i have 1 H doesn't mean you will have 1 H.

[oldpainlesskodi]

Well, for fun and giggles, just in case anyone is bored...............I'm on AMD...no issues., out of the box over 1-2 days no dropped frames. (23p)........runs and ducks for cover lol.

[huhn]

madVR still fails with creating direct3d device failed when HDR passthrough is used on my system.

[Klaus1189]

@huhn: Please let me know if should add it in the first post ;)
If so, what was the first version affected, ...

[videoh]

Madvr abandonware Show some respect for the great things he has done for us all.

[huhn]

@huhn: Please let me know if should add it in the first post ;)
If so, what was the first version affected, ...

no clue only happens with overlay rendering(love this mode and i absolutely hate AMD for never adding it even intel has it...) and is most likely just madVR using the nvidia API wrong after the changes to it or overlay been abandoned.

we will get a new madVR version soonish he is currently active in the test build thread. we will know more soon enough.

[Klaus1189]

What benefits has overlay mode?

Posts: 1,000 ... Yee-haw ;)

[huhn]

doesn't care about display composition.
doesn't care about other programs running.
can't produce tearing.
ignores the WDDM the wddm banding bug was a great one...
and it's fast. these are the first thing coming to my mind.

[CZ Eddie]

I realize it is IMPOSSIBLE to know for sure, but what are everyones thoughts on RTX 3000 driver compatibility with MADVR?
What are your personal expectations on how well the 3000 series *might* work out of the gate?

Anyone think Madshi might may already be working on driver compatiblity?

I'm wanting to return my 2060S for a 2070S but part of me wants to wait a few more weeks for the 3070 to hit major resellers.

[SamuriHL]

If what was posted earlier was true that nVidia knows about the bug with madvr and is putting a work around in the next driver build, then the 30x0 cards will work fine. If not, I'm going to cry big boy tears as I'm getting a 3090 to replace my 2070.

[huhn]

madVR uses directx 9 mostly and nvidia will support that so that most of the story.

[CZ Eddie]

This is good to know. Thanks for sharing the opinions, it is reassuring.

RTX 3000 series to be announced the first week of September, per rumors.
And the 3090/3080 to be released the end of September, with 3070 to follow sometime in October (all per rumors).

I have never paid attention to GPU/video card releases before now.

[SamuriHL]

Don't expect these cards to be cheap. Latest rumors are 1399 for the 3090 I want and 799 for the 3080. Those are msrp so expect the real prices to be higher. Again rumors but they are from typically reliable sources. But Nvidia has a habit of not setting actual prices until hours before the reveal. I'm on vacation that day so I'll be watching the reveal live. Can't wait. As soon as they put the cards up assuming they're not 2 grand I'll be pre-ordering one.

Sent from my SM-G975U using Tapatalk

[brazen1]

Video cards are as overpriced as stock in the company. I won't be buying into either.

[CZ Eddie]

Well, I'm unhappy with how loud the fans are in my MSI Ventus RTX 2060 Super when it is hitting anything above ~90% GPU usage.
I could hear it from across the room during quiet scenes!
So, I wanted to upgrade to a 2070 Super with better fans.
But before I could make the move, it went out of stock because of 3000 coming out.

So I will probably wait for a 3070 Super with good fans to be released and then buy it.
Should be under $600 I think.
The one I am hoping for is the MSI Super Gaming X since it is known for less fan noise in the 2000 series and is typically sold at my preferred major reseller.

Sorry to derail driver talk with hardware talk.

[SamuriHL]

The 3070 will be 599 MSRP. Not sure you'll find one for that price. Yes, these cards are overpriced. By a lot. If you want to see the price come down, wait until the Navi reveal and the PS5/Xbox Series X come out. nVidia will adapt or lose at that point. If I were just looking at gaming performance OR video performance, I'd not even CONSIDER a 1000+ priced card. But I'm of the opinion that I'll be able to max out whatever I want in madvr and avoid buying the PS5 and/or Xbox Series X by buying a 3090. For *ME* it is tolerable. And certainly beats spending 12 grand on an Envy.

[VBB]

Video cards are as overpriced as stock in the company. I won't be buying into either.

I'm with you on that. Will be using my 960 (same as yours) a little longer. Honestly, it does everything I really need, and all I watch is 1080 and above on a 65" 4K screen. I'll probably buy a new TV before I buy a new video card :p

[ryrynz]

I'm with you on that. Will be using my 960 (same as yours) a little longer. Honestly, it does everything I really need, and all I watch is 1080 and above on a 65" 4K screen. I'll probably buy a new TV before I buy a new video card :p

You say this.. and then you get a better card (1060 or above) and do NGU upscaling with SSIM2D downscaling and you go.. Oh, okay :P YMMV

[chros]

So I will probably wait for a 3070 Super with good fans to be released and then buy it.
Should be under $600 I think.
The one I am hoping for is the MSI Super Gaming X since it is known for less fan noise in the 2000 series and is typically sold at my preferred major reseller.
It won't solve it: the only solution is to underclock your GPU with MSI Afterburner (look for youtube videos).
Mine is a 1060 6GB Gaming X with 2 fans on it, that's supposed to be 120W (!) card and if it works on full power you can easily hear it and gets really annoying shortly.
Hence I underclock the GPU core to 1569MHz @ 0.8V, this way the fans max RPM is only 700 and they don't spin all the time. And I don't have any other fan in the Ryzen PC :)

I'm with you on that. Will be using my 960 (same as yours) a little longer. Honestly, it does everything I really need, and all I watch is 1080 and above on a 65" 4K screen.
Yes, same here, no way I spend more than 250 for a freaking GPU :)
As I said before, the only good thing about the state of madvr is that we don't have to upgrade our GPUs until AV1 decoder will be required :)

[VBB]

The one thing my card can't do (without some serious sacrificing, if at all) is tone map HDR to SDR, which is not a priority for me at this time. I've been using chroma and luma NGU low to high, depending on the source. No need for any downscaling with what I watch. I also don't use any additional processing, but I do use a LUT as well as smooth motion. Leaves me with plenty of headroom for render spikes. There might be a RTX 3070 in my future, once the dust has settled :)

[chros]

Just to do some proof of what I'm saying, I enabled Windows HDR and loaded up an HDR mkv in PowerDVD 20 Ultra. As mentioned, no HDR metadata. Dropped back to 442.92 studio driver, all metadata restored, both in madvr hdr output and in PowerDVD + Windows HDR. nVidia broke it. They need to fix it.
Thank You for checking, just upgraded from 385.28. I had small issues with it on LTSC 1809 in the last couple of months.

I can confirm that everything is working with 442.92 SD (except for 3d that I don't care about) :
- HDR10 metadata
- custom res 23p with CRU
- 12 bit falls back to 8bit at 60Hz, and back
- MSI Afterburner for underclocking/volting
- nvidiainspector for monitoring

Note that there's no performance difference between the 2 using d3d11native.

I still have audio misfunctions (I had the same with 385.28 as well, so maybe M$ screwed up something):
- set a speaker config, finish, if you open it up again it shows the default values (stereo)
- same for sample rate
- sometimes it resets to default without opening up the setting panel

[SamuriHL]

If we get a new driver I'll have to hook the fury back up. Right now it's not in my chain as I had some nasty issues going on after a power surge and had to pull everything apart. But when we get the next driver I'll throw it back in the chain and test it. I don't suspect we'll get another driver until after the 1st when the new cards are announced. I need to remind myself to DDU the driver once I get the new card unlike what I did when I went from the 1060 to the 2070. LOL I seriously hope they get this fixed by then.

[chros]

No need for any downscaling with what I watch. I also don't use any additional processing, but I do use a LUT as well as smooth motion. Leaves me with plenty of headroom for render spikes. There might be a RTX 3070 in my future, once the dust has settled :)
Yes, same here, except for Smooth Motion :D and I only use enhance details and crispen edges, nothing else.
We will see how power hungry the new GPUs will be, I'll never (!) buy a card that is more than 120W (!) that's for sure :)

[SamuriHL]

Oh then you will never be getting a 30x0 card. ;) These things are rumored to be EXTREMELY thirsty. That's the reason for the 12 pin power connector. Somewhere north of 300 watts on the 3090. I can't imagine the 3060 will be below 200 watts but you never know I guess.

[chros]

I can't imagine the 3060 will be below 200 watts but you never know I guess.
:D I hope you're joking ... :D You know, below 200W and 250 quid :D

[huhn]

you are ware a 2060 is at 160 watt?
and the 3090 is rumoured to have a new connector that just shipped with the card and the "only" reason i hope to use this connector instead of 2 8pins is the very tiny PCB. getting 375 watt into a GPU these days is nothing special. maybe nvidia just likes following specs there was something about 2 8pins to be out of spec or something.

There are cards that use two 8-pin connectors, but this has not been standardized yet as of 2018, therefore such cards must not carry the official PCI Express logo. This configuration allows 375 W total (1x75 W + 2x150 W) and will likely be standardized by PCI-SIG with the PCI Express 4.0 standard.
https://en.wikipedia.org/wiki/PCI_Express

and these cards should be supported on PCIe 3.0 systems so here you have it.

[chros]

you are ware a 2060 is at 160 watt?
Yes, that's why I didn't care and I will not care in the future either.

[huhn]

just get a smaller card when they are released

[SamuriHL]

Yea sadly I wasn't joking. LOL And definitely not on price either. They should in theory be making some non-RTX cards again that should fit the bill on price and wattage. Maybe.

[chros]

Thank You for checking, just upgraded from 385.28. I had small issues with it on LTSC 1809 in the last couple of months.

I can confirm that everything is working with 442.92 SD (except for 3d that I don't care about) :
- HDR10 metadata
- custom res 23p with CRU
- 12 bit falls back to 8bit at 60Hz, and back
- MSI Afterburner for underclocking/volting
- nvidiainspector for monitoring

Note that there's no performance difference between the 2 using d3d11native.
Actually I found a bug, not a small one, with madvr hdr output and 442.92 SD:
- start a HDR10 clip
- don't close the player
- start an SDR clip
- TV stays in HDR10 mode, it doesn't switch back to SDR! :mad:

It works the other way around: SDR -> HDR10
This was fine with 385.28.
Let me know if someone can reproduce it ...

[oldpainlesskodi]

Unable to replicate this on AMD. I normally use Kodi DSplayer, but tried as per your instructions using HPC-HC x64, and no problems.

[chros]

Unable to replicate this on AMD. I normally use Kodi DSplayer, but tried as per your instructions using HPC-HC x64, and no problems.
Thanks, but this is an nvidia driver bug :)

[oldpainlesskodi]

Ah ok.

[CZ Eddie]

The RTX 3000 card specs have been leaked:

https://videocardz.com/newz/nvidia-geforce-rtx-3090-and-geforce-rtx-3080-specifications-leaked

----------
NVIDIA GeForce RTX 3090 features GA102-300 GPU with 5248 cores and 24GB of GDDR6X memory across a 384-bit bus. This gives a maximum theoretical bandwidth of 936 GB/s. Custom boards are powered by dual 8-pin power connectors which are definitely required since the card has a TGP of 350W.
----------
The GeForce RTX 3080 gets 4352 CUDA cores and 10GB of GDDR6X memory. This card will have a maximum bandwidth of 760 GB/s thanks to the 320-bit memory bus and 19 Gbps memory modules. This SKU has a TGP of 320W and the custom models that we saw also require dual 8-pin connectors.
----------
The GeForce RTX 3070 also launches at the end of next month (unless there is a change). We can confirm it has 8GB of GDDR6 memory (non-X). The memory speed is estimated at 16 Gbps and TGP at 220W. The CUDA Core specs are still to be confirmed.
----------

https://i.ibb.co/vq8y1dd/Untitled.jpg

[SamuriHL]

Now they just need a freaking working driver for them.

[Manni]

Now they just need a freaking working driver for them.

Yes, I'll wait for you to confirm this before I consider ordering one :)

[SamuriHL]

Yes, I'll wait for you to confirm this before I consider ordering one :)

Yea that's understandable and totally fine. :) I'll hook the Fury back up before the new card comes. I plan on pre-ordering it Tuesday once they unleash them. Should be fun.

[chros]

Actually I found a bug, not a small one, with madvr hdr output and 442.92 SD:
- start a HDR10 clip
- don't close the player
- start an SDR clip
- TV stays in HDR10 mode, it doesn't switch back to SDR! :mad:

It works the other way around: SDR -> HDR10
This was fine with 385.28.
Let me know if someone can reproduce it ...

It works fine after restarting the PC :) , I don't know what happened (maybe experimenting with Windows HDR toggle?). I'll keep an eye on it.
Note that I only installed the GPU driver and HDMI audio driver with NVCleanstall, nothing else (not even PhysX), everything else is removed.

@Klaus1189: you can put the following utils into the first post for nvidia cards:
- NVCleanstall (https://www.techpowerup.com/download/techpowerup-nvcleanstall/) for removing bloatware from driver
- MSI Afterburner (https://www.msi.com//page/afterburner) for under/over-clocking/volting
- nVidia Inspector (https://www.guru3d.com/files-details/nvidia-inspector-download.html) for monitoring (nvidiaInspector.exe -showMonitoring)
- CRU (https://www.monitortests.com/forum/Thread-Custom-Resolution-Utility-CRU) for creating 23p custom resolution (guide (https://forum.doom9.org/showthread.php?p=1868998#post1868998))

[Klaus1189]

Done. If you or anybody else want anything else or some further descripion, please let me know and I add it.

[chros]

You say this.. and then you get a better card (1060 or above) and do NGU upscaling with SSIM2D downscaling and you go.. Oh, okay :P YMMV
How do you do 1080p -> 4k like this with madvr (NGU Sharp)?
There's no option for "always" at quadrupling sections, only at doubling section, so it only doubles it.

[SamuriHL]

Looks like I'm going with a 3080 this gen. I think it's going to be quite the card. Now, let's get those drivers.

[chros]

Looks like I'm going with a 3080 this gen. I think it's going to be quite the card. Now, let's get those drivers.
:D Yes, you don't have to go for the overpriced 3090 ($1500).

Just watching the live event (commented by 2 guys), the cuda core counts weren't right, this is huge (https://www.nvidia.com/en-us/geforce/graphics-cards/30-series/compare/)! (and here (https://wccftech.com/nvidia-geforce-rtx-3080-8-gb-official-launch-price-specs-performance/))
- 1070 / 1080 / 1080TI: 1920 / 2560 / 3584
- 2070S / 2080S / 2080TI: 2560 / 3072 / 4352
- 3070 / 3080 / 3090 : 5888 / 8704 / 10496

I'll be curious about the lower cards later (3060, 3050, etc), they should be the real killers...

It's on Samsung 8nm, probably that's why they're not that good with power.
They mention HDMI 2.1 as well, the question is whether it's 40 or full 48Gbps :) And whether they enable (finally) 10 bit output for gaming GPUs as well ...

Another thing popped into my mind is whether these early cards will get AV1 decoder or just the seconds batch of cards (e.g. 3060) as nVidia used to do it in the previous generations (if there will be an AV1 decoder at all).

Lastly, as I said before, we don't have to upgrade GPUs for madvr itself :)

[SamuriHL]

YOU don't have to upgrade the cards for madvr, but, for my use case I most definitely do. :) If I set my desired settings in madvr for some of the upscaling I do, it turns my 2070 into a slideshow. AV1 decoding was listed on the 3090 but I did not see the same listed on the 3080. That would suck if it's not there. The HDMI 2.1 bandwidth question is a valid one for sure. Power I'm not concerned with as I have an 850 watt gold PSU already. The price differential to me make the 3090 not quite worth it as I have no intention of going to 8K in the next 5 years. Given what I saw of the raw power of the 3080, I think I'm good with that decision. And if I change my mind later down the road, I'll have only spent 699 vs 1499. :D

[SamuriHL]

Ah, perfect, thanks for that. That's great news. I think I'm going to be quite happy with the 3080!

[chros]

https://www.nvidia.com/en-us/geforce/news/rtx-30-series-av1-decoding
Ampere GPU family all have AV1 hardware decoding.
Cheers, that's good to hear. This will be the only reason when I'll upgrade.
Hopefully they talk about bitstream support, not the handicapped one :)

YOU don't have to upgrade the cards for madvr, but, for my use case I most definitely do. :) If I set my desired settings in madvr for some of the upscaling I do, it turns my 2070 into a slideshow. AV1 decoding was listed on the 3090 but I did not see the same listed on the 3080. That would suck if it's not there. The HDMI 2.1 bandwidth question is a valid one for sure. Power I'm not concerned with as I have an 850 watt gold PSU already. The price differential to me make the 3090 not quite worth it as I have no intention of going to 8K in the next 5 years. Given what I saw of the raw power of the 3080, I think I'm good with that decision. And if I change my mind later down the road, I'll have only spent 699 vs 1499. :D
:) Yes, 3080 is more than enough.
Although I'm not sure why you need such processing power for madvr, my 1060 can double/quadruple everything (!) with luma NGU Sharp (from Low to High, depending on resolution and framerate), and that's more than enough with a 65" screen from normal viewing distance. :) Of course, it's your choice...

About the HDMI, Vincent states (https://youtu.be/-McftdJMXP4) that the new Xbox will have only 40Gbps :)

[SamuriHL]

I am well aware you don't understand why I need the raw power in madvr. LOL I record a lot of shows on my TiVo, strip the commercials, and throw them on my NAS. Because my cable company sucks, they re-encode half the stuff to 720p 60. I upscale this to 4k 60. The settings I like to use after playing around with them make my 2070 cry for help and report me for abuse. LOL! So I have to compromise. Which still looks good, don't get me wrong, but, why should I want to compromise when I know madvr is capable of giving me the picture quality I want? Besides the fact that I do game and these new cards will save me from having to buy a new console. MS couldn't care less if you buy their hardware so long as you're buying their games. And that's the right approach. But back to madvr, I don't want any compromises. I wanna be able to set the quality I want and forget about it. The 2070 just isn't quite there. A 3080 is 50% faster than the 2080ti and a full 100% faster than the 2080. I will be quite pleased I think.

As for HDMI 2.1, not relevant to me until I upgrade my panel again and that's another 3 or so years away unless someone REALLY impresses me with the perfect screen.

[chros]

Yes, I won't upgrade my B8 either soon.
About madvr: do you actually see the difference between your current and your wishful settings? :) I mean 720p is already crap, nothing can make wonders out of it. What's your current settings with 720p@60?

[SamuriHL]

Yes, I see a VERY noticeable improvement with the settings I desire. The settings I have on there right now aren't bad but as I said, I've seen what can be done with higher settings. You're not wrong that 720p kinda blows, but, it can be vastly improved. And that is one of the motivations for a faster card.

[ryrynz]

:D Yes, you don't have to go for the overpriced 3090 ($1500).


How can it be overpriced if there is zero competition? It's priced to sell and sell it will.

https://www.nvidia.com/en-us/geforce/news/rtx-30-series-av1-decoding

Ampere GPU family all have AV1 hardware decoding.

Expected and glorious. It's time for H.xxx to die figuratively speaking.

[NikosD]

How can it be overpriced if there is zero competition? It's priced to sell and sell it will. According to specs, the 3090 card of 1500$ is ~20% faster in TFLOPS and ~23% faster in memory bandwidth than the 3080 card of 700$.
The only thing that "justifies" the >2x price is the amount of memory (2.4x)
Who needs 24GB of Video RAM ?

[ryrynz]

The 20% justifies it too for people who want the best. It's not economical but it's not outrageous look at the Titan releases. They have to add something extra and this step in GB does start lead the way forward which has to be done. Sales will also determine its true market price. As for who needs it? Developers only I guess.. I'm unaware of any game that approaches 16GB let alone 24 but if you hold on to this card for a few years it's possible with new games or with high end texture packs for older ones you'll use greater than 16GB.

[nevcairiel]

Price/performance always goes off the rail at the top end. Part of that is that making that big of a chip with everything functioning is just more expensive then a cut-down one, part is the memory, and last part is indeed the premium tax for having the best available.
Compared to last gens Titan, it is however almost "affordable".

[chros]

How can it be overpriced if there is zero competition? It's priced to sell and sell it will.
Like the RX20 cards, right? :) I understand your point, there a lot of new features in them.
I meant compared to 3070/3080.

How do you do 1080p -> 4k like this with madvr (NGU Sharp)?
There's no option for "always" at quadrupling sections, only at doubling section, so it only doubles it.
And you forgot to reply to this one :)

[NikosD]

2080 Ti had a ridiculous price of 1200$ when it was released a 70% more than 2080 with just a 35% more performance at 4K.

Now 3090 which is the card above 3080 has a 1500$ price tag which is more than 110% than 3080 with even less performance difference of around 20% ?

So the 35% performance difference had a 70% price difference for 2080/2080 Ti but now the 20% performance difference leads to 110% price difference which is unacceptable.

The only real reason for this, is that probably AMD is going to compete with 3080 but not with 3090, so this is a price from a ruthless monopoly.

And the leather-jacket-man knows it.

[huhn]

the memory is one major reason. that's a lot of useless memory for most user.
the other is the DIE is gigantic lowering yield drastically.
compared to the RTX titan (i'm pretty sure this is not the biggest card they are going to release at all) cost 2.5 as much as a 2080 ti with barely any difference in performance but tons of Vram (professional will make use of them).
it's not mean to be competitive it's just "look at what we can do".

having one very expansive card but the rest is just amazingly priced is just fine.just ignore this one card.

this is what the industry needed.
more performance per buck (this time massively).

about new features.
the 2000 series mostly added RT cores which didn't work out this well... they can't do Pathtracing at good frame rates today and they may never be able to do pathtracing with good frame rates but these cores while barely used where new.

tensor cores where already on the market before they just reused them properly and that just great.

INT32 pretty much ignored but it seems ampere has some very very cool trick up with INT32. time will tell but they where new with turing and they seem to be "gone/different" with ampere.

a very small update to the decoder and a update to the encoder.

what has the 3000.
HDMI 2.1. TV with 2.1 are out for a while now that's actually right now useful.

AV1 decoding and "massive" update to the encoder.

support for direct access to the hard drive i doubt we can use this for hardware decoder. that's quite a big thing i guess consoles have it so high chance to be generally useful.

reusing the old INT 32 cores as FP32 cores or something like that?

i can not give nvidia a lot of credit for HDMI 2.1 that's not a lot of "work" like RT cores so i have to agree the feature set on the 2000 series was bigger but the 3000 series has some quite nice stuff too.

[NikosD]

How do the buyers of 2080 Ti at 1200$ feel now ?

Especially the recent ones.

If you' ve bought recently a new custom 2080 Ti at 1500$ and see the 3070 with 500$ having equal performance (or a little more according to nVidia) how that makes you feel ?

Sucker.

[huhn]

bad day?

[chros]

How do the buyers of 2080 Ti at 1200$ feel now ?
:D Loosing $700 just in 1 week :D

AV1 decoding and "massive" update to the encoder.
What do you mean about the encoder part? Do you have a link? Thanks

[brazen1]

Who in there right mind would spend this much money on... wait for it... A VIDEO CARD? And to think I thought the stock market was the main thing disconnected from reality. Welcome to the new mainstreet bubble pushing the envelope. I couldn't agree more. "Suckers".

[NikosD]

Rumors of extremely limited supply of Ampere cards, almost non-existent for all remaining 2020.
The usual nVidia tricks to raise the hype with pre-orders or strategic move to see Big Navi in action ?
Ebay is flooded by used 2080 Ti cards selling below 500$
Things look ugly for customers of nVidia, once again.
Will AMD Save Our Souls regarding GPUs just like it did with CPUs ?
I'm praying for the AMD we trust.

[brazen1]

Since you guys like to impulse buy regardless of cost and feed from the hand, here's a 16TB SSD. It's only 4 grand. https://www.tomshardware.com/news/teamgroup-announces-153-tb-sata-ssd-for-client-pcs

You could boycott the thieves and cure their insanity you know. I am.

[el Filou]

Who in there right mind would spend this much money on... wait for it... A VIDEO CARD? And to think I thought the stock market was the main thing disconnected from reality. Welcome to the new mainstreet bubble pushing the envelope. I couldn't agree more. "Suckers".It's the same as people who spend tens of thousands on sports/luxury cars. Some people have the money and prefer to buy a vastly more expensive model to have a moderate improvement in quality/performance. They exist, but it's a small percentage of customers.

[ryrynz]

And you forgot to reply to this one
How do you do 1080p -> 4k like this with madvr (NGU Sharp)?
There's no option for "always" at quadrupling sections, only at doubling section, so it only doubles it.


No can do ATM, given 8K TVs are now very much a thing and the performance now afforded by the 30 series, madshi should ideally add this feature for always - supersampling quadrupling and beyond.

Re:3090
To give you an idea The 2080Ti is used by about over 850K users on steam, you could multiple this more than a few times to guess cards in the wild.
The 3090 will have hundreds of thousands of sales eventually, maybe half a million or so. Will probably go a ways to helping the more aggressive pricing of this series this time round.
This card is in part funding these price points, it needs to sell. People be making out like Nvidia going to sell a half dozen of these max [emoji14]

[SamuriHL]

I'll be honest, one of the things I was considering when I was looking at the 3090 was exactly that...supersampling 4k to 8k and downscaling it back to 4k like some of you do for 1080p->4k->1080p. I've no idea what that'd look like (and neither does anyone else because it currently can't be done) but the thought most definitely crossed my mind. Alas, given what they showed with the 3080, I'm sticking with that this round. Then I can work on getting the intel side of the house to pcie 4.0 at some point. Maybe down the road madshi can figure out how to benefit from RTX IO. LOL :D

[brazen1]

You're comparing apples to oranges. Cars come in all shapes and sizes. The majority of people in this world just desire something to get from A to Z in and could care less about all the prestige and worthless bells and whistles to show off because anything decent is unaffordable. So they pay through the nose for an 'entry' level car because it's the only low price there is. They have no choice. Even buying a used car is overpriced. Are they happy about making monthly payments for the rest of their life on something that depreciates as soon as they drive it? Yeah, I'm sure they're jumping for joy. Glitzy bling is reserved for folks who spend beyond their means and that's who are aimed at. The internet prowls for subjects. Sure, those that actually have more money than they need (or earned) exist but they aren't interested in video cards. An entry level video card, of all things, costs $500 now? LoL. How on earth will we make our Lamborghini payment this month?

[SamuriHL]

The 3070 is hardly the "entry level" video card. That honor will go to the 3060 who's price has yet to be unleashed. But I suspect that's where the "fight" will be so I wouldn't be entirely shocked to see it unleashed for 299 USD.

[nevcairiel]

Yeah a x70 model has never been entry level, thats squarely in the mainstream, but also the model of card they are likely to sell the most of.
There will be more models in the future at lower price/performance points, 3060 is definitely on the roadmap, maybe a 3050.

The prices of the established models didn't actually increase at all this generation, eventhough the performance jumped almost 80%. 3070 and 3080 have the same price as a 2070 or 2080 respectively.
The only "new" thing is the 3090 which sits somewhere between the old top-end Ti and the Titan. And you don't have to buy it if you don't want to. But even the Titan RTX had a market, and that one cost 2500 dollars - with the 3090 you get that level of performance/memory for a 1000 less as well.

There is also a good chance there will be either a 20GB 3080, or a 3080 Ti if you wanted something higher but not quite as expensive as the 3090, but those would require some patience.

And your statements that anyone that buys luxury products does it "beyond their means" is just asinine. Especially in computer hardware. IT jobs pay well around here, and those are the people that also like hardware.
Personally, I only use money I make through my hobby projects to upgrade my PC, and I can easily afford a 3090 if I want one. Which I might, baring reviews. Buying hardware every couple years isn't that much money, even if one indulges in "luxury". But then having this hardware is also what makes me part of that money, so :o Although I don't have (or need) a car, so much money saved there. City living is neat.

[VBB]

A 3060 at $299 sounds very likely. I'm actually quite happy with Nvidia's pricing this time around.

[el Filou]

You're comparing apples to oranges. Cars come in all shapes and sizes.Just like PC hardware.Sure, those that actually have more money than they need (or earned) exist but they aren't interested in video cards.How do you know that? Just like nobody "needs" an expensive car to get from A to B but there are still people who buy expensive cars, nobody "needs" a high-end graphics card but there are still people who buy them. Heck, you don't even "need" a car, you could take public transport or cycle or walk, just like I know people who game on Pentium or Celeron laptops with the lowest-end Intel iGPU and are just happy they can game. it's just how things are.

[VBB]

Or someone like me, whose madVR rig consists of a 1st gen Intel Core I7 920 and a EVGA X58 mobo from 2008 ;)

[videoh]

Things look ugly for customers of nVidia, once again. Things look very sweet and beautiful to me.

[videoh]

"Suckers". So sorry you are too poor to ride the bleeding edge, but sour grapes are unbecoming. I will pray for you.

[SamuriHL]

If I had the patience nev I'd wait for one of those higher memory 3080 or 3080 Ti. But I don't. Lol. I've not completely ruled out a 3090 yet but I see no compelling reason for one, even in this hobby, over the 3080.

Sent from my SM-G975U using Tapatalk

[NikosD]

Things look very sweet and beautiful to me. You can also burn your money in your fireplace instead of buying 2080 Ti or 3090.
Noone can stop you.
AMD took advantage of Intel's stagnation at 14nm and Skylake architecture since 2015 and attacked ferociously on 2017 with Zen architecture.
It took only 2 years for Zen 2 to destroy Intel for good on 07/07/2019.
What a day!
Intel is AMD's bitch since then.
nVidia made a huge mistake with Turing architecture on September of 2018.
The rasterized performance difference of 1080 Ti and 2080 was ZERO.
Now, if AMD doesn't take advantage of this huge mistake of nVidia two things could be true:
1) They are somehow incompetent to build GPUs for the high-end segment which is very bad for the industry.
Radja did a big harm with Vega architecture, but in 2020 it's not a real excuse.
2) They have left the high-end most profitable segment to nVidia in order to keep the consoles and lower segments which is extremely bad for the industry because GPU cards are a duopoly (after decades of nVidia's criminal behaviour) and with 80% of discrete cards to nVidia nowadays, it's actually a covered monopoly.

I'll wait for Big Navi in order to decide for my next card and see what is really going on with AMD GPU department.

[huhn]

the fact that navi had the same price hike as the turing cards is irrelevant yes?

intel CPU sales are not really going down at the moments it's the prices they reduced they are still an option and sometimes they even have the better product.

[chros]

You can also burn your money in your fireplace instead of buying 2080 Ti or 3090.
Noone can stop you.
AMD took advantage of Intel's stagnation at 14nm and Skylake architecture since 2015 and attacked ferociously on 2017 with Zen architecture.
It took only 2 years for Zen 2 to destroy Intel for good on 07/07/2019.
What a day!
Intel is AMD's bitch since then.
nVidia made a huge mistake with Turing architecture on September of 2018.
The rasterized performance difference of 1080 Ti and 2080 was ZERO.
Now, if AMD doesn't take advantage of this huge mistake of nVidia two things could be true:
1) They are somehow incompetent to build GPUs for the high-end segment which is very bad for the industry.
Radja did a big harm with Vega architecture, but in 2020 it's not a real excuse.
2) They have left the high-end most profitable segment to nVidia in order to keep the consoles and lower segments which is extremely bad for the industry because GPU cards are a duopoly (after decades of nVidia's criminal behaviour) and with 80% of discrete cards to nVidia nowadays, it's actually a covered monopoly.

I'll wait for Big Navi in order to decide for my next card and see what is really going on with AMD GPU department.
I agree with you completely.

the fact that navi had the same price hike as the turing cards is irrelevant yes?
They got greedy :) And it was a huge failure. Let's hope this year they won't make the same mistake.

intel CPU sales are not really going down at the moments it's the prices they reduced they are still an option and sometimes they even have the better product.
Not with the missing pci express 4.0 :) That's the final nail in the coffin. :D
Will they go bankrupt now? No, not now. But if they do things like this they definitely will!

[chros]

A 3060 at $299 sounds very likely. I'm actually quite happy with Nvidia's pricing this time around.
I doubt it unfortunately, more like 350/400.
If the 3080 is 30% faster than 3070, and 3070 is like 2080Ti, then probably 3060 will be like 2080 which is 30% slower than the 2080Ti.
Let's hope for the best :)

[chros]

The HDMI 2.1 bandwidth question is a valid one for sure.
Vincent states (https://www.youtube.com/watch?v=GnRdJQMf6C0) that it's full 48 Gbps.

[SamuriHL]

Beautiful. I'm super excited to get my hands on a 3080 and see what it can do.

Sent from my SM-G975U using Tapatalk

[chros]

Beautiful. I'm super excited to get my hands on a 3080 and see what it can do.
:) Let us know about all your experience then!

[SamuriHL]

That I shall if I'm able to get my hands on one. They are going to be scarce until next year sometime.

Sent from my SM-G975U using Tapatalk

[nevcairiel]

They said that for so many launches before, and I never had issues getting the exact model I wanted without significant wait time. I wonder if such scarcity just affects some countries more then others, for reasons.

[SamuriHL]

That's entirely possible. Of course, with this pandemic still raging in quite a few countries, it's anyone's guess. Plus they're on a different Samsung 8nm process that they don't yet know what the yields are going to be. This would not be the time to screw up supply else they will end up losing sales to consoles which seems to be the point of these new cards in the first place.

I STILL want a fixed driver though. Windows has taken to continually reinstalling the latest broken driver because I bothered to install the non-studio driver this past time I fixed it. I'd prefer they just fix the damn thing so I can move forward because in a couple weeks, if I'm able to secure the card I want, I'm going to NEED a new working driver.

[chros]

Well, good luck then, you'll need it :)

[VBB]

I'll be waiting for feedback from Asmodian and the rest of the C9/CX owners over at AVS. Proper, full chroma along with banding-free 10-bit at 120Hz is my main reason to upgrade eventually. Of course, I'll also need a new TV to go with that :)

[chros]

As it seems we have to wait for couple of generations for this with LG if it will ever happen :D (I mean HDR10 issues in PC mode)

[SamuriHL]

Yea, I have very little hope of that happening like anytime soon if ever. We're on how many generations of LG OLED now and it doesn't seem to really improve all that much. But yes, having a card that's ready will be nice. I don't plan on upgrading from my C8 unless I see a massively compelling TV come out. I'm not going to hold my breath waiting for that. LOL

[nevcairiel]

I still hold out hope that 2021 will be a good year for TVs, at least for those of us that didn't buy into OLED yet

[chros]

I still hold out hope that 2021 will be a good year for TVs, at least for those of us that didn't buy into OLED yet

I'm not sure Nev, I couldn't choose today either, just to mention a few issues:
- no judder free 23p playback in PC mode with Panasonic/Philips
- microstutters with HDR using normal mode on Panasonic (not with SDR)
- screwed LLDV implementation on Sony (posterization/banding in every DV content!)
- HDR10 posterization / banding using PC mode on LGs

Pick your poison :) Anyway, ask us when you want to buy, we know a lot about these sets :)

[SamuriHL]

Hey I'm open to anything that improves my PQ, so, I'm not married to OLED. It's just at the time I bought it, there was no better picture (that I could afford). Coming from plasma, OLED made the pain easier to swallow. LOL Maybe I'll go 4k short throw laser for my next display. LOL

[nevcairiel]

I'm not that overly married to PC mode, I can barely see a difference in 4:4:4 chroma on a 1080p display at that size, with 4K you get more chroma on the same size still, so maybe it'll be fine..

[QBhd]

I'm not that overly married to PC mode, I can barely see a difference in 4:4:4 chroma on a 1080p display at that size, with 4K you get more chroma on the same size still, so maybe it'll be fine..

I have long ago ditched PC mode on my C8... Banding is far, Far, FAR more of a deal breaker than the lack of 4:4:4.

QB

[SamuriHL]

Totally agree. And once I learned how to calibrate the c8 without introducing banding 444 was off the table. We'll see if these new cards do 10 bit bit that seems highly unlikely.

Sent from my SM-G975U using Tapatalk

[SamuriHL]

Omg as soon as I post that I see a new video claiming the did add 10 bit. Well alrighty then!

Sent from my SM-G975U using Tapatalk

[huhn]

who would have seen this coming after nvdiia added HDMI VRR for LG TVs and the use of a LG 8K TV as a demo that they would support LG TVs unbelievable.

[VBB]

huhn, why do you have to be sarcastic all the time? You give the rest of us Germans a bad rep ;)

As for 10-bit RGB/4:4:4 via HDMI, I'll believe it when I see it in a working driver!

[chros]

:D True.

[QBhd]

Totally agree. And once I learned how to calibrate the c8 without introducing banding 444 was off the table. We'll see if these new cards do 10 bit bit that seems highly unlikely.

Sent from my SM-G975U using Tapatalk

Are you saying you calibrated out the banding that is introduced by using PC mode? I can't see how that is possible...

QB

[NikosD]

No changes for the HW encoder in Ampere.
Exactly the same like Turing.

[nevcairiel]

Are you saying you calibrated out the banding that is introduced by using PC mode? I can't see how that is possible...

QB

Imagine there is a comma there:
And once I learned how to calibrate the c8 without introducing banding, 444 was off the table.

Thus, not using PC mode.

[SamuriHL]

Are you saying you calibrated out the banding that is introduced by using PC mode? I can't see how that is possible...



QBOh hell no I'm not saying that at all! I'm saying when I first got into calibration a year ago the method and software I used created some artifacts in non pc mode. No, pc mode can't be fixed.

Sent from my SM-G975U using Tapatalk

[SamuriHL]

Imagine there is a comma there:


Thus, not using PC mode.Exactly. Bad grammar is all. Lol

Sent from my SM-G975U using Tapatalk

[NikosD]

A few more clarifications regarding the sneaky moves of nVidia regarding Ampere.

The cores are not exactly capable of 2x FLOPS versus Turing.
Each core is capable of INT + FP (like Turing) OR FP + FP (Ampere only) in every cycle.
So, if ALL instructions are FP then 2x FP is possible, but they are not.
And because there are also INT instructions, Turing (using INT + FP) is faster than Pascal.

So, the max TFLOPS number is wrong (in general for every case)

Also, 3080 has the same family GA102 chip like 3090.
It's the maximum configuration reserved only for Ti or Titan in the past.
For example, 2080 has the TU104 and not the larger TU102 chip that it was reserved for 2080 Ti
1080 has the GP104 chip and not the larger GP102 which was reserved for 1080 Ti.

What does this mean ?
It means that the performance difference of 3090 vs 3080 will be as I have already reported very low (~20%) almost half than 2080 Ti vs 2080 or 1080 Ti vs 1080 because they share the same huge family chip, like never before.

Why does nVidia scaled up for the first time the x080 card using the same family chip like the faster Ti/ Titan ?
Because nVidia is afraid of Big Navi performance and wants to be sure that 3080 is as fast as it could be, faster than ever before, pushing the performance to the edge, matching a normal 3080 Ti (if it could exist)

nVidia wants the official flagship 3080 to be faster than Big Navi by all means even using a bigger chip that used to be leveraged by Ti/ Titan and is selling it at the same price of 2080, which is insane.

nVidia definitely counts on the idiots buying 3090 for 20% more performance than 3080 with a 110% difference in price, in order to balance the reasonable price of 3080 and 3070.

I'll be back.

[huhn]

The cores are not exactly capable of 2x FLOPS versus Turing.
Each core is capable of INT + FP (like Turing) OR FP + FP (Ampere only) in every cycle.
So, if ALL instructions are FP then 2x FP is possible, but they are not.
And because there are also INT instructions, Turing (using INT + FP) is faster than Pascal.

great update isn't it? INT 32 is barely used at all they are just like AMD cores now. so that means the max Flops on both are wrong? and only turing is correct because they are dedicated cores sleeping all the time?

[nevcairiel]

FLOPS literally stands for "floating point operations per second", which is always a theoretical max value independent of any other constraints or workloads - so it cannot be wrong. For the context of Ampere, a single core is capable of twice the FLOPS as a Turing one.
FLOPS are literally calculated by something rather simple like "core count * calculations per tick * clock speed", it does not account for anything else. Its not measured. Its not based in reality. Its just math.

FLOPS are a very theoretical value, its not particularly meaningful for real-world performance. Never has been, on any vendor. But companies (all of them) also like to use it to boast, because its easy to make those high numbers.

Its sorta funny to see people scramble to make up reasons why everything is terrible.
Sure, the 3090 is too expensive, but it practically has the specs of a Titan RTX with all that memory and costs almost half of that. As everyone keeps saying, the performance gap is likely small enough to just get a 3080 instead, so might as well treat it like a Titan and ignore it. :)

The 3080 and 3070 are priced/spec'ed very competitively. Of course waiting for RDNA2 might be sensible if you have the patience, maybe it won't be too long anymore. But even without that, the 3070/80 are very worthy replacements of the 2070/80.

[huhn]

i agree with most points but Flops are "useful" getting a rough estimate of the performance between cards from the same generation and only the same generation.

[chros]

Thanks Nev for FLOPS info.

A few more clarifications regarding the sneaky moves of nVidia regarding Ampere.

The cores are not exactly capable of 2x FLOPS versus Turing.
Each core is capable of INT + FP (like Turing) OR FP + FP (Ampere only) in every cycle.
So, if ALL instructions are FP then 2x FP is possible, but they are not.
And because there are also INT instructions, Turing (using INT + FP) is faster than Pascal.
Yes, something is weird with such a huge cuda core count jump.
I understand that it doesn't scale linearly, but I'd wait for 1.5x return out of 2x.
In reality probably it will only be 1.25-1.30. So where the rest of 20% went?
Some think that by adding double of the micro-cores on a smaller node, IPC suffers about 30% decrease compared to Turing.

[huhn]

a 2080 ti has 4352 cores. that means 4352 FP32 and 4352 INT32 cores so it technically has 8704 cores (50% are very small and mostly "useless/barely used" cores).
a 3080 has 8704 that means 8704 FPS32 and 4352 INT32 see the image below.

a cluster now has 2xFP32 not a core is doing 2xFP32. and we got this new not confirmed image: https://cdn.videocardz.com/1/2020/09/NVIDIA-GeForce-Ampere-SM-Layout_HardwareLuxx.jpg they just added the same FP32 cluster again on a different datapath where only the INT32 cores or the FP32 cores are usable so the INT 32 cores are not gone.

a 3070 has 5888 cores while they clearly don't scale linear to a 3080 with 8704 cores but it will be far more then 50 %... but this is comparable they are the same they physical there.

IPC. if nvidia claims are true and are generally applicable across the board (which you should never believe in).
a 3070 is supposed to be as fast as a 2080 TI.
assuming no INT32 was used when we compare 4352 cores at ~2000mhz vs 5888 at ~1700mhz then the IPC for FP32 is 13% lower on ampere and this is the worse case scenario treating the double FP32 as real cores as we should be.

[NikosD]

FLOPS are literally calculated by something rather simple like "core count * calculations per tick * clock speed", it does not account for anything else. Its not measured. Its not based in reality. Its just math.

FLOPS are a very theoretical value, its not particularly meaningful for real-world performance. Never has been, on any vendor. But companies (all of them) also like to use it to boast, because its easy to make those high numbers. OMG, I hope Top500.org is not reading this thread and this post!
The whole world of SuperComputers nowadays are based on real-world benchmarks of FLOPS, meaning that besides the theoretical value of FLOPS that someone can calculate of a processor, the whole world besides you probably, is interested in how many actual FLOPS the processor is capable of, given a specific workload or even a favor workload.
For example, at the last presentation of Radja showing Xe GPU, he used an OpenCL tool that is capable of trying different worksets in order to find maximum theoretical but actual values of FLOPS of a CPU/GPU, meaning that the figure is a best case but real world scenario.
So, we have a theoretical number of FLOPS, we have a theoretical best-case fully optimized number of FLOPS and we have a real-world actual sustained FLOPS leveraging common case workloads as those of benchmarks or real-world apps, like 3D games.
In conclusion:
FLOPS are not only theoretical but also real world and we are mostly interested in real-world measurement of FLOPS.
They are measured in a lot of cases and a lot of different workloads and scenarios.
They are actually the most used benchmark of SuperComputers since the beginning of their existence and of every actual processor (CPU/GPU) ever built, with hardware support of FP.
They are based on reality, because every actual program which uses FP numbers they are actually doing a FLOP (Floating Point OPerations) in hardware that supports FP numbers.
It's not just math, but are useful math in order to test/benchmark real-world performance of workloads/applications like 3D Games since we are talking about graphics cards.
That post that I just replied, could easily be one of the most misleading posts ever written in this forum!
I hope Top500.org measuring the Top500 Supercomputers of the world haven't committed suicide by laughing or crying, reading the post above.

[huhn]

you know what compare a vega 64 and a 3700XT performance and flops and then say that again...

GPU driver and game engines cheat at every corner. just buy a card based on FLOPS because they are the metric.

ridiculous...

[el Filou]

OMG, I hope Top500.org is not reading this thread and this post!
The whole world of SuperComputers nowadays are based on real-world benchmarks of FLOPS, meaning that besides the theoretical value of FLOPS that someone can calculate of a processor, the whole world besides you probably, is interested in how many actual FLOPS the processor is capable of, given a specific workload or even a favor workload. [...blahblah...] I hope Top500.org measuring the Top500 Supercomputers of the world haven't committed suicide by laughing or crying, reading the post above.That's why Nev is only talking about *theoretical* FLOPS, which are the *only* value ever advertised in consumer GPU marketing materials. When Microsoft says "Xbox Radeon GPU has 12 TFLOPS" or NVIDIA says "RTX 3080 has whatever FLOPS", it's not a benchmark, it's a simplistic theoretical formula like Nev explained. He never said you can't benchmark specific code and get a number out of them, but that's not what GPU vendors are advertising.Its sorta funny to see people scramble to make up reasons why everything is terrible.It's because those people are AMD fanboys who just want to rejoice in seeing their competitors not absolutely dominate for once.
The fact some are saying AMD is "destroying" Intel in CPUs while they only got competitive again is saying a lot.

[NikosD]

There are no performance numbers for 3090 yet.
I haven't seen a single game measured with this card and it's going to be released only one week later than 3080.
nVidia is clearly hiding 3090 because the performance difference from 3080 is so small to justify the 110% price difference.
The good news is that AMD is going to release Big Navi or 6000 series probably on October 7th, just two weeks later than 3090.
And Big Navi cards are not going to be paper launch cards like Ampere, because the superior TSMC 7nm process has already produced 1 billion chips of various kinds.
Very good yields for TSMC 7nm, with great performance and efficiency.

Boycott 3090 cards.

Friends don't let friends buy 3090 cards.

Wait for Big Navi and decide with all cards on the table, but not 3090 cards.

Everything but 3090 for God's sake!

[SamuriHL]

I don't think even nVidia thinks anyone will really buy a 3090. It's more the "look at me" attention seeking card. The rumor today is that AMD was planning a 16gb card for 599 and they're now talking about bumping that down to 549. As I said before, this only means good things for consumers. I was originally looking at the 3090 because I genuinely didn't expect the 3080 to come out so strong but since the launch event I'm only looking at the 3080 at this point. Competition is what forced nVidia's hand here. They would have been quite happy to sell you a 10-20% bump over comparable 20x0 cards and raise prices while doing it, but, the fact that they absolutely killed it while keeping prices the same is a huge benefit to us. And the real world benchmarks for the 3080 are starting to come out. Not the 100% over 2080 claim that they're making, but, somewhere in the neighborhood of 75-92% in real world terms of AAA gaming titles. I mean, you gotta give that one to them. That's absolutely unreal and hasn't been seen since 2004. I'm not discounting AMD, but, I think they were likely expecting the 10-20% usual bump over the previous gen and thought they could sneak in comparable performance at a cheaper price. I doubt they have anything that's going to best a 3080 in terms of performance. Maybe I'm wrong. That'd be nice if I am.

[NikosD]

nVidia talked about almost 2x performance of 3080 vs 2080 not 2080 Ti, which is close but not 2x.
I expect around 70% on average for the 15 most popular games e.g techspot's suite.
The difference of 3080 vs 2080 Ti could be around 40% to 50%.
Big Navi could be close to 3080 even though nVidia named the 3080 Ti (GA102) released card as 3080 card.
They changed level but AMD can keep up, I think.
We'll see in a month...

[SamuriHL]

Yes I know, I edited my post as that was a typo on my part. It's still highly impressive.

I don't think AMD can shift on a dime. nVidia played their hand first. It's not like AMD was sitting on their hands this whole time. I don't know what they have planned, no one really does, but, given past history I just find it difficult to believe they're going to match a 3080. Maybe they will and that would be good. I absolutely believe they will beat the 2080ti.

[nevcairiel]

The technical embargo on Ampere expired this evening, although info was only given to the press after the event (ie. a few days ago), so expect the good tech publications to have articles out after analyzing the info, whenever they manage to get through all of that.
Benchmarks are likely still out until closer to launch, when cards are also being sampled to reviewers.

Highlights:
- The 10000+ CUDA cores are real, there is that number of FP32 capable ALUs that the scheduler can all individually use as the work load permits (128 per SM, Turing had 64 per SM, Pascal however also had 128)
- RTX IO works in the CUDA cores, and comes to Turing as well. The INT32 cores are likely a big factor here.
- RT and Tensor can now be used simultaneously in a core, which boosts tensor+RT performance, perfect for RT+DLSS rendering
- The 3080 has 96 ROPs, the 3090 has 112. For comparison, the 2080 Ti had 88. The ROPs are what pushes the final pixel to the screen and a key performance component next to the memory and cores.

I do sort of want a 3080 with more memory, but we'll see on reviews if that comes close to being a concern.

[SamuriHL]

Holy God, nVidia. I'm really impressed. Yes, the memory thing is a concern. I'd like to see 16 or 20 but oh well.

[nevcairiel]

My gaming system is on 1440p and I have no huge plans to change that, since at this screen size and viewing distance 4K does nothing but eat my FPS, so that saves me a bunch of memory. Will likely be fine. Would've just been nice.

Memory is a technical challenge at this point. They can't just offer a 3080 with 16GB or such, not without affecting everything about the card.
GDDR memory is all linked up to individual memory channel, each being 32-bit in width (technically 2x 16-bit on GDDR6, but same thing). And memory chips come in 1GB.

So 10GB? Fine. 10x 32-bit = 320-bit bus, 10 chips. All fine.
20GB? Sure. Do the same thing, connect two chips per channel.
16GB? Uhh. 8 channel with 2 chips each? But that reduces bandwidth. Full 16 channel? Way too expensive to build on the board.

You also cannot put more chips on some channel only, then you get the 970 situation where memory is off-balance and some parts are slower.

11 or 12 GB might've been possible by using the full memory controller loadout that the 3090 has (12 x 32 with 2GB per channel), but they likely needed to cut there to get the chip yield up.

20GB is probably too much, while 10GB might feel slightly constrained. But the nature of GDDR memory controllers dictates this, unfortunately.

Not sure if AMD is once again going to use HBM, which has similar constraints except that it comes in multiples of 4 or 8, so they can do 16GB. But HBM is also very expensive.
Or they could go for lower bandwidth and more capacity (ie. the 8 channel approach) with GDDR6.

[SamuriHL]

I game at 4K but I still don't think it'll be a problem.

[huhn]

- RTX IO works in the CUDA cores, and comes to Turing as well. The INT32 cores are likely a big factor here.

good point but doesn't nvidia have fixed function for decompression and compression similar to the lossless memory compression?

[NikosD]

The first game ever that used CUDA cores do decompress textures, was RAGE on 2011.

With RTX IO, nVidia looks cheap again by excluding Turing GTX cards.

[ryrynz]

Turing will support RTX IO (https://www.nvidia.com/en-us/geforce/news/rtx-io-gpu-accelerated-storage-technology/).

[huhn]

let's see if AMD will look cheap with navi.

[NikosD]

nVidia at the first comment:
RTX IO is supported on all GeForce RTX Turing and NVIDIA Ampere-architecture GPUs

[chros]

Boycott 3090 cards.
Friends don't let friends buy 3090 cards.
Everything but 3090 for God's sake!
:D

I game at 4K but I still don't think it'll be a problem.
I think 10GB is just on the limit :(. 1080Ti had 11 and Titan X(p) 12.

Not the 100% over 2080 claim that they're making
:D I just watched again Jim's (AdoredTV) 2 years old nVidia turing marketing survival guide (https://www.youtube.com/watch?v=b_x1JGG4JC8) video in 25 mins when he only talks about 1 (!) slide :D (he talks about memory bandwidth) It's hilarious to see how they operate.

They would have been quite happy to sell you a 10-20% bump over comparable 20x0 cards and raise prices while doing it, but, the fact that they absolutely killed it while keeping prices the same is a huge benefit to us.
...
I mean, you gotta give that one to them. That's absolutely unreal and hasn't been seen since 2004.
Yes, some already said that it's the first generation for a long time (since Fermi?) when a xx80 class GPU gets the 102 die! (with Turing/Pascal it had 104 die). That means almost the full die in a generation! It's basically a Ti class card in the previous generations. :)
That's why the claimed huge performance uplift compared to the previous generations!
I'm pretty sure they were/are scared of something :) (consoles/amd/etc ?)

And that's one of the reasons why the power usage is in the clouds with the 3080!
Along with Samsung 8nm, that Jim state in his new Ampere analysis (https://www.youtube.com/watch?v=1milKOPa9zA&t=2s) that it's only around a half-node shrink compared to TSMC's 12nm!

One more thing to note: it can happen (due to how nvidia treats/treated their partners in the present/past) that they will be stuck with Samsung for a long time! And this can be a real issue in the long run (2-4 years later).
They won't be able to do this very same trick 2 generations later, because they already gave the full die ... :)

- The 10000+ CUDA cores are real, there is that number of FP32 capable ALUs that the scheduler can all individually use as the work load permits (128 per SM, Turing had 64 per SM, Pascal however also had 128)
Interesting about Pascal, thanks.

- RTX IO works in the CUDA cores, and comes to Turing as well. The INT32 cores are likely a big factor here.
What do you think, can this be interesting for us, playing back (high bandwidth) videos or not?


let's see if AMD will look cheap with navi.
:) It will be interesting to see what will happen, there are controversial "insider" information, some state it's nowhere near, some state it'll be competitive. I hope the latter, it's good for the industry and us in the long run.

[nevcairiel]

What do you think, can this be interesting for us, playing back (high bandwidth) videos or not?


Its doubtful that its useful for video. Afterall video decoding is already on the GPU, with near to no CPU load, and the bandwidth requirements arent going to massively increase. If anything video bitrates are reducing, or at best stagnating.

[chros]

Thanks.

Yes, some already said that it's the first generation for a long time (since Fermi?) when a xx80 class GPU gets the 102 die! (with Turing/Pascal it had 104 die). That means almost the full die in a generation! It's basically a Ti class card in the previous generations. :)
That's why the claimed huge performance uplift compared to the previous generations!
And if I look at it this way, that means the 3070 is a xx80 class card and hopefully the 3060 will be a xx70 class card for (hopefully) $300-350! The latter will be a real deal for lot of users with hdmi 2.1 and av1 decoder ...

One more thing to note: it can happen (due to how nvidia treats/treated their partners in the present/past) that they will be stuck with Samsung for a long time! And this can be a real issue in the long run (2-4 years later).
They won't be able to do this very same trick 2 generations later, because they already gave the full die ... :)
And if I think of this more, I'm pretty sure the next gen nvidia GPUs will "only" bring the Turing type of increase tops, nothing like we see now.
So, unless AMD's RDNA3 (not 2) will bring massive change, these 3000 cards will be a good value for a long-long time.
Maybe I should also buy a 3060 (although I said I won't :) ), but let's wait for the specs and real benchmarks first, and hopefully some madvr user reports as well ... :)

[NikosD]

NVLink SLI is reserved for 3090 card only.

But why ?

Because people shouldn't have the chance to combine 2x3080 cards with the price of a single 3090 and get ~80% more performance than 3090.

I think a heavily overclocked custom 3080 card could overcome 3090's performance.

[QBhd]

...

Why? Because SLI/CF is totally dead from a gamer perspective... More problems and headaches and games don't really leverage the ability. For compute workloads, yes. Hence, the new "Titan" (3090) has it.

QB

[ryrynz]

get ~80% more performance than 3090.


Two 2080s in SLI couldn't even beat a 2080Ti on average and most of the time managed only somewhere in between a normal 2080 and a Ti.
3080s in SLI would be a serious waste of cash and the above reply from QBhd factors in as well. I think from a logical standpoint this decision makes sense, SLI is basically dead in it's current form (with gamers)

[NikosD]

3080 has the same chip like 3090.

2x3080 combined would be a killer combination for the use of NVLink SLI

20GB of 2x3080 VRAM vs 24GB of 3090 VRAM and 80% more performance of 2x3080 vs 3090 for the same price.

For the use of NVLink SLI.

[butterw2]

hopefully the 3060 will be a xx70 class card for (hopefully) $300-350! The latter will be a real deal for lot of users with hdmi 2.1 and av1 decoder ...


It seems these are the 2 only new features of rtx 3000 (vs rtx 2000). Maybe a used rtx 2060 Super 8GB can be had at a more reasonable price once the 3060 and 3050 and navi2 are out ?

I bought my current PC new, monitor/SSD-HDD included, for less than the starting price of the "cheap" RTX3070. It also uses less power than the RTX3080. So yes it has less TFlops, but it is also silent and has no driver issues (don't update).

I can't help noticing Nvidia does a form of planned obsolescence with the GDDR6 quantity on their cards. Rtx 2060 only had 6GB VRAM. While 8/10GB is enough right now (for 4K gaming), what will the situation be in 2 years time, given that the next-gen consoles come with 16GB ?

More relevantly how much VRAM do you actually need on the gpu to be able to make real use of AI features ? Will MS DirectStorage (RTX-IO) be applicable to Machine learning to overcome VRAM limitations ?

PCIe4, M.2 NVMe SSD, zen3 compatibility look like useful features for future upgrades.

[NikosD]

For those who think with an eye on the next-gen games at 4K resolution, a fast enough Big Navi with 16GB VRAM will be a better choice than 3080 at 10 GB VRAM.

If it manages to be competitive though.
Which I think it will.

[chros]

For those who think with an eye on the next-gen games at 4K resolution, a fast enough Big Navi with 16GB VRAM will be a better choice than 3080 at 10 GB VRAM.
Like last year, right?! :D We have to wait for it to see ...

It seems these are the 2 only new features of rtx 3000 (vs rtx 2000). Maybe a used rtx 2060 Super 8GB can be had at a more reasonable price once the 3060 and 3050 and navi2 are out ?
Well, it depends for what. If you want to fool around (coding) with the RT cores, then probably yes, because you can't play games with it :)
But even then if you can get (let's say) with $100 plus a truly useable card with the aforementioned features then I'd go for it instead. (If you don't need the RT cores at all then just go for a used Pascal, they will be dirty cheap :) )
I hope nvidia won't dare to come out with a 6GB 3060 :D

I can't help noticing Nvidia does a form of planned obsolescence with the GDDR6 quantity on their cards.
I'm also pretty sure that's the case! :) As others said, there will higher versions as well but for higher price (maybe called Ti/Super/etc.) :)
And they have to think of the next gen GPUs as well, otherwise they won't have any card left in their hands :)

[NikosD]

No Founders Edition RTX 3000 series card will be sold in Australia.

Aussie buyers have to wait for custom cards from OEMs

Don't ask me why.

Ask the leather-jacket-man instead.

Happy days of new nVidia products!

[butterw2]

hopefully the 3060 will be a xx70 class card for (hopefully) $300-350! The latter will be a real deal for lot of users with hdmi 2.1 and av1 decoder ...

A 300$ card from Nvidia ? Maybe a custom rtx 3050, with 6GB GDDR6 ? Can do raytracing at 1080p with DLSS enabled in supported titles such as fortnite. :D

[chros]

:) Yep, $300 3060 8GB is a bit unlikely, but who knows? If 3080 $700 and 3070 $500 ...
I think it will all depend on AMD (this time again).

[ryrynz]

I can't help noticing Nvidia does a form of planned obsolescence with the GDDR6 quantity on their cards

The way you talk about it is like the performance absolutely tanks or the card just doesn't render certain elements or even stops working if the required texture memory is higher than the card has onboard, none of these situations is the case.
At most you'd lose maybe ~20% or less FPS if you had a game needing say 16GB or more texture memory, it may actually ending up being less on these new cards with RTX IO.
Anyway, you can always just lower the texture quality down a notch of two and you're golden. It's up the the devs how much memory they want to allow a program to use and up to you to set the value appropriately.
Games are made for today's hardware not tomorrows, purchase what suits your needs, Less out of the box performance but with more RAM? Looks like it's AMD for you :P
Remember the market is designed to make you want to buy new hardware and companies are rewarded for selling it to you, you can't expect a capitalist market to operate any differently than that.

I do feel that 10GB on the 3080 seems a tad on the low side for a card of this calibre, however Tensor Memory Compression is going to reduce texture memory requirements upwards of 20%
so there's that too. There will be plenty of reports of how this holds up to the most demanding games later on this or early next year, it'll will probably do better than you think.

[NikosD]

After the announcement of a real rocket from Sabrent using Phison E18 controller for the PCIe v4.0 NVMe SSD, which eats Samsung 980 Pro for breakfast, the last obstacle for this year's upgrade has been lifted.

I'm ready to move to PCIe v4.0 platform for both Graphics Card (less impact) and SSD (could be more useful with suitable software)

I was hoping for a battle in the CPU arena between Rocket Lake and Zen 3, but unfortunately Rocket Lake moved once again to Q1 2021.

The battle on GPU arena will be intensive between Big Navi and Ampere cards and we expect it around October.

All reports say that Navi21 will be faster than 3070 and probably competitive with 3080 with more VRAM and better price as always.

For my new rig and my budget, I have to wait for 3050 Ampere card (if ever is going to be released) and the RDNA2 card of the same budget around 250$

I'm very pleased with 1080p gaming and I need nothing more.

The only case that I would wait for Rocket Lake or buy an Ampere card would be, if RDNA2 doesn't eventually support AV1.

Then probably one of my two SoCs (CPU or GPU) should support AV1 so it's either iGPU Xe inside Rocket Lake or Ampere 3050.

If RDNA2 with Big Navi supports AV1, I will probably go ALL AMD after many many years since AthlonXP and ATi Radeon 9500 -which wasn't AMD(!) actually.

Around December close to the holidays season, a new rig could be a nice gift!

[NikosD]

We had also another release this week, which is actually a non release but rather an announcement.

But I will leave Charlie speak with the voice of reasoning:
More to the point, Intel can’t actually supply Tiger Lake CPUs. SemiAccurate’s checks with OEMs all say they Intel is shorting them on supply by 7-digit numbers in several cases.

If you are a glass half full person, you could say demand is high.
If you are realistic, you can say that yields are still not viable on 10nm, and wafer starts are… the number is shockingly small.

Because of this, we'll talk about DJs, how often people re-encode and run AI over their entire photo collection, and more ‘realistic’ use cases!

Just don’t ask about ship dates, prices, yields, or the important bits, and really don’t ask about the specs on the ‘tests’ they presented.

[Magik Mark]

Thank You, guys, CRU 1.4.1 with DisplayID also solved my above mentioned issue. That's what I did:
- wrote down the madvr timings, removed the madvr created custom resolution, then restart
- in CRU 1.4.1
-- add DisplayID to Extension blocks and set it as the 1st entry
-- set madvr timings and set Pixel clock and Native as well

It's a big help to get rid of a major annoyance, now I can set 23p as the default refresh rate of the TV :)

Couple of notes:

The difference between CRU and nvidia created (by madvr) custom timing is:
- nvidia adds a new one to the already existing ones: that can confuse applications which one to use (e.g. madvr)
- CRU replaces an existing one with the custom one on the OS level, so there won't be any confusion

When you create custom timings with madvr (http://madvr.com/crt/CustomResTutorial.html):
- always start from the default one (EDID) and pay special attention to the "sync" attributes (e.g. + , +):
-- don't select a mode that is different!
- don't deviate a lot from the default EDID values
-- remember, this approach is modifying "back porch" (horizontal and vertical values) only!
- the above 2 wrong settings can easily modify the response of the panel! (e.g. modifying gamma curve, causing black crush, etc.)
- if you get >= 10 hours after the first optimisation then it's already really good, you don't need to do more runs

Examples, original EDID timing:

EDID/CTA:
front sywi back pixels sync
hor 1276 88 296 1660 3480 5500 +
ver 8 10 72 90 2160 2250 +
picl 296.70 mhz
23.9757575757576
OSD result: 23.97791Hz (~4.41 minutes)


Result after the first optimisation run:

Custom:
front sywi back pixels sync
hor 1276 88 356 1720 3480 5560 +
ver 8 10 72 90 2160 2250 +
picl 299.94 mhz
23.9760191846523
OSD result: 23.97537Hz (~22 hours - 1 day)



About nvidia driver versions:
- some of them can block custom timings completely (doesn't matter how they were created)
- so always check in madvr OSD whether your timing is applied if you install a new driver

Btw, I still use the Display Changer II (https://forum.doom9.org/showthread.php?p=1863306#post1863306) util to easily manage 2 displays.

Have you tried using 47.952 instead of the usual 23.976. It's basically doubling the latter.

So far no success for me

[chros]

No success with which? :) I haven't tried 48p because LG TV has 23p but not 48p.

[NikosD]

A quick recap of Ampere RTX 30 series launch before the first real benchmarks on September 17th:

As I have told you from the beginning, nVidia decided to go nuke, giving the largest die of GA102 at both 3080/3090 chips.
It also went nuke by releasing both cards at >300 Watts for desktop gaming which is a record high for nVidia previously reserved only for two chips combined (like GTX 690)

The leather-jacket-man with a Duke Nukem EGO, didn't manage to get TSMC 7nm and decided to go with Samsung at 8nm.
The comparison of TSMC 7nm vs Samsung 8nm for Ampere using GA100 vs GA102 gives a 47% higher density for TSMC and better efficiency.

ASUS announced a RTX 30 series card with LEDs, flashing after a system crash indicating that you have to change PSU in order to feed properly the power hungry card.

nVidia fan boys used to blame ATi/AMD when it raised power consumption in order to be competitive with nVidia.

Now we see the opposite, but noone complains.

As I have already mentioned, Ampere's TFLOPS are "fake" meaning they advertise a maximum theoretical value not reachable in real-world.
So the first calculations give 1TF of Ampere to be around 2/3 TFLOPS of Turing aka 3080 performance of 30TFLOPS is approximately around 20TFLOPS of Turing.

That's why even nVidia says that 3080 card with 30TFLOPS is up to 2x faster than 2080 with just 10TFLOPS and not 3x.

In real-world games, in average of at least ten modern games, I expect 3080 to be around 70-80 percent faster than 2080.

Judging by MS Flight Simulator 2020 and recently released Marvel game, the 8GB VRAM of 3070 and even 10GB VRAM of 3080 could be a problem for 4K max settings even from now, not in the future.

The good thing about this release is that I didn't read anywhere the embarrassing and disgusting "Just buy it!" from Tom's Hardware two years ago with the release of Turing RTX 20 series.

Digital Foundry was paid by nVidia for a biased presentation of 3080 vs 2080 benchmarks, but this is minor.

And what about the competition ?

The Big Navi21 with 80CUs could be around 20+ TFLOPS of RDNA2 architecture which is closer and probably a little faster than Turing's TFLOPS so it could compete with 3080.

All AMD's Big Navi cards could potentially have lower power consumption than nVidia's Ampere cards for all tiers.

Also, it seems that Big Navi21 could have 20GB of VRAM competing with 3080 and Navi22 16GB of VRAM competing with 3070 card.

So, lower TDP and double VRAM for AMD with close to target performance.

But a very bad thing appears more and more in rumors.

Due to the large die of Big Navi21, they may want to reserve the quantities of TSMC's 7nm for more profitable and smaller console and CPU (Zen 2/ Zen 3) chips than GPUs that have little margins.

VERY BAD THING for GPU competition in the high segment.

We'll see...

[Magik Mark]

No success with which? :) I haven't tried 48p because LG TV has 23p but not 48p.Using CRU to make custom refresh rates in LG Oled. Is it possible for you to make experiment on this? 48fps

Sent from my SM-G988B using Tapatalk

[chros]

Using CRU to make custom refresh rates in LG Oled. Is it possible for you to make experiment on this? 48fps
The problem is, as the guide suggests, that we should start from the EDID of a given mode and only change the "back porch", but we don't have any initial values for 47p.
But why don't you use 23p on the C9?

As I have told you from the beginning, nVidia decided to go nuke, giving the largest die of GA102 at both 3080/3090 chips.
It also went nuke by releasing both cards at >300 Watts for desktop gaming which is a record high for nVidia previously reserved only for two chips combined (like GTX 690)
Good summary.

nVidia fan boys used to blame ATi/AMD when it raised power consumption in order to be competitive with nVidia.
True.

Now we see the opposite, but noone complains.
I do, not that it matters much :)

So, lower TDP and double VRAM for AMD with close to target performance.
That's what I'm interested in, but let's wait to see it happen :)

VERY BAD THING for GPU competition in the high segment.
I don't think it's bad, for me power efficiency is way more important. The only question is whether they can achieve it.

And what about DLSS (marketing BS) regarding to gaming? Am I the only one who feels like cheating? :)
I mean the need of DLSS clearly demonstrates that even these GPUs can't deliver native 4k content with reasonable fps. And they started to talk about 8k :D

Lastly, pricing: lot of folks are happy that nvidia haven't raised prices but a 3070 for $500 is already really high. They basically kept the Turing prices (previously a xx70 class GPU did cost around $400).

[huhn]

xx70 is just a number.
the problem with turing and the reason for the low sales number was you got the same performance for your money making them a terrible upgrade.

the ampare GPU are suppose to bring much more performance on the table for the same money if it is called 3070 for a 1000€ or 500€ doesn't matter what matters is that there is more performance for your money smaller cards will follow soon enough.

the problem with amd is not just the power number it's the performance per watt. that doesn't mean there is a problem if they have a card with 300w+ in there hands as long as there are smaller one with a reasonable power usage are following.

maybe ignorance reaches a new high here but NAVI is already 7nm and it can't beat turing in performance per watt. it is impossible to fix that clearly not but it doesn't change current facts.

[SamuriHL]

Couple that with the fact that we don't yet know (at least I don't anyway) what they plan on putting in for encoder/decoder support. And that matters. A lot. At least to this crowd. Traditionally they've fallen short on the encoder/decoder side of the equation. Couple that with Huhn's accurate portrayal of their performance per watt, uh, challenges....and it doesn't usually spell good things. I still maintain they're going to fall short of the 3080's performance. It might be "close enough" to make the case for it. Especially if they crank up the memory, but, will it be GDDR6X or the equivalent? At the price points they're going to have to compete at, I have a hard time believing they're using a 7nm process AND fast memory. I bet they sacrifice memory speed to give you gobs of it. 16gb at near 3080 performance, right? Let's hope that's not closer to 3070 performance, although it could be. If they could match or beat the 3080 performance with 16gb of ram they'd not be talking about cutting prices to 550....just above the 3070's price. What's that say to you about where we can expect the card to fall? All conjecture at this point. Nothing concrete until we see it from the electric bike's mouth, as it were. ;) :D

[ryrynz]

Calling it, Nvidia has the crown for a long time yet. 16GB at near 3080 performance? DOUBT. Competitive prices? Yes @ 3070 levels and below. Nvidia is not worried in the slightest about Big Navi.
This gen is a big deal for various reasons, will be a long time till you see anything like this again, new memory standard required.


VERY BAD THING for GPU competition in the high segment.


Overall prices are great, nobody is really complaining about these. Sure 3090 price is high but we're talking top of the line fastest GPU in the world, they can milk it.

Nvidia has competition, they've priced this aggressively to beat the competition.. This is a great release all round and still people complain about various minor things as always, so tiring.

[SamuriHL]

I mean I tend to agree with the assertions that they'll be slightly faster than a 3070 but not near the 3080. Which is why, IMO, the leaked price is probably accurate. But we have nothing concrete so it IS just conjecture.

[NikosD]

A few things to correct, because ignorance reached the sky and is still going up.

So, before ignorance reaches the moon I have to say:

1) GDDR6X is a custom-made memory produced by Micron only and while we don't know what will happen in the future, the only customer for now is nVidia.
So Big Navi is going to have GDDR6 or if AMD surprises once more will have HBM2.

2) There is memory faster than GDDR6X and it's called HBM2 and HBM2e from Samsung, SK Hynix and AMD

3) NAVI was based on RDNA and according to Techspot the NAVI10 (5700 XT) chip of 10.3 billion transistors with a TDP of 225 Watts has the same performance with the TU104 (2070 Super) of 13.6 billion transistors with a TDP of 215 Watts

So, extremely close I could say essentially the same performance per watt for RDNA and Turing.

4) AMD has promised 50% better performance per watt of RDNA2 versus RDNA1

5) Big Navi is based on RDNA2 architecture.

6) Do the math.

The completely ingnorant and insignificant troll who is comparimg TFLOPS with game engine cheating (Oh Jesus!) should stop trolling below my comments and read a little technical stuff before exposing himself so badly.

Just my two cents...

[NikosD]

Nvidia is not worried in the slightest about Big Navi. and Nvidia has competition, they've priced this aggressively to beat the competition This is the definition of contradictory opinions.
You can't say these two phrases in the same post.
One of them is wrong and obviously it's the first one.
Sure 3090 price is high but we're talking top of the line fastest GPU in the world, they can milk it. This was and still is one of the biggest problems regarding customer behavior against Intel and nVidia compared to AMD.

Just look at the price of the fastest CPU in the world of servers - the EPYC with 64 cores - the fastest CPU in the world of High End Desktop - the Thread ripper with 64 cores and zero competition - and the fastest CPU in the world of desktop - the Ryzen 3950X with 16 cores.

Who is milking and what is the reaction when AMD tries to put the price a little higher ?
The war of trolls.
But for Intel and nVidia is just Wednesday.

Be more responsible customers not fan boys and support the competition.

[ryrynz]

and This is the definition of contradictory opinions.


No. These are priced attractively to sell now, that's the point. AMD will always be a competitor, it's to tempt people to upgrade now and upgrade they will, just look at all the cards going up for sale now on the second hand markets. AMD will lose some GPU market share likely this year and next as a result, it's a great release for Nvidia and sales will surge.

[NikosD]

No. These are priced attractively to sell now, that's the point. AMD will always be a competitor, it's to tempt people to upgrade now. AMD will lose some GPU market share likely this year and next as a result. nVidia is afraid of Big Navi and tried to make AMD believe that is not worth it to fight in the high-end segment since Big Navi of 80CUs will be a big die eating a lot of TSMCs wafers of AMD and with the aggressive pricing of 3080 and 3070 it needs to cut the margins a little more.
I said earlier that if AMD is going that way, it will be bad for customers leaving the monopoly of high-end segment to nVidia.
AMD has to compete with everything it has in its hands, releasing Biggest Navi of 80CUs and we have to support it.
If it does this, it will certainly INCREASE its GPU market share as there will be competitive products in the high-segment for the first time after many years.
If AMD doesn't release a high-end product after the extremely aggressive move of nVidia releasing a 320 Watt GA102 die product calling it just "3080" in the same price of the poor 2080, then it will certainly lose GPU market share and will support unintentionally the monopoly of nVidia in this segment.
I hope the margins to be OK for AMD and release the Biggest Navi of 80CUs and customers support it.

[NikosD]

And after writing the above post, I feel obliged to ask you to keep a moment of silence for those who bought Turing RTX 20 series cards, especially the 2080 Ti owners probably while reading the "Just buy it!" article of Tom's Hardware and after the extraordinary fear of nVidia because of Big Navi which made it release Ampere this way.

RIP brothers & sisters, owners of RTX 20 cards.
May the ebay be with you.

Also, I have to ask you to pray with me for all those nVIDIOTS who are going to buy 3090 cards for gaming.
They are the future RTX 20 owners with nVidia's next GPU release, but they don't know it yet -although they should.

Pray for them until the next RIP in around two years...

[chros]

And after writing the above post, I feel obliged to ask you to keep a moment of silence for those who bought Turing RTX 20 series cards
...
RIP brothers & sisters, owners of RTX 20 cards. May the ebay be with you.
:D That's a good one :) Let's be frank, those cards were crap without having competition.

Btw, M$ just confirmed the price of the new small Xbox: $300.
I hope that's a good news for the small cards in this generation. :)

Pray for them until the next RIP in around two years...
I'm sure that there won't be anything big in next years, see these posts why: link (https://forum.doom9.org/showthread.php?p=1922785#post1922785) and link (https://forum.doom9.org/showthread.php?p=1922791#post1922791)

3) NAVI was based on RDNA and according to Techspot the NAVI10 (5700 XT) chip of 10.3 billion transistors with a TDP of 225 Watts has the same performance with the TU104 (2070 Super) of 13.6 billion transistors with a TDP of 215 Watts
Huhn has a point, those were on 7nm already, that was not impressive at all. Let's see this year :)

with the aggressive pricing of 3080 and 3070
Guys, don't call this aggressive, it's still freaking high :)

xx70 is just a number.
That's where you are wrong: it's *not just* a number.
It's a price bracket and a performance bracket as well in a given generation!

if it is called 3070 for a 1000€ or 500€ doesn't matter
It does matter and matters a lot! See the 2 links above.

[huhn]

what link?

and no matter how you turn it it's just a name it does not mean that with a xx70 you will get a payable but not the best card if you interpret it as that then that's your own problem.
they can go and name then as they want and they can price them as they want as long as you get more performance per dollar you are not robbed.
if they sell you a product for half the price with the same performance as the last flag ship and massively cheaper in price and performance as the concurrence then this is fine no matter how it is named.
if you want a card for 350-450 wait for one but as long as this card also has the same performance increases per dollar is totally fine.

[NikosD]

Troll has a point, those were on 7nm already, that was not impressive at all. Let's see this year :)
In the name of Mighty Thor and the Elder Gods, how much power efficiency gain do you think AMD pulled out from TSMC 7nm vs TSMC 12nm of Turing ?
AMD targets clocks and performance mainly for its GPUs and power efficiency later.
You can't have both performance and power efficiency.
You target performance or power efficiency using a specific TSMC node.
They needed a small and fast die with IPC and power efficiency comparable to Turing and they did it.
We don't compare performance per area here, where the NAVI10 using TSMC 7nm simply destroys Turing with TSMC 12nm.
It's a TDP of 225 Watts for NAVI10 using fewer transistors vs TDP of 215 Watts for Turing that really matters.

[huhn]

you are comparing a 12 nm chip which actually is 16nm with 7 nm and you are not winning in clock rate and in power consumption.
the transistor number are 10800 to 10300.
does NAVI have dedicated INT32, RT and tensor cores?

just to hammer this in while a huge part of the turing cards has "wasted" or not often used transistors NAVI could still not win with a massively better process process.

and i have a hard time ever believing 7nm 251mm² chip is cheaper then a 12nm 445mm²...

[chros]

huhn is right in the above.

and no matter how you turn it it's just a name it does not mean that with a xx70 you will get a payable but not the best card if you interpret it as that then that's your own problem.
It's not my problem :) You don't understand what I'm telling. I don't mind that.

[NikosD]

As I wrote above TU104 of 2070 super has 13.6 billion transistors not 10.8 vs 10.3 billion of NAVI10 of 5700 XT.

If I have to repeat my posts two times in order to understand it correctly, it's a huge waste of time for me.

@ryrynz for example didn't read my post that RTX IO isn't supported by Turing GTX and he replied that it is supported by Turing RTX which is obviously a different card.

Read my posts two and three times before you reply.

[SamuriHL]

Especially if they crank up the memory, but, will it be GDDR6X or the equivalent? At the price points they're going to have to compete at, I have a hard time believing they're using a 7nm process AND fast memory. I bet they sacrifice memory speed to give you gobs of it.

A few things to correct, because ignorance reached the sky and is still going up.

So, before ignorance reaches the moon I have to say:

1) GDDR6X is a custom-made memory produced by Micron only and while we don't know what will happen in the future, the only customer for now is nVidia.
So Big Navi is going to have GDDR6 or if AMD surprises once more will have HBM2.

2) There is memory faster than GDDR6X and it's called HBM2 and HBM2e from Samsung, SK Hynix and AMD



Read my posts two and three times before you reply.

Yes, reading skills definitely have fallen by the wayside here. :) There's a reason I put "or the equivalent" in my sentence that you appear to be calling ignorant. We don't know what memory AMD will be utilizing. I VERY much doubt that it's going to be HBM2 at the price points that are rumored. It's likely going to be GDDR6, but this is pure speculation at this point. And I maintain what I said...to compete with the 3080 it will need faster memory, not just gobs of slower memory. I have no doubt AMD is going to best the 3070 with their top end Big Navi card. But I have SERIOUS doubts about it reaching the 3080. That's just my opinion until we see some comparisons in the real world for both cards instead of paper specs, for which we don't even have on Big Navi yet. Everything right now is mere speculation and conjecture.

[huhn]

a 5700 Xt is about as fast as a 2070 not a 2070 S.

[NikosD]

Zen3 on October 8th and Radeon RX 6000 on October 28th.
It's official from AMD:
https://videocardz.com/newz/amd-to-introduce-zen3-on-october-8-radeon-rx-6000-series-on-october-28

I can see that rumors will be huge till end of October for RX 6000.

[nevcairiel]

I hope Zen3 actually launches in October as well, and not very much later. Been eagerly awaiting those to rebuild my setup from ground up.

[chros]

And nvidia will lunch the 16GB and 20GB versions of their 3070/3080 in the middle of October as well, then 2 weeks later the 3060.
Actually this is a really clever move: they can sell the higher memory configs more ($100+ ?). Of course they could have give it away for "free" :D
And they can say: "we have a $500 card that is really good, if the memory isn't enough for you then buy the higher models" :)

[huhn]

yeah not going to get a new GPU before i get a new CPU. i really need zen 3.

[videoh]

RIP brothers & sisters, owners of RTX 20 cards. My 2080 Ti is not going to stop doing all the awesome things it does (and has done) for me just because a 30xx is released. Worst case, I can move it to replace the 1080 Ti in my backup machine (which will in turn replace the 1050 Ti in my second backup machine). But more likely I'll put my 3090 Ti in a new kick-ass machine with latest CPU goodness and PCIe 4. I've been benefiting from my 2080 Ti for almost two years so there's no reason for any regrets. Long live nVidia!

All this guy has is rhetoric, insults, forlorn hopes, and sour grapes. I will pray for him.

[NikosD]

All of you guys are a bunch of clueless nVidia fan boys, irresponsible customers and have real issues with facts and the real-world.

I see no real and solid arguments against mine.

All I see is the reproduction of the mainstream propaganda of nvidia's marketing BS.

You are living in a bubble and you are responsible for the monopoly of Intel and nVidia with all of the consequences of these monopolies - high prices and milking.

I feel pity.

[SamuriHL]

I hope Zen3 actually launches in October as well, and not very much later. Been eagerly awaiting those to rebuild my setup from ground up.

Traditionally I've stuck with Intel and do like my 9900k. However, they have definitely fallen behind and I no longer even remotely care about SGX and "official" UHD playback anymore. (I was only moderately interested before when I got the 9900k but now not even a little bit). So AMD is an option for me and I really want to see how balls out these Zen3 procs are. PCIe 4 nvme. Mmmmmmm :)

[huhn]

All of you guys are a bunch of clueless nVidia fan boys, irresponsible customers and have real issues with facts and the real-world.

I see no real and solid arguments against mine.

All I see is the reproduction of the mainstream propaganda of nvidia's marketing BS.

You are living in a bubble and you are responsible for the monopoly of Intel and nVidia with all of the consequences of these monopolies - high prices and milking.

I feel pity.
the unbeatable i'm right you are wrong argument i'm disarmed here.

from the guy with an intel CPU btw.

[SamuriHL]

Latest leak/rumor shows big navi with 16gb of GDDR6 at 256 bit bus. It's entirely possible this is not finalized, but, if it is, that hardly is going to beat a 3080.

[chros]

Do you have a link?

[NikosD]

Latest leak/rumor shows big navi with 16gb of GDDR6 at 256 bit bus. NO NO.
I have written before a few times that Big Navi21 of 80CUs should have 20GB of VRAM and larger bus.
The 16GB VRAM Navi should be the one competing with 3070.
If this that I describe is not the case, then AMD has lost this battle by far.

[ryrynz]

You are living in a bubble and you are responsible for the monopoly of Intel and nVidia with all of the consequences of these monopolies - high prices and milking.


Nvidia is going hard and delivering good prices too for what's being developed, Intel milked quad cores for far too long and didn't innovate quickly enough, they're paying the price now and so will anyone that languishes too long.

Competition is always there waiting in the wings. Like I said, nobody is complaining about how things are priced and what we're getting in any hardware market.

Can't have a fairy tale capitalist market, it's impossible. There are hardware limits, and you can't know the entire costs of running a large scale business.
The market is far from the doom and gloom you seem to view it as.


If this that I describe is not the case, then AMD has lost this battle by far.

The battle is far from lost, I'm sure they'll have a reasonably successful launch, now that I know that news of Big Navi will come very shortly, perhaps AMD will not lose market share as I expected but it will still be a tough battle.

This launch may be where things begin to turn around for them depending on the benchmark results and internet feedback.
We do want them to do well, the sentiment towards them is getting much better now they are delivering what the market wants, I hold hope.

[NikosD]

Competition is always there waiting in the wings. Like I said, nobody is complaining about how things are priced and what we're getting in any hardware market. NO.
Competition in the x86 CPU arena stopped in 2006 with the release of Core 2 Duo CPU and the dramatic mistakes of AMD.
x86 CPU arena like GPU arena are both near to monopoly situations or pseudo-duopoly if you prefer.
Intel has allowed to exist extremely few licenses and only AMD survived from this situation.
AMD brought up competition again, after 11 YEARS in 2017 with Zen architecture.
From 2006 to 2017 there was ZERO COMPETITION due to the pseudo-duopoly or covered monopoly of a Goliath that is Intel and David that is AMD.
No other real competitors in x86 CPU arena. ZERO.
This is a situation caused by anti-competitive practices of Intel for decades.
Take a look at the ARM architecture CPU arena.
Apple, Samsung,Qualcomm, Huawei (before Trump), MediaTek and Marvel, Ampere, Nuovo and so many other options for ARM.
THIS IS HEALTHY COMPETITION.
x86 "competition" is a disease, an illness, a disgusting, manipulating thing of Intel and its fanboys supporting this.
If AMD didn't bring Zen architecture after 11 YEARS, NOBODY of you would complain about the disease of Intel's monopoly and domination.
A lot of you are very unhappy with the situation and you cry like babies during the middle of the night, because Intel has become AMD's bitch since 2017.
You are Intel's slaves...You belong to Intel due to a Stockholm syndrome probably.
Take a look at this forum posts, not only this thread and you will find many of Intel's slaves.
The same situation is for the GPU arena.
In the beginning there were many GPU companies but the leather-jacket-man did many and very bad things and like a miracle nVidia managed to have another covered monopoly or pseudo-duopoly of AMD and nVidia this time.
You know, like AMD and Intel in x86 CPU arena.
AMD is too small to compete simultaneously with these two large companies.
Actually is too small to compete separately with each one, so to fight two wars at the same time is impossible.
The situation is ideal for both covered monopolies of Intel and nVidia to fight a little, tiny company like AMD.
They can legally say that they are not monopolies, because look there is AMD, but in the real-world we all know that Intel had 99% of the server CPU market before 2017 and nVidia still has 80% market of the discrete GPUs
If these are the numbers of YOUR COMPETITION, then you have clearly an issue of understanding the reality.
The battle is far from lost, I'm sure they'll have a reasonably successful launch, now that I know that news of Big Navi will come very shortly, perhaps AMD will not lose market share as I expected but it will still be a tough battle. NO NO.
If AMD releases the Biggest Navi, the highest and fastest of its RX 6000 products at the 3070 level of performance, then we will have a RDNA aka NAVI situation again.
I will say this one more time:
If the fastest product of Big Navi is at the 3070 level, then this means that RDNA2 (RX 6000) will compete with Ampere, like RDNA (RX 5000) competes with Turing aka 80% of discrete cards to nVidia.
WE DESPERATELY NEED the release of Big Navi of 80CUs that I personally believe can compete with 3080 and Big Navi using 20GB of VRAM with lower power consumption vs 10 GB of VRAM and higher power consumption of 3080 and this could bring some GPU market to AMD along with the lower tier models.
AMD and we as customers we can't afford AMD to bring its fastest RX 6000 model competing with the lowest of Ampere high-end cards.
THERE IS NO COMPETITION in a situation like this.

[huhn]

there is no need to have the biggest card they have barely any market share:
https://store.steampowered.com/hwsurvey/Steam-Hardware-Software-Survey-Welcome-to-Steam

it's pretty much look how great the 2080 ti is i get a 2060.

[NikosD]

I' ve just read a recent statement from the co-founder of Arm company, Hermann Hauser, to NS Tech regarding the possibility of nVidia buying Arm.

He said:
They are the semiconductor company that can buy Arm to destroy it—and it is very much in its interest to destroy Arm because they [would] gain a lot more than the $40 billion that they would pay for it
This statement pretty much summarizes what kind of bastard company ENVYdia really is.

[SamuriHL]

Do you have a link?https://wccftech.com/amd-radeon-rx-6000-big-navi-graphics-card-pictured-16-gb-gddr6-memory/

Remember these are leaks and speculation. Nothing to get all hot and bothered about just yet. But if this is true, and that's not guaranteed, then they will be competing with the 3070. Figure slightly faster with more ram for about 50 bucks higher. This may explain the supposed leaks by Lenovo of a 16gb 3070ti.

Quite frankly I want amd to succeed. I always have. I go back a very long ways in pc building and started out with amd for over 10 years until Intel annihilated all competition. Anyone remember cyrix? I used amd GPUs until the Nvidia 4xx cards, as well. I'm very seriously looking at the zen 3 but if these leaks turn out to be accurate about big Navi then I can't yet take them seriously for GPU.

Sent from my SM-G975U using Tapatalk

[butterw2]

Nvidia going with Samsung for Ampere is actually good for competition (vs TSMC) in the sub-12nm semiconductor manufacturing market.

If price/performance isn't there in mid/entry price range, don't buy, buy used, or buy a console instead. Not much point in complaining about lack of competition, Nvidia has a rockstar (or at least leather jacket) CEO and wants to become a datacenter company.

[chros]

I don't think that AMD is much bothered by desktop GPUs. They have the console market now and that was an important step for them, that's where the money comes from (along with CPUs, datacenter products). The desktop market compared to the rest is mimimal.

[nevcairiel]

I don't think that AMD is much bothered by desktop GPUs. They have the console market now and that was an important step for them, that's where the money comes from (along with CPUs, datacenter products).

"now"? :)
Consoles from the last ~7 years were based on AMD APUs, and both the XBox 360 and Wii before that used ATI GPUs - in combination with other CPUs, however.

NVIDIA, on the other hand, has been making more money from datacenter the last quarters then gaming products.

[SamuriHL]

But nVidia has the SHIELD and, uh, the Switch. Come on, those are killer! :D LMAO

[el Filou]

The comparison of TSMC 7nm vs Samsung 8nm for Ampere using GA100 vs GA102 gives a 47% higher density for TSMC and better efficiency.You can't compare the transistor and power budgets of a compute-only chip with a full graphics chip.As I have already mentioned, Ampere's TFLOPS are "fake" meaning they advertise a maximum theoretical value not reachable in real-world.As people have explained to you already but you seem to ignore, that's the case for all TFLOPS figures advertised by all GPU manufacturers.We don't compare performance per area here, where the NAVI10 using TSMC 7nm simply destroys Turing with TSMC 12nm.
It's a TDP of 225 Watts for NAVI10 using fewer transistors vs TDP of 215 Watts for Turing that really matters.Consumers (except maybe you?) don't care about "performance per chip area", only about performance per price or sometimes performance per watt.In the beginning there were many GPU companies but the leather-jacket-man did many and very bad things and like a miracle nVidia managed to have another covered monopoly or pseudo-duopoly.What bad things? Releasing technically superior (both in hardware and drivers) products that customers bought while Number Nine, Rendition, Intel, S3, PowerVR and Matrox (in decreasing order of awfulness) couldn't keep up and while 3dfx sabotaged itself by deciding to cut its board manufacturing partners? :rolleyes:

[NikosD]

As people have explained to you already but you seem to ignore, that's the case for all TFLOPS figures advertised by all GPU manufacturers. You and anyone else who have the guts to compare by any means the RDNA/TURING/RDNA2 TFLOPS vs Ampere TFLOPS and find them equal, I would declare you utterly ignorant and dangerous trolls of tech, that you should abandon writing in technical forums.
Nobody has explained to me anything of value.
You are still writing your opinions and mainly garbage.
Like this one you posted here.
I'll repeat it in case you eventually understand it.
Not all TFLOPS are equal.
It has already been reported by nVidia itself as I wrote before that the 30TFLOPS 3080 card has "up to 2x" the performance of the 10TFLOPS 2080 card.
Judging by the architecture of the Ampere cores (SM) it has already been processed that in real-world 1TFLOP of Ampere is 2/3 TFLOPS of Turing.
So the 3080 card of 30 TFLOPS produced by Ampere cores is equal around to 20 TFLOPS produced by Turing cores or "up to 2x" of the 2080 card.
This is the real-world performance of Ampere card.
Who cares what nVidia writes in its specs as "maximum theoretical value TFLOPS" of Ampere ?
Consumers (except maybe you?) don't care about "performance per chip area", only about performance per price or sometimes performance per watt.What bad things? You have serious issues understanding simple sentences not even technical.
I referred to performance per area as a metric that we don't care here for comparison reasons.
I mentioned it as a metric we don't compare, but instead we tried to compare performance per watt.
Next time be more careful with what you read, understand and write.
Unless you want to be the second troll of this forum.

[NikosD]

I have to be honest to this forum, like I always try to be.

The last few days dozens of tweets and more than 5 technical hardware news sites that I respect their opinion, have reported that Big Navi of 80CUs, the largest Navi21 according to rumors, is going to have 16 GB of VRAM and 256 bit bus.
Also, most of them if not all, seem to agree that Big Navi of 80CUs is going to compete with 3070 after all and not 3080.
On the other hand, I against all these respectful sites, believe that if AMD decides/wants/needs to release the Biggest/Baddest/BigFucking NAVI of all time using 80CUs (probably Navi21) is going to have 20 GB of VRAM and 320 bit bus.

So, it's ME against THE WORLD.

We'll see...

Now, according to a tech journalist of Extremetech site called Joel Hruska, Big Navi of 80CUs is going to be slower than 3070.
He claims that because of 3070 having 5888 cores and Big Navi of 80CUs having 2x40CUs = 2x2560 cores = 5120 cores, it will struggle to compete even with 3070.

Thank God there are actually so incompetent tech journalists comparing wrong figures in such ridiculous way that I'm now more confident that the 80CU RDNA2 chip is going to have 20GB of VRAM and will come close to 3080 with a suitable clock.

I'll leave it to the users of this forum to understand why this "tech journalist" is writing nonsense.

After so much training by me, you should be able to figure it out.

If not, I failed as a teacher.

[chros]

"now"? :)
Consoles from the last ~7 years were based on AMD APUs, and both the XBox 360 and Wii before that used ATI GPUs
:o Oh, sorry, I stopped playing a long time ago.

I just realized why nvidia was/is scared: M$ will provide netflix type of subscription for games: $25 for small and $35 for big Xbox per month for 2 years including the device. I think this is big for gamers (well not for me :)).

[NikosD]

A really interesting read for this afternoon:
https://www.mooreslawisdead.com/post/nvidia-s-ultimate-play

[ryrynz]

Therefore, I think this is Nvidia’s Ultimate Play: Intentionally causing an initial dearth of Ampere stock, allowing “supply and demand” to inflate the street price of Ampere when those beautiful $699 Founders cards instantly sell out, and then ultimately forcing AIBs to sell most of their models well above MSRP due to the required BOM Cost-Downs…that will be in ample supply once the street price is elevated.

So letting the market do it's thing, this is what a smart company does.. I applaud them for it, all hail Capitalism.

[NikosD]

These are the tricks of an abusive, ruthless, manipulative company acting as monopoly to the free market and NOT LETTING free, capitalistic market work as it should, in order to keep margins and prices high, while it looks like is selling in lower prices.

Marketing propaganda and nVidia fan boys help a lot here to increase the hype.

Hail to nVidia means hail to crippled "free" market, hail to high inflated prices for the customer.

Anyone who doesn't have nVidia shares in his hand or not being a mindless fan boy should see this and criticize this behavior.

[chros]

3060Ti leaks (https://www.youtube.com/watch?v=4-BWfmBP_WE), just like last year the 2060S, it's pretty close to 3070 (same die, lower cuda cores, etc.). If that will be the case then I don't see the point of 3070 since it's not good enough for 4K gaming with its 8GB, so a 3060Ti would be cheaper to get (let alone madvr usage). The only question remains with how much: if they follow last year prices (just like they did with higher end models) then it will cost $400. And that's way too high for me ...

The other thing that there are really confusing "leaks" about RDNA2, some state it's crap, some state it's an Ampere killer :D
How on earth can this happen with AMD at this stage that I don't understand ... :)

[SamuriHL]

Because there's 2 chipsets for AMD, and I suspect the leaks so far are on the lower end one. NO ONE should be buying a 3060 or 3070 until the AMD reveal. I'm still planning on getting my 3080 this week if I can find one. On the lower end cards, where AMD will absolutely compete, there is going to be a price war I think. Fun time to be a consumer.

[CZ Eddie]

Just noticed these up on the USA BestBuy.com website.
It says all three will be launching this coming week on the 17th.
https://www.bestbuy.com/site/searchpage.jsp?st=rtx+3080&_dyncharset=UTF-8&_dynSessConf=&id=pcat17071&type=page&sc=Global&cp=1&nrp=&sp=&qp=&list=n&af=true&iht=y&usc=All+Categories&ks=960&keys=keys
https://i.ibb.co/hH12Kwp/Untitled.jpg

[huhn]

i'm not far away from not caring about AMDs next GPU anymore after the nightmare of my 480 and the current nightmares with my 5700 XT.

freesync is a lie on these cards cap a game to 200 FPS make an overlay on and see the monitor OSD displaying between 120-240 while the game and the overlay clearly displays 200 FPS all the time the card isn't outputting it it just outputs what ever it wants to. take a much slower nvidia card and see how the monitor OSD and the overlay/ game FPS are spot on with an AMD verified freesync screen there are no word for this.

the pathetic QA. they just sometimes release driver where you can't switch the refresh rate they didn't even check if the OS get's reported the correct refresh rates i mean the driver can't switch refresh rate so no point in checking if the OS can right?

virtual super resolution = stick at highest refresh rate because reasons.

reporting the wrong native resolution they love to do that with TV by saying 4096x2160 is "native" so you get from time to time a user asking why is my screen cut of...

there wasn't a signal day in the life of the 5700 XT where deint ever worked correctly.

[chros]

Interesting to see the different cooler designs, for the 1st time nvidia did something completely different than its AIBs :) I'll be curious about the tests regarding to this, which will be more efficient.

[chros]

i'm not far away from not caring about AMDs next GPU anymore after the nightmare of my 480 and the current nightmares with my 5700 XT.
:D I can understand, no one can blame you ...

[SamuriHL]

The FE card has been up there since shortly after the reveal. That's likely where I'm going to try to order one from on Thursday. I can't wait and hope I'm able to grab one.

[NikosD]

The guy from MLID that I posted his article, says that Big Navi is not going to be that "Big" eventually.
The die is small and for the first time power consumption could be lower than the competing Ampere cards.
But he also says that memory bus and memory bandwidth will be lower, so no 320bit I guess - probably it's going to be 256 bit and 16GB even for Big Navi21.
At the same time he seems positive that is going to crash 3070 and compete directly with 3080.
With so low memory bandwidth ?
All these estimations are for rasterized performance.
Ray tracing is going to be faster than Turing and slower than Ampere, according to him.
I will definitely wait for RDNA2 cards which are ready months ago and will be released with polished drivers, probably for the first time in history.
But AMD cards are famous for keeping their value during their long life of supporting drivers and performing better in the long run than nVidia cards.
For example the RX 470/RX 570 card is clearly faster these days than 1060 3GB card and RX 480/RX 580 card is faster than 1060 6GB card.
These nVidia Pascal cards were faster than Polaris cards using their initial drivers.
Anyway since I saw also the video of Steve from Gamers Nexus bashing Radeon Marketing team for the hilarious history of misleading hype and underperforming cards, I hope this time history will not repeat itself with "Poor Volta" ridiculous hype.

I'm confident that RDNA2 cards will be more competitive than ever to nVidia's top Ampere cards.

[varekai]

Hello!

GPU NVIDIA Geforce GTX 960, drivers 398.11 win7 x64.
Would this old GPU benefit with latest drivers 452.06?

[chros]

I doubt it will.

Something else: nvidia just bought ARM (https://www.youtube.com/watch?v=1nYPNXwpNS4) for $40 billion. I have mixed feelings about this ... I like one of the comments: "The birth of Skynet" :D

[ryrynz]

Anyone who doesn't have nVidia shares in his hand or not being a mindless fan boy should see this and criticize this behavior.

The consumer cares only about what they can buy and it's perceived value. Everyone with a 30 series Geforce is going to be very happy indeed. *Begin applause* the world moves on as the market continues as it always has.


Something else: nvidia just bought ARM (https://www.youtube.com/watch?v=1nYPNXwpNS4) for $40 billion. I have mixed feelings about this ... I like one of the comments: "The birth of Skynet" :D

An excellent move, finally the beginning of the end for x86, lots of cash to put into finally putting this old bird to roost. I called this happening back in 2006.

[SamuriHL]

Where the hell is the new driver?! I'm starting to get impatient with them. I know it exists because all the people out there who've gotten prerelease cards are using them. We need to see if their claims of a madvr fix holds true.

Sent from my SM-G975U using Tapatalk

[NikosD]

nVidia is going to acquire Arm, if some specific companies and countries allow it.

It's not finished yet.
It seems that nVidia wants to become "Apple of silicon" but I don't know what real Apple and China say about this.

Apple is the worst company in the world, but this is a different discussion.

[VBB]

Hello!

GPU NVIDIA Geforce GTX 960, drivers 398.11 win7 x64.
Would this old GPU benefit with latest drivers 452.06?

It's safe to use up to 446.14. Anything newer breaks HDR with madVR.

[SamuriHL]

^^^^

Yea! That's why I want the new driver so badly. LOL

[chros]

The other thing that there are really confusing "leaks" about RDNA2, some state it's crap, some state it's an Ampere killer :D
How on earth can this happen with AMD at this stage that I don't understand ... :)
This is what Coreteks thinks (https://www.youtube.com/watch?v=BR70xbcwB6U) about RDNA2 GPUs. In short: big Navi will be only 3070 + 15% (with AMD preferred titles!) and won't be much more power efficient than Ampere.
(He also tries to answer my question above.)

Well, if that will be the case then only the pricing can save AMD from another disaster :)

[SamuriHL]

That's pretty much what I've been speculating this entire time. I think they beat the 3070 but fail to match the 3080. And they will be forced to lower prices before launch to compensate.

Sent from my SM-G975U using Tapatalk

[el Filou]

Oh look, AMD has teased a 3D render of their cooler design and confirmed 2x 8-pin power.
It would be pretty funny if NVIDIA announced an RTX 3060 on the day of AMD's Navi2 presentation (or just after).

[chros]

It would be pretty funny if NVIDIA announced an RTX 3060 on the day of AMD's Navi2 presentation (or just after).
They will do it before it (AMD is on okt 28).

And here's Jim's ranting about Ampere (https://www.youtube.com/watch?v=tXb-8feWoOE&t=269s): "Ampere is no Pascal", "Ampere is bad choice for everyone" :)
He talks about power consumption that is mainly caused by Samsung 8nm.
He doesn't mention 1 thing though that Coreteks talks about in his RDNA2 video: the power hungry TSMC 7nm+ :)

I still think that lower end Ampere cards (3060/3050/etc) will be the go to for madvr if madvr receives any update at all in the future and those cards will get all the goodies of the higher models (HDMI 2.1, AV1 decoder, etc).

[huhn]

there is no point in saying ampere is power inefficient when it is currently the most power efficient card "on the" market. (we see tomorrow if that's true...)

but what do you expect from someone saying "these clocks weren't fast enough to feed these numbers of shader" ok that's how a shader work sit's the feeding speed thank you...

what we currently know about 7nm is that it performance terrible in navi and can't even beat 16 nm TSMC at the same clock speeds.

and even on zen 2 the clock speeds are not that impressive it's competing with a process that's been refined for like ever but still not that impressive if you consider that 7nm is a much better process but power efficiency is quite impressive.

until AMD releases a GPU which is more power efficient using 7nm his "argument" is absolutely worthless. and even if it is more power efficient which it easily could be that again is not that great if the heavily needed process is to expensive.

[varekai]

It's safe to use up to 446.14. Anything newer breaks HDR with madVR.
OK, thanks for the info.
Do you think there will be any performance improvements with that driver?

[el Filou]

Not with madVR and on Win7 and an older GPU like 960, I don't think so.

Apart from obviously supporting newer GPUs, newer drivers are only good for:
- better supporting newer games
- supporting newer APIs, or WDDM versions in Win10
- sometimes providing better default video timings or functionalities (like some driver a few years ago added some LPCM sound formats for HDMI)
- security issues if you worry about that

But they also sometimes break things in point 3.
If you don't have any bugs bothering you and don't care about the above, I don't see a need.

[NikosD]

Interesting guy that one from Coreteks, although a little nVidia biased.

So, I agree with him like MLID guy that due to internal reconstruction of caches, the Big Navi21 can be faster (a lot ?) than 3070 with the same 256bit memory GDDR6 and with 16GB VRAM
No 320bit and 20G VRAM as I thought before in order to compete with 3080.
I lost this bet against the world, because probably it doesn't need 320bit to come close to 3080.
But he says only +15% over 2080 Ti.
That could be the case before Ampere's announcement with Big Navi21 initially at 275W TDP.
The last conversations at AMD, as MLID and Igor said also, have put Big Navi21 to 300W just a little less than 320W of 3080.
The higher clocks of Big Navi21 at 300W will bring performance a lot closer to 3080 just as power consumption.
I also believe like him that AMD will bring HBM2e Big Navi on Radeon gaming cards to beat 3090 but nVidia could bring an Ampere Titan card to respond.

Two bad things that I kept from Coreteks' analysis that really annoyed me and possibly could be true but I hope not:

1) No smaller/ cheaper cards than Big Navi21 for 2020!
Even the Big Navi22 will be released on late Q1 2021 or early Q2 2021 according to him.
That's a disaster (if true) for smaller wallets.

2) He said that AMD will repeat itself like Vega64 LC (Liquid Cooling) and 5700 XT Anniversary card for the highest Big Navi21 300W card that competes with 3080.
And the "normal" Big Navi21 at probably 275W will be like 3070 (+15% ?)
That's not good but it's not bad.
AMD still (before RDNA2 architecture) hasn't managed to beat 1080 Ti Pascal card.
So, let's wait for tomorrow's real-world reviews on nVidia's cards in order to see the perfect samples of Ampere 3080 GPUs that leather-jacket-man has sent to reviewers in order to be highly overclockable and with the lowest power consumption possible, how will stand against 2080/2080 Ti.

Is it possible for 3080 to be twice as faster in rasterized games than 2080 on average for the ten- fifteen most popular games ?
No.

But measuring the real-world performance of 3080 will probably unlock Big Navi21 benchmarks a lot sooner than 28th of October...

[varekai]

Not with madVR and on Win7 and an older GPU like 960, I don't think so.
But they also sometimes break things in point 3.
If you don't have any bugs bothering you and don't care about the above, I don't see a need.
Thanks for the info.
"if it ain't broke, don't fix it"

Best regards,
varekai

[Pastol]

Hi.

What is the alternative to madvr on the latest nvidia drivers for hdr material

[SamuriHL]

Hi.

What is the alternative to madvr on the latest nvidia drivers for hdr materialIMO the isn't one. The problem exists in other players too.

Sent from my SM-G975U using Tapatalk

[NikosD]

Reviews are out for 3080.

14 Game Average of Hardware Unboxed:

1440p
3080 is 20% faster than 2080 Ti and 50% faster than 2080

4K
3080 is 30% faster than 2080 Ti and 70% faster than 2080

More or less, the expected results.

[nevcairiel]

More interesting, in heavy OpenCL/CUDA workloads (eg. Blender tests), there is a strong >100% uplift over 2080. This might benefit compute-heavy scenarios like running very heavy madVR settings quite a bit, I could imagine, since any other factors that could limit performance are pretty much non-existant and the huge number of FP32 cores can really do their work.

[huhn]

there are benchmarks where the 5700 XT wins against a 3080 so production benchmarks are sadly hard to judge. as far as we know madVR is shaders "only" so yes could be great but well polaris anyone?

and yes it's the most efficient card on the market.

[NikosD]

After three reviews I have seen so far, it's pretty obvious that nVidia failed this time badly to deliver a gaming (Geforce) card.
Ampere architecture is basically a computing architecture for data center and this is obvious in gaming as well.
The performance of 1440p - a really high resolution for most gamers as still the majority uses 1080p monitors - is not so good compared to 2080 Ti and even 2080.
The power consumption is huge, analogous to the Data Center destination of Ampere architecture.
Real-world power measurements of 3080 give numbers between 330W to 350W as peak power consumption vs 250Watts of 2080 Ti.
The result is that 3080 is only 8% better in performance per watt compared to 2080 Ti at 4K resolution.
The 3080 is strictly a 4K card not a reason to buy it if you have a 1440p monitor or below.
But at the same time 4K is going to need more than 10GB of VRAM in the near future, so a disastrous combination.
AMD has a real opportunity here to attack viciously this non gaming friendly card with Big Navi.
And with AV1 HW acceleration added to AMD Radeon RX 6000 cards, I have really all the good reasons to wait for a good price to performance VFM Big Navi card than to buy an Ampere card to run Blender!
For God's sake!

[SamuriHL]

Yes....you ALL want to wait for AMD. /waves hand like Jedi

Then I can get my 3080 tomorrow. :)

[huhn]

well i wonder how high the number of user is that buy a 3080 for 700 and have a 1080 screen?

[nevcairiel]

If I actually planned on gaming at 4K, I might wait for a 20GB variant, or a Ti, or whatever they are going to have, but right now I'm on 1440p, and likely to stay there for the time being. But I also have other reasons to get a 3080, because I do use productivity tools, and of course I need something to work on AV1 decoding, so that I can finally make a worthwhile LAV release again. :)

I agree that on 1080p its mostly nonsensical, you can get a 3070 there for high-end performance.

[NikosD]

Interesting that Digital Foundry accepted their mistake or "mistake" to benchmark Doom Eternal to highest Texture Settings at 4K that were simply above the 8GB of 2080 and the performance difference from 3080 was about 100% or even more due to limited memory of course.
Using a slightly lower setting "ultra" in order for textures to fit in the 8GB VRAM the performance difference is around 70% like the average of 4K gaming.
If 2080 was a PCIe v4.0 card tested on Ryzen platform in order to support PCIe v4.0, the performance difference could be less.
But I also see some reviewers like Digital Foundry that are trying a little desperately to increase the hype, comparing a 3080 vs a 1080 card of 2016!
Oh, come on guys...You can do better than this!

[chros]

For 1080p probably a 3060 will be enough :)
The rest is really good news, this card is a beast in every way.

[SamuriHL]

Yea for 1080p 144hz gaming I have my monster Alienware Area 51m with a 2080 so I'm covered there. I agree about the memory, Nev, but, I'm not waiting. I think that 10gb of GDDR6x will be fine for the majority of 4k gaming. And if after a year or 2 it isn't, ok, well I can upgrade again to something better at that point. I don't necessarily buy the October surprise rumors about a 20gb 3080 unless AMD hits a homerun which is looking increasingly unlikely. AV1 support would be nice, though. And if you have a 3080 I can be sure it'll work well for me when you release it. Sweet :D

[SamuriHL]

Are they seriously going to wait til tomorrow to release the damn driver? SIGH

[CZ Eddie]

https://us.download.nvidia.com/Windows/456.38/456.38-desktop-win10-64bit-international-whql.exe

https://uk.download.nvidia.com/Windows/456.38/456.38-desktop-win10-64bit-international-whql.exe

https://us.download.nvidia.com/Windows/456.38/456.38-win10-win8-win7-release-notes.pdf

Nvidia released GeForce 456.38 WHQL driver.

Fixed Issues list includes:
[madVR][MPC-HC]: Various HDR issues occur when using the madVR renderer with MPC-HC. [3038381]

https://videocardz.com/driver/nvidia-geforce-game-ready-456-38-whql

[SamuriHL]

At least the driver is released. But no humans were able to buy any cards. They were sold out in microseconds, all snatched by bots and are now up on eBay. WTF Nice job, nVidia. Way to push people to consoles. At least those pre-orders are easier to get.

[CZ Eddie]

This one shows available at Best Buy
https://www.bestbuy.com/site/gigabyte-geforce-rtx-3080-10g-gddr6x-pci-express-4-0-graphics-card-black/6430621.p?skuId=6430621

[CZ Eddie]

This one shows available at Best Buy
https://www.bestbuy.com/site/gigabyte-geforce-rtx-3080-10g-gddr6x-pci-express-4-0-graphics-card-black/6430621.p?skuId=6430621

Not anymore.
Keep checking other cards on BB because I saw another one come up available that had been out of stock earlier.

I'm waiting for something to show up on Dell so I can use my $400 in coupons & credits I have available.
I have no idea when they will start selling them.

[SamuriHL]

Yea they're releasing them periodically but before I can even get through the cart it's gone. It's insane.

[huhn]

HDR works "fine" on this driver even with overlay rendering.

edit: this is not an indeep check if meta data and such.

[SamuriHL]

I'll check metadata tonight. I have to hook the integral 2 back up to do it, but not a big deal.

[chros]

I'll check metadata tonight. I have to hook the integral 2 back up to do it, but not a big deal.
Thanks, can you also check the rest of the bullet points (https://forum.doom9.org/showthread.php?p=1921810#post1921810) as well?

Jim explains Ampere cuda cores (https://www.youtube.com/watch?v=nJBggUfYozY) one more time. Interesting stuff in it:
- 3080 has ~3x cuda cores of 2080, but only 70% faster
- according to nvidia, ~25% of instructions are int out of all
- since the new path can either run fp or int, and int are 25%, this path can be occupied 50% by int

[nevcairiel]

- 3080 has ~3x cuda cores of 2080, but only 70% faster


Games (and game-like benchmarks like 3DMark) are often being limited by something else rather then pure FP32 performance. More shader-intensive workloads would benefit from the additional cores more. Maybe future more graphically intensive games will benefit more from it.
Some productivity benchmarks have shown the real power of the massive number of FP32 CUDA cores, but of course for that to be meaningful an application has to be able to use it.

Interestingly, AMDs Fiji architecture from years ago had the same issue. They had a lot more potential shader power then they could get games to use. Todays use might be limited, but future use, who knows. Free shaders for RTX IO and stuff. But of course one shouldn't buy hardware on account of future what-ifs.

(Also, I can't watch Jims videos, his accent just drives me crazy)

[SamuriHL]

Thanks, can you also check the rest of the bullet points (https://forum.doom9.org/showthread.php?p=1921810#post1921810) as well?


I can check that metadata is being sent, someone else will have to test the rest of the bullet list.

[huhn]

like rop, TMU, cache and so much more.

memory bandwidth is not comparable too because they changed the lossless compression over the years.

[NikosD]

You don't need to watch Jim.
You only need to read my posts here, I've been telling you the same things from day one BEFORE he releases his videos.
Now, regarding the gaming workload analysis, the 25% of INT commands mentioned above is wrong.
The figure given by nVidia when Turing was released back then at 2018 that was the first architecture to execute INT+FP at the same time was around ~35 % or so to speak 1/3.
That's why the 30TFLOPS become 20TFLOPS and the number of CUDA cores follow the same pattern of 2/3 real-world number compared to Turing.
Now I have some info/news/prediction for Big Navi.
It seems that the model that will be announced on 28th of October will be a 60CU - 64CU die that could possibly reach 2.4GHz for the 3 fans edition of 300W+ TDP.
It will have 256bit bus with 16GB of GDDR6.
I'm under the impression that will beat 3080 at both 1080p and 1440p resolutions but will lose a little at 4K.
AMD will release later the full die of 80CUs using HBM2e memory and will beat even the 3090 card even at 4K.

[SamuriHL]

So, mostly good news with this new driver. More or less. When you play an HDR title now, it'll start out as SDR and then switch to HDR mode and then enable HDR. Those of you who aren't familiar with the work around for the previous few drivers of starting a movie, stopping it, and playing it again are probably wondering wtf I'm talking about, but, they've essentially appeared to "automate" that process. The metadata is being sent correctly now, as well. Hooray for small miracles. Now for the not so great stuff. If you bring up your player's UI, it'll take you out of HDR into SDR. But we've seen that behavior quite a bit before. It's whatever. I can live with this current driver. And hey, if I ever do manage to land a 3080 in the next year, I'll be all set. LOL

[VBB]

Thanks for testing so quickly, bud! So, are we back to using NVidia's API for HDR with this driver, or does it still switch to OS HDR? The player UI bug has been around since early this year, so I guess that hasn't been fixed yet.

[Klaus1189]

Any improvements of the out-of-the-box 23 Hz refresh rate? Dropped/repeated frames...

[huhn]

i have nor rebooted yet but i'm back to 23.980

[Klaus1189]

Please reboot :)

[huhn]

same.

[Klaus1189]

Have they added new options, like 10 bit? Huge new branch so possible in my head and a workaround for madVR ... very nice of Nvidia.

So the new driver behaves like that?
445.75 2020-03-23 This version and all 445.XX and higher: Playback at greater than 8-bit drops back to 8-bit (and out of HDR) in FS windowed mode when player GUI is triggered, regardless of bit-depth of the file. Can be circumvented by using FSE mode or by setting madVR to 8-bit.

[huhn]

i can only tell you that "everything" is fine at my end with overlay rendering if it comes to HDR.

and they did not add 10 bit in this driver and they may never add 10 bit for HDMI 2.0 signals jusdt for HDMI 2.1 (they are packet differently).
edit: please report if 23.976 is 23.976 on your system witht he newest driver so i can report it to madshi. one case is far to little to come to a conclusion

[oldpainlesskodi]

See, I can't help myself. Lord forgive me.....AMD=HTPC.

[Klaus1189]

If someone got a 3080 please let us know then if there is 10 bit available. You will of course need a display with HDMI 2.1 input like newer TVs and please report your refresh rate here, info is needed if 23.976 is 23.976 on your end, see previous post of huhn.

[huhn]

See, I can't help myself. Lord forgive me.....AMD=HTPC.

so you found a way to send BT 2020 to the display at will ?

[DMU]

@huhn
what are you speaking about?

[oldpainlesskodi]

@ Huhn...that's loaded question, so Touché. And, no, I see no reason to send BT2020 at will.

[huhn]

3D LUT gamut mapping.

it's "important" for PJ where some devices trade CR for gamut.
currently AMD choice the gamut as they wishes which can work but... with nvidia you can force BT 2020 under any circumstances which is quite a nice feature for calibration.

the real bad things about AMD currently are deint which is a joke but no one laughs.
the just bad HDR api which only works with 10 bit which is just dumb and needs FS if i recall correctly...

taking information from mclingo and the current user in the madVR thread there could be a horrific issue in the API.
i didn't had a HDR TV in the past just a 4K TV so i couldn't test it. BTW. my old TV had a better CR in BT 2020 mode i didn't test the new one yet waiting for some serious firmware fixes first before i touch calibration again.

[DMU]

the real bad things about AMD currently are deint which is a joke but no one laughs.
the just bad HDR api which only works with 10 bit which is just dumb and needs FS if i recall correctly...

All of the above can be applied to the Microsoft (WinAPI) HDR API. HDR only with 10 bits and in fullscreen. And given what is happening with the Nvidia drivers, they will soon come to this too.

[huhn]

win API clearly works in 8 bit windowed and so on so what do you mean?

[DMU]

This is not true. Are we talking about HDR API? Here's an example (https://github.com/microsoft/DirectX-Graphics-Samples/tree/master/Samples/Desktop/D3D12HDR) from Microsoft. Try to get HDR with true metadata on it at 8 bits and not in fullscreen.

[VBB]

Huhn, curious which new TV you ended up getting...

[huhn]

well as long as you are in window mode the HDR meta data would be wrong the file is not the only thing displayed and in the end no proper screen cares about meta data it has to measure far more accurate for every frame anyway. your sample is 2 years old and the windows DWM is evolving with pretty much every major windows update.

and i never even hinted that the windows API is any good.

edit: XH90 for my 120 HZ needs.

[oldpainlesskodi]

Oh boy....AMD, MadVR FSE, 10bt (10 bit at driver level too for 30hz and below....) passthrough AMD api HDR = perfect.

No need to continue this debate, I was just bored.

Update - I don't like anything other than pure refresh rate matching - on movies, anything else is a cluster f&^k.

[DMU]

your sample is 2 years old and the windows DWM is evolving with pretty much every major windows update.

This sample was updated (https://github.com/microsoft/DirectX-Graphics-Samples/tree/master/Samples/Desktop) 2 months ago. ;)

[huhn]

well look closer because tehy change something in thsi project doesn'T eman they touched the HDR code.

the d3d12HDR.ccp was lostz touched 2 years ago with this world changing edit: "Replace tabs with spaces"
and we have this change: <WindowsTargetPlatformVersion>10.0.17763.0</WindowsTargetPlatformVersion>
<WindowsTargetPlatformVersion>10.0.19041.0</WindowsTargetPlatformVersion>

the UWP version was changed but that'S not what you linked.
that still doesn't mean i'm correct i don't have the time to really check if this example you post is 2 years old i be like always be wrong.

[SamuriHL]

I'm showing 23.97585 on my 2070.

[huhn]

Oh boy....AMD, MadVR FSE, 10bt (10 bit at driver level too for 30hz and below....) passthrough AMD api HDR = perfect.

No need to continue this debate, I was just bored.

Update - I don't like anything other than pure refresh rate matching - on movies, anything else is a cluster f&^k.

so an API that only works in two mods is perfect compared to an API that literally can do HDR with everything under pretty much every condition?

if an image with zero control over it is perfect than yes the AMD API can do that and nothing else.

[Klaus1189]

I'm showing 23.97585 on my 2070.

Out of the box? No custom tweaking?

[SamuriHL]

Out of the box? No custom tweaking?

Correct. I don't use CRU or anything else. Just out of the box, that's what I'm getting.

[Klaus1189]

@SamuriHL: Thanks

@all: please post your refresh rate for huhn and madshi

[jmone]

Here is a summary of my quick test with the new nVidia drivers after a couple of hours playing a movie. I was hoping for no frame repeats/drops @ 23.976 (4K) out of the box with no tweaks... close but no cigar. Also the Refresh Rates move around a little over the playback period.
- 23.97531 to 23.97536 : 1660Ti to a Sony OLED - 1 Repeated Frame in 2hrs of playback
- 23.97452 to 23.97458 : 1660Ti to a JVC PJ - 2 Dropped Frames in 2hrs of playback

Also, It seems the ability to custom tweak from inside madVR is broken on this driver. You can "Test" just fine, but when you go to save it fails "The GPU driver rejected this mode, for unknown reasons".
https://jmone.org/jmonenextcloud/index.php/s/SztdBYbtekQR4H6/preview

[cableffm]

23.97576 / 1 drop in 1.1hours
without tweaking.

sometimes it's moving between .97572 and .97576.

(1 repeated in 5 mins before)
2080ti-->Sony Oled
456.38
12bit/RGB full/23p
HDR output in madvr

Win10 2004, no HDR probs :D n
NVHDR works, no switch to OSHDR

[VBB]

Same here with my GTX 960. Went from 4.5 minutes to a little over an hour. I'll take that :)

Also, for me HDR goes straight into HDR mode, not SDR first.

[NikosD]

Paper launch, limited stock by design, scalpers with bots, ridiculous prices.

They reminded me a recent article:
https://www.mooreslawisdead.com/post/nvidia-s-ultimate-play

Also, the queues outside physical retail stores were not good reminding me Apple's situation.

Some people discovered a new use for 3080 at 50100$
Money laundering...Interesting.
https://i.postimg.cc/m2mnrK3X/75242-02-zotac-geforce-rtx-3080-trinity-graphics-card-on-ebay-for-over-50-000-full.png

[VBB]

I think this also happened with previous drivers, but for some reason I cannot get YCbCr 4:4:4 8-bit to play nice with HDR in normal or PC mode. It always ends up with everything being green. If anyone knows how to prevent that from happening, please let me know. I tried exclusive, windowed, FS windowed, D3D9 overlay. All the same.

[nevcairiel]

Some people discovered a new use for 3080 at 50100$
Money laundering...Interesting.


I read an article just now that apparently some people are running anti-resale bots that make fake accounts and jack up the ebay listings of the cards to ruin the scalpers days, since it wont actually be sold anytime soon that way.
They just need to keep it up for long enough until supply normalizes, and the scalpers might be stuck with their cards and no profit to make.

[chros]

Also, for me HDR goes straight into HDR mode, not SDR first.
Which OS version do you use?

Some productivity benchmarks have shown the real power of the massive number of FP32 CUDA cores, but of course for that to be meaningful an application has to be able to use it.
Good point, do you have a link?

(Also, I can't watch Jims videos, his accent just drives me crazy)
:D Scottish, I really like it, I got kinda used to it during the years due to couple of colleagues. But check out one of his early videos :D
I like to read your insight as well, maybe you should also start a podcast :) (I got to the point that I mostly interested in theories rather actually using stuff :D )

You only need to read my posts here, I've been telling you the same things from day one BEFORE he releases his videos.
Your write ups would be also interesting to read if you wouldn't want to push them into others' throat :)

Now, regarding the gaming workload analysis, the 25% of INT commands mentioned above is wrong.
The figure given by nVidia when Turing was released back then at 2018 that was the first architecture to execute INT+FP at the same time was around ~35 % or so to speak 1/3.
No, it's ~26% (https://youtu.be/nJBggUfYozY?t=724) (not that it matters much): they didn't say 36 out of 100, but additional 36 on top of 100, meaning 36/(100+36) = ~25.9.
But of course it's only an average, as Nev said it can be completely different based on a given application, e.g. can be even different when using jinc vs NGU Sharp upscaling in madvr.

I'm under the impression that will beat 3080 at both 1080p and 1440p resolutions but will lose a little at 4K.
AMD will release later the full die of 80CUs using HBM2e memory and will beat even the 3090 card even at 4K.
I hope so, I like competition! Although I'm not sure about the HBM2 ones.

[NikosD]

We are talking here about gaming Geforce cards aka gaming workloads only.

And still, nobody has explained why Ampere is underperforming so badly in lower resolutions like 1080p and 1440p and this has nothing to do with CPU bounded gaming cases at lower resolutions.

This thing doesn't scale proportionally.

And the gaming workloads using 36 INT instructions per 100 FMADD instructions can't explain why in the best case scenario of 4K resolution, the 3080 of 30TFLOPS is on average only 70% faster than the 2080 of 10TFLOPS card.

A lot of hardware inside Geforce Ampere is underutilized in gaming workloads and at the same time the figures written in Ampere's specs are no real-world or even optimized or even best-case maximum figures.

The 3x of TFLOPS number hasn't been reported anywhere even in the most compute-intensive floating point productivity benchmarks like CUDA rendering using V-RAY and Blender.

[SamuriHL]

Same here with my GTX 960. Went from 4.5 minutes to a little over an hour. I'll take that :)

Also, for me HDR goes straight into HDR mode, not SDR first.I'll try rebooting and see if that fixes it for me.

Sent from my SM-G975U using Tapatalk

[NikosD]

Judging by the eBay reselling of 3080 cards at 1700$ to 2000$ (initial price 700$) and the pre-orders of PS5 reselling at 800$ to 1000$ (initial price 500$) it is safe to say that marketing departments should be closed and overloaded hype should be banned from everywhere.

But all these tricks are the base of the so called "free market" in order to create inflated demand, resulting in inflating prices.

[SamuriHL]

We'll see if the scalpers are brave enough to try this s**t with AMD cards. LMAO

[VBB]

Which OS version do you use?

Windows 10 Enterprise 64-bit 2004 19041.508

[VBB]

chros, SamuriHL, when you get a chance, could you please test YCbCr 444 8-bit with HDR to see if you also get that green screen issue? Thanks!

[Klaus1189]

https://www.youtube.com/watch?v=yE9n_llJOVQ
https://www.youtube.com/watch?v=qHogHMvZscM

[chros]

@VBB: Sorry, but I haven't upgraded it and I don't have time to fool around with it, hopefully others will.

We'll see if the scalpers are brave enough to try this s**t with AMD cards. LMAO
:D

[el Filou]

So, are we back to using NVidia's API for HDR with this driver, or does it still switch to OS HDR?After reading this KB article: https://nvidia.custhelp.com/app/answers/detail/a_id/5072 (linked to in the release notes for the last drivers), as they're suggesting to sync the Windows HDR setting with the application HDR setting (in this case, madVR), I am under the impression that NVIDIA is trying to dynamically switch Windows HDR itself on or off. Don't know if this is correct but clearly something has changed compared to the old NVHDR way that directly controlled the display from the app.Oh boy....AMD, MadVR FSE, 10bt (10 bit at driver level too for 30hz and below....) passthrough AMD api HDR = perfect.
No need to continue this debate, I was just bored.I get how people with some hardware that suits their use case can think some other hardware has bugs that they consider showstoppers, but sincerely, the best way to avoid tiring debates is to not start them by proclaiming that a personal use case equals the perfect hardware config for everyone. ;)

[NikosD]

People and users of this forum don't seem to understand even the basics of the covered monopoly game of Intel and nVidia.
Both these companies during the last few decades (!) were spending a huge amount of money in marketing and PR (BRIBING) of tech journalists and OEMs in order to do two simple things:

1) Avoid AMD's products inside their devices (OEMs) even if they are better (recent example of Ryzen processors in laptops, EPYC processors in servers etc and Threadrippers in HEDT) and exaggerate over any weakness (tech journalists) compared to theirs (Intel & nVidia) - recent example of 5% faster Intel processors in 1080p gaming.

2) Produce a lot of hype for their products even for useless features (hairworks, physx, RTX in 2018 for nVidia/ AI, AVX-512 inside Tiger lake and desktop processors for Intel etc) and bury or underestimate any AMD's better feature.

So, even after a desktop processor of 8C/16T for desktop Zen1 Ryzen which later became 16T/32T for Zen 2 even after a 32C/64T processor for servers which later became 64C/128T processor and the same for HEDT market with Threadrippers, the OEMs and tech journalists didn't have any intense positive reaction.

No overloaded hype was created, no inflated prices for AMD products.
Even for the best of the best CPU processors that mankind has ever created by far - the whole Zen 2 architecture implementations - no particular hype was ever created by OEMs and tech journalists.
Because Intel spends BILLIONS of $$$ to BRIBE these two - OEMs and tech journalists.
The margins of AMD are still low - even for Zen 2 CPUs.
So, after the next release of the absolute monster of CPU processors, the divine Zen 3 CPUs that will be released on 8th of October, no extraordinary hype will be created, no limited stock by design, no inflated prices.
Just the best processor you could ever have, in right stock and price.
The same with GPUs although not even this time nVidia is in position of Intel, unfortunately...

[VBB]

After reading this KB article: https://nvidia.custhelp.com/app/answers/detail/a_id/5072 (linked to in the release notes for the last drivers), as they're suggesting to sync the Windows HDR setting with the application HDR setting (in this case, madVR), I am under the impression that NVIDIA is trying to dynamically switch Windows HDR itself on or off. Don't know if this is correct but clearly something has changed compared to the old NVHDR way that directly controlled the display from the app.

Yeah, I'm sure MS is pushing OS HDR, and Nvidia will likely abandon their API eventually. With this new driver, though, at least madVR is still able to use NV HDR.

[el Filou]

2) Produce a lot of hype for their products even for useless features and bury or underestimate any [competitor]'s better feature.Congrats, you just described marketing. Every company is doing it when they can, even AMD. Remember the x3 multitexturing no devs used? Truform tessellation? AVIVO that wasn't even fixed-function video decode? TrueAudio? The one-die 4-cores when Intel "used glue" for its 4-cores even though it didn't matter much as the market for 4-cores was so small at the time? The supposedly "future-proof 8-core" FXs ?So, even after a desktop processor of 8C/16T for desktop Zen1 Ryzen which later became 16T/32T for Zen 2 even after a 32C/64T processor for servers which later became 64C/128T processor and the same for HEDT market with Threadrippers, the OEMs and tech journalists didn't have any intense positive reaction.
That's because for most consumer workloads until now, adding more cores had diminishing returns apart from things like two-pass video encoding (live encoding for streaming is best done on the GPU), 3D rendering, and pure compute, which are things most people aren't doing. Maybe it will change now even consoles have 8 cores and games start using them, but until now it wasn't a game changer. And no, the journalists did have intense positive reaction for the high-core count AMD models, but only for specific use cases."best of the best CPU processors that mankind has ever created" ... "absolute monster" ... "divine"You do realize you're talking like someone who's in a cult, right?

[Klaus1189]

If there wasn‘t AMD, mainstream Intel would still have 4c/8t...

[huhn]

yes AMD was the reason that intel didn't need to bring more cores to the market.

[almawdah]

So what the recommended , can we upgrade the driver to latest driver for nvidia to use it with MadVR ??

[Manni]

So what the recommended , can we upgrade the driver to latest driver for nvidia to use it with MadVR ??

I've only tested 8bits RGB with 456.38 (Studio / Standard) as that's what I use and don't have the time to do any other tests at the moment, but at least with that I can confirm that HDR works fine (with nvAPI) and that if using passthrough the HDR10 metadata is sent properly.

I don't know how it behaves in 12bits or in YCC.

[EDIT: Also I'm still using CRU so I haven't tested native frame rate.]

I don't have the issue of the "double switch" for HDR. It switches seamlessly between SDR and HDR, or SDR BT2020 and HDR passthrough.

So at least at first sight, it seems good to go :)

I'll post if I find any issue.

My full config is in my sig, YMMV.

[el Filou]

If there wasn‘t AMD, mainstream Intel would still have 4c/8t...But they only started releasing mainstream CPUs with more than 4 cores when AMD became competitive overall, it wasn't just the mere existence of more than 4 cores from AMD that made them do it: Phenom IIs had 6-core versions on the mainstream platform, but that didn't cause Intel to release mainstream 6-cores, they reserved them for the HEDT platform.
My point is that what's important is if the number of cores can actually benefit your use cases. It's good that AMD was able to release a 16-core on a mainstream platform with Zen 2, but with mainstream apps (i.e. Web, games, productivity apps) you hardly see the benefit anyway, heck in some cases the 12- and 16-cores can be slower than the 8-core. So while you have to salute the technological prowess, and 16 cores can be useful for content creators or devs or people who run a whole lab in VMs, I see no need for journalists to have an "intense reaction" (or is that a euphemism for something?).

[almawdah]

I've only tested 8bits RGB as that's what I use and don't have the time to do any other tests at the moment, but at least with that I can confirm that HDR works fine (with nvAPI) and that if using passthrough the HDR10 metadata is sent properly.

I don't know how it behaves in 12bits or in YCC.

I don't have the issue of the "double switch" for HDR. It switches seamlessly between SDR and HDR, or SDR BT2020 and HDR passthrough.

So at least at first sight, it seems good to go :)

I'll post if I find any issue.

My full config is in my sig, YMMV.


Thanks very much brother


Sent from my iPhone using Tapatalk

[Manni]

Thanks very much brother


Sent from my iPhone using Tapatalk

You're welcome :)

One last thing, I'm still using CRU so I haven't tested native frame rate.

I'll edit my post to reflect that.

[SamuriHL]

chros, SamuriHL, when you get a chance, could you please test YCbCr 444 8-bit with HDR to see if you also get that green screen issue? Thanks!

Oh for crying out loud. LOL Yea I can look into it in a bit. That requires a lot of setup. lol

[VBB]

:D What do you have to change? I appreciate it. I'll do some more testing tonight. Perhaps it just requires a reboot on my end, which I didn't get to last night.

[SamuriHL]

I still haven't rebooted, either and probably should. As for what I have to change, a bunch of crap. LOL First YCbCr 444 mode, then changing to PC mode on the TV, which apparently reset all my damn picture settings completely screwing up my calibration (SIGH, screw you LG)....Now that I fixed all that and set it up, I threw on 1917 just to test it. No green, and PQ looks pretty damn impressive. This isn't a movie where I'm likely to notice a lot of banding so, um, you're making me mad by having me test this seeing how good it COULD look if LG hadn't completely blew it on PC mode. LOL So uh yea thanks for that :D

[VBB]

LOL well thanks for letting me now that it works for you. I'm sure you do know that PC mode is completely separate as far as settings and calibration go. On my C7, the two modes even need different brightness settings to have the same black level.

Also, I noticed again last night that YCbCr 4:4:4 actually has very, very little banding in both SDR and HDR, in PC mode. That's why I'm now trying to get it to work.

[SamuriHL]

Yes I do know that which is why I said I had a bunch of setup to do. :) Not HARD to do but I just had to remember everything to set without looking. Anyway, it wasn't a big deal and I was able to test it.

I'd have to do a lot more testing on that theory. PC mode has been an unmitigated mess for LG. ALL of us have WANTED it to work but we've always found issues. I'm assuming you're sticking to 8 bit for this venture yes?

[VBB]

Oh, 8-bit for sure. It's pretty much established now that PC mode switches the TV into 8-bit mode anyway, and of course my old C7 and GTX 960 can't do full chroma at higher than 60Hz anyway. And, as we know by now, the new Nvidia driver did not enable 10-bit for anything lower than 120Hz outside of 4:2:2. So, nothing's really changed for us C7/C8 users, except a much more accurate 23Hz mode :)

I'm trying to switch from normal HDMI mode, 4:2:2 10-bit at 60Hz, with smooth motion, to PC mode 4:4:4 8-bit, with frame rate switching.

[SamuriHL]

So far I like this new driver a lot. I should grab the studio version just to have it. The improved 23Hz mode is a welcome change. We can thank madshi for pushing them on that one. :)

[VBB]

Definitely. BTW, now that I'm thinking about using PC mode again, has anyone ever had an issue with the lack of 5:5 pull down, since Real Cinema is permanently OFF?

[SamuriHL]

I use real cinema always so I can't tell you on that one. But that could be a real problem.

[jmone]

Here an update on my testing with the new nVidia drivers after a couple of hours playing a movie. I was hoping for no frame repeats/drops @ 23.976 (4K) out of the box with no tweaks... close but no cigar. Also the Refresh Rates move around a little over the playback period.
- 23.97525 to 23.97434 : 1660Ti to a Sony OLED - 2 Dropped Frame in 2hrs of playback (Decoding Audio)
- 23.97531 to 23.97536 : 1660Ti to a Sony OLED - 1 Repeated Frame in 2hrs of playback (Decoding Audio and Running JRivers "Video Clock")
- 23.97452 to 23.97458 : 1660Ti to a JVC PJ - 2 Dropped Frames in 2hrs of playback (Bitstreaming Audio)

My conclusions so far are:
- Out of the box: The drivers seem to adjust the Refresh Rate during playback to try to avoid Repeated/Dropped frames and it works pretty well... or is it just clock Drift? Anyway, the good news is that over the length of a 2hr movie I get 2 Dropped Frames on a couple of different setups.
- JRiver Video Clock: When I played a movie also using JRivers "Video Clock" (similar to Reclock), I then got 1 Repeated Frame. I don't know if they are "fighting" each other but historically, Video Clock would give me 0 Dropped/Repeated frames on my OLED Setup. Maybe JRiver Video Clock needs a tweak to work a bit better with this new driver???
- madVR Custom Modes: Can't save any custom modes anymore to tweak the refresh rate which previously also got me to 0 Dropped / Repeated frames (on my Bitstreaming PJ setup).

So it is a big improvement but for the perfectionist..... still a bit to do (I guess I'll have to look at CRU unless others have suggestions???)

[SamuriHL]

The madvr custom modes has been broken for a LONG time now. If you want to do custom you need to use CRU.

[*MidnightWatcher*]

Performance in 12-bit now with the new drivers is the same as the previous drivers in 8-bit. Only one frame repeated on average during a movie, which is not perfect but pretty close.

[Calvi]

New driver works as good as 442.92 for me.
Zoomplayer osd does not drop out of HDR, same as above.
HDR works perfectly in full screen windowed MadVR/ZP.
And 23 gives me a dropped frame every 1-4 hours which is not perfect but close enough for me.
Will keep these drivers for a while.

[jmone]

Yeah - I spend the better part of this arvo trying to get to 0 dropped / repeated frames (CRU) but could not do better than what the new driver does out of the box, so I'm calling it a day on this (well for now anyway).

[NikosD]

RTX ray tracing is still a bonus feature even after two years of its introduction and not a different, more realistic way to render games.
Only a few games adopted the idea and even fewer added the whole package and not just one or two.
The owners of RTX 20 series Turning cards paid a lot for this hype following the "Just buy it" order of Tom's Hardware and nVidia fan boys.
The new 2nd gen RT cores and the new 3rd gen Tensor cores of Ampere cards don't help so much.
The difference of RTX OFF to RTX ON of Ampere and Turing is less than 10%, so you almost get the same performance drop for these two generations.
No, Ampere generation doesn't have the x2 performance gain compared to Turing for RTX that nVidia says.
It's actually ~10%.
Unfortunately the DLSS performance gain of 3rd gen Tensor cores of Ampere vs 2nd gen Tensor cores of Turning is even worse.
Almost not difference at all.
So no.
Ampere is not a game changer for RTX and DLSS features and that's why nVidia didn't sell these features for its hype.

[Klaus1189]

https://www.youtube.com/watch?v=_SO2b_VIOXI

[SamuriHL]

Ray tracing was never going to take off until consoles make it a commodity. Which is what we have now. Most new games coming in the next few years will now support it by default. Having seen what it can do with my own eyes on my 2080 I'm quite excited for this future. Even amd figured out they needed to support it. We'll see how well they did with performance.

Sent from my SM-G975U using Tapatalk

[QBhd]

AMD was always going to support it, it has been part of their "roadmap" for years... they just didn't rush it and try and gouge the consumer by promoting it early as a selling point and a must have technology. Saying AMD "figured out they needed to support it" is a little misleading and makes it sound like they are scrambling to copy NVidia. Don't forget who is working closely with MS/Sony on these consoles, it's AMD; so they have known for years that RT will be in demand for the future. RDNA2 was always going to be where AMD started with RT, and I am glad they did not panic and push it forward in the 5700XT.

QB

[NikosD]

The first leaked gaming review of 3090 is out and it is very, very bad for the flagship card.
The 3090 of 1500$ card is only 5% to 11% faster at 4K than 3080 the 700$ card which is worse than 20% that I speculated here.
The price to performance ratio compared to 3080 could be one of the worse ever for flagship cards.
That's why the objective tech journalists have already started to compare it with 2080 Ti and Titan RTX.

[videoh]

Not everyone interested in these high-end cards is a pimply gamer living in mom's basement. I haven't played a game in decades but still find the cards incredibly useful.

[SamuriHL]

AMD was always going to support it, it has been part of their "roadmap" for years... they just didn't rush it and try and gouge the consumer by promoting it early as a selling point and a must have technology. Saying AMD "figured out they needed to support it" is a little misleading and makes it sound like they are scrambling to copy NVidia. Don't forget who is working closely with MS/Sony on these consoles, it's AMD; so they have known for years that RT will be in demand for the future. RDNA2 was always going to be where AMD started with RT, and I am glad they did not panic and push it forward in the 5700XT.

QBThat's not quite what I meant. I wasn't implying they all of a sudden threw it in there last minute. I simply meant that they've yet to compete at the high end for a very long time and even they know that ray tracing is the next evolution. This has been talked about for 20 years but the tech wasn't powerful enough to go there until now. This is the very beginning of a new era. It's exciting to see these new cards and consoles support it but there's a long way to go here. Nvidia kind of blew any advantage they may have had with 2 years of Turing but there was no way they could drive an entire gaming industry in their own with over priced cards. Now the playing field is level and we'll see if amd can capitalize.

Sent from my SM-G975U using Tapatalk

[chros]

Definitely. BTW, now that I'm thinking about using PC mode again, has anyone ever had an issue with the lack of 5:5 pull down, since Real Cinema is permanently OFF?
No issues with frame rate matched content. But you won't like PC mode, no matter what output you set. :) Even ycbcr 4:4:4 8bit has posterization with this sample (https://www.avsforum.com/threads/2020-lg-cx%E2%80%93gx-dedicated-gaming-thread-consoles-and-pc.3138274/post-59700054), although half of the issues are gone that are present with rgb 12 bit.

[VBB]

chros, see what I just posted over in the OLED thread here, and let me know if you agree with my findings after you test it. Curious to hear what you guys think.

And I agree after having watched a bunch of stuff, the lack of Real Cinema in PC mode is not noticeable.

[jmone]

I wonder if they have plans for a GTX 1660Ti style of card this round (SANS ray tracing and hence at a cheaper price point)? For me the GTX 1660Ti's are a great price point for madVR.

[ryrynz]

I would very much expect so. Could be unlikely next generation though.

[*MidnightWatcher*]

I picked up a used 2070S recently for $300. Very satisfied with its performance over my 1060. I think there comes a point when diminishing returns don't justify spending more on faster cards that have virtually no visible picture quality improvements with madVR and slighter higher settings. For those looking to upgrade a lower end card for madVR, buy a used 20 series and use the money saved for something else.

[chros]

use the money saved for something else.
Good idea! That's why I won't replace my 1060 until it's absolutely necessary :D
Enjoy your new card!

[NikosD]

After reading a few custom 3080 card reviews and the leaked review gaming numbers of 3090, the picture is clear now.

Cards based on the same die of GA102 have all similar performance, no matter what.
Even 3090 is in the same ballpark and custom 3080 cards have almost no added performance to 3080 FE.
The 3080 FE cards are already pushed to their limits, they work on maximum performance.
I expect the same for 3090 custom cards.
So, nVidia is not just afraid.
They are terrified.
They are all in for the Ultimate Play against AMD GPUs.
Maybe now is the time for AMD.
They could have come so close to nVidia, probably closer than the last decade.
The next 40 days look too far away like leaks regarding Big Navi performance.
One of the best kept secrets of the industry and most anticipated ones.
They haven't said "poor Volta" or something similar this time, let's hope they are serious.
They way they handle things this time could bring a lot of frustration if RDNA2 is not competitive with Ampere.
More than ever...

[chros]

If RDNA2 won't be fast enough then they will sell it cheaper and there you go, we got a competitive product :) It's all about the price.
About nvidia, I'm still shocked about the >=320W power consumption. What will happen 2 years later?! :)

[Klaus1189]

I'm still shocked about the >=320W power consumption. What will happen 2 years later?! :)

Exactly that is driving me crazy for the past week.

And 10% plus performance of the 3090 over the 3080 for 1500€ vs. 700€ ...

[SamuriHL]

I'm not buying a 3090 that's for sure. I want a 3080 founders edition.

Sent from my Pixel 3a using Tapatalk

[huhn]

why would you want one of these terrible founders editions...

[butterw2]

Its all about available price, genuine features and stable performance, not gimmicks, marketing tricks and manufactured hype.

[SamuriHL]

Cause I don't think they're terrible. The new cooler design is effective and performance seems to be slightly above some of the tested AIB cards so far. I have a 2070 FE that I love.

[huhn]

so a cheaper card that has massively lower temps and notable lower rpm on the better cooler is not clearly better?

it'S normal for the memory to reach 104 on a 3080 FE... that 6 missing for thermal throttling while even early tested cheap AIB are at 60-70.

the 2000 series was worse you pay 100 extra for a relative bad cool and pretty much no chance to re paste the card because the construction is terrible.

there are some good niche things about them. it's easy to get a water block and the VRM are never bad for extreme over clocking. but that's it the cooler performance is underwhelming the price is a joke and the construction an insult.

[nevcairiel]

The FE with its industrial design looked good at first, and its still the best FE NVIDIA ever made, but at least the better AIB cards beat it by a rather large margin still both on temps and noise. Not sure about entry-level AIB cards.

[huhn]

asus tuf was tested and it has a MSRP at 700 (not that it means anything at the moment sadly) and it was better at pretty much every metric sometimes by quite a margin. the VRM are technically worse but that doesn't matter if you are not going to like sub zero overclocking which was for obvious reason not tested (FE could win there).

[SamuriHL]

The asus cards are on my list. Right now availability is what matters most to me. But I'm seriously considering the tuf of tuf oc if I can find one. I could end up with a zotac for all I know. Lol

Sent from my SM-G975U using Tapatalk

[el Filou]

1 Repeated Frame in 2hrs of playback (Decoding Audio and Running JRivers "Video Clock") [...] historically, Video Clock would give me 0 Dropped/Repeated frames. Maybe JRiver Video Clock needs a tweak to work a bit better with this new driver???I don't own JRiver so can't tell, but I'm using ReClock which works on the same principle, and if I install a new driver that changes the display mode timings I have to clear ReClock's timings database so it can re-learn the clock timings of the new driver. If I don't do that it doesn't sync correctly. Maybe JRiver needs something similar.About nvidia, I'm still shocked about the >=320W power consumption. What will happen 2 years later?! :)I'm not shocked, they used the biggest die usually reserved for the TITAN, they're just selling it to more reasonable price. In the past that spot in the portfolio was taken by the dual GPU cards that were sold for about the same price and had 300+ TDPs too. The better (i.e. liquid) cooled version of Vega 64 was 345W.
It's not ideal, but nothing we've not already seen. (Edit: now, of course the problem NVIDIA possibly has is performance/watt, the reviews of the 3070 will be interesting)

[chros]

Well, the previous 3 gen Titans (Maxwell, Pascal, Turimg) were 250W this is 80W more, that's huge! And the 3080 corresponds to the xx80Ti that have even less W.
I don't think that the 3070 will be weak, these cards are beasts :)

[huhn]

thanks for the 23p reports i had to two times purge the driver to get it working correctly on my system again. this should maybe be remembered so a user with bad "timing" can try to DDU to fix this.

at the second try i used the none DCH driver.

[NikosD]

I really DON'T LIKE the news/leaks/rumors regarding the upcoming Big Navi release.

It seems, according to many rumors, that second best RDNA2 based die will have 40CUs - the Navi22
This is exactly the same number of CUs as RDNA Navi10 die of 5700.

But we already know that 3080 is around 100% faster than 5700 XT

So we need what AMD promised for the Big Navi - Navi21- the x2 die of Navi10 which means 2x40 = 80CUs
This seems to be impossible to happen if Navi22 is 40CUs.
If, eventually, the Big Navi release goes like this:
1) Big Navi21 -> 60-64 CUs
2) Navi22 -> 40 CUs
3) Navi23 -> 32-36 CUs

then it's probably going to be a fiasco.

What IPC gain and clock increase is required for the RDNA2 vs RDNA in order to reach Ampere using the above number of CUs ?

It doesn't look feasible to reach Ampere's performance using this number of CUs.

[Ejziponken]

Using latest nvidia driver and the HDR isn't 100% fixed. The gui menu still takes me out of HDR without exclusive fullscreen even when set to 8 bit in madvr and nvidia.
Also Im getting new bugs like when minimizing mpc running a HDR movie, my TV gets really weird colors, until I close the app or go fullscreen again.

I tried to reinstall twice using DDU.

[JonnyVee]

I’ve been a long time MadVR user and I’ve always used it paired with a Nvidia GPU. Most recently a 2080Ti, which I sold 3 months ago while prices were still favorable.

The lack of supply for the 3080 (and overpriced 3090 and gimped 10 GB model rather than 20 GB) has me thinking of an AMD 6900XT.

I’m sure it’ll be fine for 60+ fps gaming and VR for my son — but the only thing I care about is MadVR for 1080p to 4k upscaling and 4K HDR to SDR conversion. Am I giving anything up in MadVR by going with AMD over Nvidia from a MadVR perspective?

I have a JVC x790 and 120” Stewart G4 130.

[huhn]

it's a card that does not exist so we don't know currently AMD can do about 99 % of all feature nvidia can do.

if you are using the option send BT.2020 to control the wide gamut filter then this will be lost. until madshi adds it if i can be added and that could take a while.

i would personally think more about gaming issue than madVR issues with amd.

[QBhd]

..... Am I giving anything up in MadVR by going with AMD over Nvidia from a MadVR perspective? ...

I've been a happy AMD user for... ever!

Currently rocking a 5700XT.

As huhn said, there is the report 2020 thingy, but I have no idea what it is used for and I can say my OLED paired with the 5700XT works flawlessly

QB

[NoTechi]

Am I giving anything up in MadVR by going with AMD over Nvidia from a MadVR perspective?

Same as QB and I am coming from a 1080. As for madvr I did not had to change any settings when switching to the 5700XT (I never used send Bt2020 with my JVC 7900).

Upgrading the gpu with more power for madvr might be still beneficial but I doubt it would be a big wow effect and most non pixel counting users most likely would not even recognize it :p

Next to already mentioned still supported 3D I just recently discovered that AMD driver has a build in support (ReLive for VR) to enable VR glasses like my Oculus Quest headset to play PC VR games like Half Live wirelessly which works really great and is an additional plus on the AMD side in regards to my personal needs.

If I would update it would be just for gaming and most likely it would be one of the upcoming AMDs. Being able to use higher settings in madvr would be just a side effect. But to be honest I am not in a hurry since I am really happy with my setup atm :)

NoTechi

[Klaus1189]

8K gaming :rolleyes:
https://www.youtube.com/watch?v=Xgs-VbqsuKo

[Manni]

I’ve been a long time MadVR user and I’ve always used it paired with a Nvidia GPU. Most recently a 2080Ti, which I sold 3 months ago while prices were still favorable.

The lack of supply for the 3080 (and overpriced 3090 and gimped 10 GB model rather than 20 GB) has me thinking of an AMD 6900XT.

I’m sure it’ll be fine for 60+ fps gaming and VR for my son — but the only thing I care about is MadVR for 1080p to 4k upscaling and 4K HDR to SDR conversion. Am I giving anything up in MadVR by going with AMD over Nvidia from a MadVR perspective?

I have a JVC x790 and 120” Stewart G4 130.

The issue with AMD vs nVidia isn't what it can do now, but what it can do when/if Madshi starts using tensor cores in new algos, though there is no guarantee that these will end up in madVR, or even that this will happen.

In any case, madshi has for a while been recommending nVidia RTX over GTX or AMD for this reason, and it's an RTX that's in Envy. Make what you want of this, but it's not hard to imagine that more testing/optimisation will be done with nVidia RTX than with anything else.

So depending on how long you want to keep your GPU and what ends up being supported in madVR that might take advantage of tensor cores, going for AMD might be a problem, or not.

The BT2020 flag is a far bigger issue now, but that's only relevant if you calibrate and need madVR to tell an HD Fury which calibration it should select automatically in your projector/display, or if your projector/display needs this info to switch to the correct mode when tonemapping HDR to SDR.

I've resolved the gimped 10Gb by trying an EVGA 3090 XC3 (delivery expected on the 5th of October), but if I'm not happy with it or if it doesn't fit in my case, I'll wait until the 20gb 3080 (ti/super/whatever) is available.

It should cost around 1000-1200, which isn't that much cheaper than the $1500 for the 3090, which I'll be using for other stuff than madVR or gaming by the way. There is no way I'll get a 10gb 3080, or anything less powerful than a 3080 to replace my 11gb 1080ti. I'm happy to stay with my 1080ti for as long as necessary...

[huhn]

there seems to be a major issue in the current driver i'm currently figuring out if it is general or only affecting some GPU like pascal.

DDU or clean install could be a major issue with that driver.

[NikosD]

Using Tensor cores for madVR sounds like a joke.
A bad joke actually.

[huhn]

using cores for neural network operation in a video render full with neural network scaler is a bad joke...

edit: about the bug. the following thing is what i encounter and i sadly need other user to reproduce it to confirm it.

i have the following issue with 456.38 the power management mode is ignored so if you are using ddu or use the clean installation option in the driver you will be stuck with optimal power.

so how did i "proof" this.
i contacted nvidia support because my GPU was stuck at 800 mhz with madVR playback the usually issue with optimal performance even through i changed this setting.
i just followed his orders by installing an older driver and setting some very question settings in the nvidia control panel and "prefer maximum performance" as he ordered. after that "worked" i was supposed to install the newest driver again. and it works now but i'm stuck at maximum performance the GPU is idle at 1.5 GHZ.

so what i like to get a report from some user here is if maximum performance works on your system and the idle GPU clock is int he GHZ range if your driver was not at maximum performance before.

if you would excuses me i have to install 452.22 set it to adaptive and install 456.38. "fun"

[chros]

8K gaming :rolleyes:
https://www.youtube.com/watch?v=Xgs-VbqsuKo
:D I saw that as well, it's a big fat joke just as we thought.

[NikosD]

... it's a big fat joke just as we thought. We ? Who's we ?

[huhn]

pretty much everyone here sad just ignore it is just expensive and has use cases for professionals that's it.

[NikosD]

3090 is a disaster.

It needs around 375W (custom 3090 is over 400W) and 1500$ to be 10% faster than 3080 of 700$ card in 4K gaming.

If 4K is a niche for most people, then 8K gaming is extra-terrestrial right now.
No monitors and most games are not ready for this resolution.

8K gaming is feasible mainly using DLSS enabled games using this new "Ultra performance" mode, rendering the game in 1440p and upscaling to 8K using 9x (!) upscaling.

Also, it's not a professional card.
Very far from it.
Steve from Gamers Nexus is not right on this one.

It's a lot slower than Turing TITAN RTX in applications like CAD, Design etc.
nVidia didn't enable driver optimizations of Quadro cards to a Geforce card like 3090, as it did with Titan series.

The result is a much slower card even than Turing Titan RTX in these cases.

Also, 3090 card has very slow FP64 performance and doesn't support GPU virtualization/sharing using SR-IOV.

nVidia probably reserved these things for Ampere Titan of 3000$

So, no.
Most of the users of this forum were wrong.
3090 is not a replacement of Titan cards.

It could be used only by content creators using Blender or other rendering software or video editing of 8K or any productivity app workload that can't fit in 10GB of 3080 VRAM.
Because the chip and the drivers are exactly the same for 3080/3090, with disabled features of real professional SW.

Prosumer Titan and professional Quadro users doing CAD, Design, AI Training should wait for 3000$ Titan Ampere or Quadro Ampere to do their job.

So if 3090 doesn't worth it for gaming and doesn't have professional/ workstation drivers for what is it worth ?

Ask the leather-jacket-man.

Or even better -gamers, prosumers, professionals could wait for RDNA2 cards.
I hope this time AMD will grab the opportunity.

[Polopretress]

using cores for neural network operation in a video render full with neural network scaler is a bad joke...

edit: about the bug. the following thing is what i encounter and i sadly need other user to reproduce it to confirm it.

i have the following issue with 456.38 the power management mode is ignored so if you are using ddu or use the clean installation option in the driver you will be stuck with optimal power.

so how did i "proof" this.
i contacted nvidia support because my GPU was stuck at 800 mhz with madVR playback the usually issue with optimal performance even through i changed this setting.
i just followed his orders by installing an older driver and setting some very question settings in the nvidia control panel and "prefer maximum performance" as he ordered. after that "worked" i was supposed to install the newest driver again. and it works now but i'm stuck at maximum performance the GPU is idle at 1.5 GHZ.

so what i like to get a report from some user here is if maximum performance works on your system and the idle GPU clock is int he GHZ range if your driver was not at maximum performance before.

if you would excuses me i have to install 452.22 set it to adaptive and install 456.38. "fun"
I confirm you that a contact of mine who has tried the 456.38 has the same behavior.
No management of the selected power mode. only stuck at the last value applied with the previous driver version.
(where is the validation team and no regression tests database on nvidia side ? ...)

It seems also that CRU is not working anymore (by the extensions blocks and displayID )

[Manni]

@NikosD

Any chance you could calm down a bit and stop the misinformation?

Custom 3090 GPUs don't all need more than 375W. Apart from the fact that I like EVGA support and 3 year warranty, I chose the XC3 because it only needs 2 8-pin power and draws a max of 375W. It's also smaller than the 3090FE, so I hope that it will fit in my case.

375W is in full load. It will draw a lot less otherwise.

Why are you so worked up about companies strategies? They are not your friend or mine. There is nothing personal. They have a single goal: maximising profits for their shareholders. AMD is no more friendly than nVidia. Who cares who is making a GPU, as long as it's the best choice for your needs?

Anyway, re your other points, I don't really play games, I don't have the time, though I occasionally play fairly old games with my daughter. Apart from madVR, I plan to use the 3090 with Flight Simulator 2020 (in 4K) and for video editing and post-production work in 4K/8K workflows (I use studio drivers). Both applications need more than 10Gb, so there are definitely some use cases for it, even if it's often more a want than a need.

If people are stupid enough to buy them to play games in "8K", then it's their problem, but I don't think many do that. Currently the 3090 is the best solution for my needs, especially as I don't upgrade often. I bought my 1080ti in 2017, I have zero interest in Turing due to the lack of HDMI 2.1 and I'm hoping to keep the 3090 (if it fits in my case) for at least 2-3 years. AMD is not even on the table due to the lack of tensor cores and BT2020 flag support, so it's nice to not even care about what they might release next month. They could release a card more powerful than the 3090 for half the price, they would still not get my money :)

When/if madVR starts using tensor cores, I want to be able to make the most of it, instead of having to upgrade once more. And being able to set the BT2020 flag is a requirement for me. I don't switch gamut or calibration manually in my cinema room, sorry. I start a film with my remote control or my iPad, and HD Fury does the rest.

Sure, I'd rather buy a 3080ti with 20Gb for 30% less money, as that would suit my needs just as well, but we don't know when it will materialise, in which quantity, and the time I waste looking on forums and resellers to find out or secure a new GPU IS money.

On the other hand it's quite nice to have the flagship with full bus and full memory (until Ampere Titan shows up, but I have no interest in that) and to stop thinking about upgrading for a while. I really don't care about overclocking, getting top scores at benchmarking, water-cooling etc.

I managed to snatch a pre-order for the 3090 during the 1-2 hour window it was available (entirely by luck, I thought I'd try once on the 24th). I was able to place an order for the model I wanted - EVGA XC3 Gaming - at the price I was expecting to pay, knowing that I can return it if it doesn't fit my case, so I did it.

My situation is slightly different than most though because it's a business purchase as I use it primarily for work, both with madVR and with video post-production. I get 20% VAT back and I save 20% taxes on company profits, so it costs me in effect 40% less than the full price. It's still expensive, but as a business expense, it's not that bad compared to a high-end laptop, projector or an AVR. I certainly wouldn't recommend a 3090 to anyone who doesn't need one and can't wait for the 3080 20Gb. I'm curious to see how much my CPU will be a bottleneck for my use case, I expect it will be. I've budgeted an upgrade to Rocket Lake next year if necessary. Again, I don't upgrade often, and I've been waiting for Intel to implement PCI 4.0 to do so, only if I need to.

Buying a 3080 with 10Gb for anything but 4K video feels short-sighted to me, many games already need more than that, so to me it's a worse choice than the 3090, even if the 3090 completely loses on the value for money side.

But we're all different, with different needs, priority, preferences and budgets.

Couldn't you accept that what's good for you might not be best for someone else, and vice-versa, and please could you stop the AMD fanboy rants?

In any case, I suggest that you start a new thread if you really want to keep going that way, because this brand war has nothing to do with the topic of this thread, and it's getting boring, frankly.

I subscribed to this thread to follow potential driver issues, not to hear fanboys ranting about their most-hated or favorite brand, or to have to justify my purchase. If AMD was giving me what I wanted, I'd most happily buy AMD. I had a Sapphire 7870 before the 1080ti and I was very happy with it (until they changed their drivers and made changing advanced video settings impossible, but that's another story).

I also tried a few AMD GPUs for my EGPU as they would be a better choice on my Macbook Pro when using MacOS, but they were all pants in Bootcamp so I sent them back.

Anyway, thanks for considering giving us a break!

[NikosD]

@Manni

I don't know if this thread is suitable for GPU discussions but the last days I have clearly stopped writing a lot of posts and certainly my posts are not so huge like yours.

You are definitely an anti-AMD guy even considering buying Rocket Lake on 2021, a non-existent CPU not mentioned anywhere officially and you didn't even mention Zen 3.

You are an Intel fanatic and anti-AMD fanatic.
That's why you cared so much replying to me with a huge post.

Also, a rather large part of your post is trying desperately to justify the worse buy ever - the 3090 card.

Are you serious saying to people that you bought a 3090 card for madVR and business ?
Are you serious that Tensor cores could be used outside very specific purposes like DLSS from Geforxe RTX cards ?
Who is saying such garbage ?

You are definitely misleading people by justifying the huge power consumption of 3080/3090 cards.

There are custom 3090 cards with 420W TDP.
Do you know what you are talking about ?

The strategies of companies like Intel, nVidia, AMD have serious impacts in our pockets and the industry in general.

We have free market system and everyone can burn his money to garbage like 3090 cards - no doubt about it.

But we have to say loud and clear that the only reason to justify buying a 3090 cars is just this - because you have money.

Posts like this one above make me wanna write even more about this subject.
Posts from nVidiatel fanatics.

[Manni]

All right, I thought there was hope with you, I was wrong. :)

I'm out of this discussion. And yes, I work in film and as a consultant for madVR Labs (though what I post here are my words), so the 3090 is 100% a business purchase. MS FlightSim 2020 is just for fun at the end of the day :)

EDIT: as for Intel vs AMD, when AMD starts supporting Thunderbolt in more than a couple motherboards, I might consider it. In the meantime, for video editing, especially with After Effects, a threadripper is a waste of money and Intel provides better performance at the same price point. I don't expect this to change with Rocket Lake (https://hexus.net/tech/news/cpu/144445-intel-rocket-lake-system-pcie-40-nvme-ssd-spotted/), but if it does, I'll reconsider. I use Intel because it's better for my needs, not because it's Intel. I mention Rocket Lake because I want (for my needs) PCI 4.0 and Thunderbolt 4.0, and as far as I know only Intel will provide that next year.

[Asmodian]

Posts from nVidiatel fanatics.

Why do you assume people have to be fanatics if they disagree with you? (Projection?) I really do not think of any of these companies as entities to be fans of. The idea seems weird to me, billion dollar corporate entities are not something that I can be a fan of.

I am buying a 3090 asap, but for no good reason. A 3080 is a much better buy for my use cases. If big Navi turns out to be better than the 3090 I will buy one.

Anyone without an AMD tattoo seems anti-AMD to you. :rolleyes:

[Manni]

Where is the like button when you need it :)

[Klaus1189]

https://www.youtube.com/watch?v=7YQ7rNgoqMA

[videoh]

If you are going to link to Germish stuff, how about an executive summary in English? The language for this forum is English.

[Klaus1189]

See in the description of the video:
https://www.igorslab.de/en/what-real-what-can-be-investigative-within-the-crashes-and-instabilities-of-the-force-rtx-3080-andrtx-3090/

[videoh]

Thank you, Klaus.

[SamuriHL]

huhn you asked why I'd even consider an FE over an AIB. Given the latest reports out there, hopefully it's painfully obvious now. The AIBs, at least some of them, it looks like they cheaped out on capacitors and the crash to desktop problems we're hearing about MAY be caused by voltage problems on those cards when they boost. nVidia used the good capacitors. ASUS did, as well, which is the only other card I was even looking at. And...I think I'll stick to that. LOL Drivers are also not super awesome for these cards combined with the problems you've found about power settings sticking. This launch feels rushed even though it shouldn't have been. We know why that is, but, nVidia really didn't have any room for a screw up here. Hopefully these issues get sorted out.

[Asmodian]

I think companies simply use release as a QA step today. In house QA at that level would be expensive, digital updating is cheap, and everyone who bought one can do it so the PR isn't too bad. Everyone who can does it now. :p

[ryrynz]

This launch feels rushed even though it shouldn't have been. We know why that is

Yeah Nvidia was scared of big Navi [emoji23]

[huhn]

huhn you asked why I'd even consider an FE over an AIB. Given the latest reports out there, hopefully it's painfully obvious now. The AIBs, at least some of them, it looks like they cheaped out on capacitors and the crash to desktop problems we're hearing about MAY be caused by voltage problems on those cards when they boost. nVidia used the good capacitors. ASUS did, as well, which is the only other card I was even looking at. And...I think I'll stick to that. LOL Drivers are also not super awesome for these cards combined with the problems you've found about power settings sticking. This launch feels rushed even though it shouldn't have been. We know why that is, but, nVidia really didn't have any room for a screw up here. Hopefully these issues get sorted out.

yes i still do the 3080 FE is part of this problem and the 3080 FE is so bad that it will thermal throttle with let's say furmark that's not the case with a decent AIB.

there where AIB so terrible that they would die if you flash a stock bios on them and run furmark and just to make that absolutely clear i said die not maybe die.

clock rate are currently taking into account why the card is crashing. AIB cards often reach higher boost clock then FE cards just by been massively cooler with the same bios it is as it is.

so just lay back wait for some test it's sadly not unusual for stuff like this to happen. someone wants an exploding EVGA card even through the VRM are totally fine? there where and there will always be terrible AIB cards too.

[nevcairiel]

There is no real information on the crashing issue at all currently. It might be driver. People might be running their PSUs close to their limit. AIB boards may be OCed a tad too aggressively and a BIOS update will sort that out. The capacitor theory is just that, a theory. I see no conclusive tests in any of that reporting.
Some AIBs are also reporting that they actually changed the capacitor layout in a final board revision to improve the situation, which might not be the one that was sent to reviewers.

The FE on the other hand is reportedly running quite hot and more noisy then good AIB cards. But I suppose if you were happy with previous FEs, this is still an upgrade over that.

And I wouldn't say the release was necessarily rushed by NVIDIA. The problem is the stringent secrecy, which didn't give the AIB much time to do full validation. So AIB cards might be rushed, because they wanted to be part of day 1 sales, but the reason for that being getting drivers, for example, only days before the actual launch.

[NikosD]

as for Intel vs AMD, when AMD starts supporting Thunderbolt in more than a couple motherboards, I might consider it.

In the meantime, for video editing, especially with After Effects, a threadripper is a waste of money and Intel provides better performance at the same price point.
I really do not think of any of these companies as entities to be fans of. The idea seems weird to me, billion dollar corporate entities are not something that I can be a fan of.

I am buying a 3090 asap, but for no good reason. Thank you.

[Manni]

Thank you.

You're welcome :)

[Manni]

As nobody seems to mind the off-topic posts, an interesting review of the 3090 (no idea if it can be trusted) seems to show that, as expected, it's more efficient compared to a 3080 (2-4%), 2080ti, Titan RTX and RX5700XT (20-33%!), despite its higher TDP: https://www.eurogamer.net/articles/digitalfoundry-2020-nvidia-geforce-rtx-3090-review

****

Watts Per Frame RTX 3090 FE RTX 3080 FE RTX 2080 Ti FE Titan RTX RX 5700XT
Death Stranding 3.209 3.349 3.562 3.670 3.846
Percentage Diff 100% 104.4% 111.0% 114.4% 119.9%
Watts Per Frame RTX 3090 FE RTX 3080 FE RTX 2080 Ti FE Titan RTX RX 5700XT
Gears 5 4.055 4.156 4.266 4.291 5.403
Percentage Diff 100% 102.5% 105.2% 105.8% 133.2%
"The results here are fascinating. Despite sporting a higher TDP, the 3090 actually consumes fewer watts to generate a single frame than the 3080 in the titles we tested: In Death Stranding, the RTX 3080 requires four per cent more power per frame, and this drops to just under three per cent in Gears 5. This may sound implausible bearing in mind the beastly nature of the RTX 3090 - but the new card's total graphics power is 12.5 per cent higher overall and it is capable of a performance bump in excess of that number."

****

If confirmed, this should be the case for madVR too: with the same settings, the 3090 should be more efficient (though a 3080 or lower remains a more sensible choice for most use cases obviously). Now of course if you pump everything up to the max, it won't be :)

I'm interested to see what I can get with 4K60p HDR with the 3090/3080, and even more so when we'll be able to send this in 4:4:4 8 or 10/12bits to an HDMI 2.1 display / projector, as well as what it can do with poor content such as upscaled SD/HD.

As some in this thread seem to have trouble understanding what I write, I'll state again that I don't advise to buy a 3090 for madVR or games. For madVR, a 3080 10Gb is a far better choice at the top end (anything lower according to your budget will do too). For high-end games, the upcoming 3080 20Gb might be a better long term investment than the 10GB version, as quite a few games already make use of more than 10Gb of RAM. If a 3080 20Gb was available to buy today, that's what I would get, not a 3090, as it would give me all I need for around 30% less money (guesstimate). That's what I will be buying in November, when it's most likely released, if I'm not happy with the 3090.

If you prefer AMD for whatever reason, by all means buy an AMD GPU, I couldn't care less, but don't complain if later down the line madVR starts using tensor cores for new algos and you can't make use of that extra power!

[NikosD]

If you prefer AMD for whatever reason, by all means buy an AMD GPU, I couldn't care less, but don't complain if later down the line madVR starts using tensor cores for new algos and you can't make use of that extra power! It started as a joke but since you mention it for a second time, which is the API inside Geforce drivers that exposes tensor cores to be leveraged by a developer outside of DLSS for games ?

[Manni]

It started as a joke but since you mention it for a second time, which is the API inside Geforce drivers that exposes tensor cores to be leveraged by a developer outside of DLSS for games ?

I have no idea and I honestly don't care. Ask madshi or nVidia :).

The only thing I can tell you (again!) is that madshi has, over the last couple of years, publicly stated that he planned to make use of tensor cores in the future for madVR/Envy algos. I work as a consultant for madVR Labs, and as far as I know this is still his intention and it has been stated publicly a few times.

For this reason, he has informally advised anyone asking him the question in forums to buy RTX over GTX or AMD if they wanted to future-proof their purchase, though there is still no commitment at this stage that 1) he will make use of tensor cores in madVR/Envy and 2) that even if he does end up using tensor cores in Envy, this will translate to madVR. His focus is currently on finalizing Envy, not on new algos, so we won't know for a while.

However, what we know is that the GPUs in Envy are all RTX and will remain RTX in the foreseeable future. The GPU in the top of the range Envy Extreme will be a 3080 10Gb whenever stocks can be secured. As a result, there will clearly be more testing going on with RTX than with anything else. There are already some features (such as supporting the BT2020 flag to signal SDR BT2020 content for automatic calibration switching or HDR/SDR switching through API) that is only supported with nVidia GPUs.

These are facts. If you don't like them, there is nothing I can do. I am not posting here for you, I'm posting here for everyone else.

If you still want to buy an AMD GPU for madVR use, why should I care? It's your money, your choice, your decision.

I only provide information so that buyers make an informed decision. MadVR use is, for many, only one factor in their decision. They might have very good reasons to buy AMD, or even Intel.

As I said, we are all different, with different preferences, priorities and budget.

This is my last reply to you. I am not pro nVidia, pro Intel or anti-AMD. I don't care about these companies. I don't own shares in any of them. In this thread, I don't represent madVR Labs, only myself. I am not sharing any insider info, only what madshi has publicly stated in many occasions, or what has already been announced in press releases.

As a consumer, I am pro-me and I buy what suits my needs, my priorities, my budget and my personal preferences. I don't expect anyone to have the same as mine.

I only expect everyone with a brain to do the same, and make their decisions based on THEIR needs, priorities, budget and preferences :).

[el Filou]

It started as a joke but since you mention it for a second time, which is the API inside Geforce drivers that exposes tensor cores to be leveraged by a developer outside of DLSS for games ?Do you by any chance know about something called CUDA? It's an API provided by NVIDIA to develop natively for their hardware. (I'm sure you also know about something called Google that would give you the answer to the question you're desperately trying to troll with)

[NikosD]

I have no idea and I honestly don't care. Ask madshi or nVidia :).

For this reason, he has informally advised anyone asking him the question in forums to buy RTX over GTX or AMD if they wanted to future-proof their purchase, though there is still no commitment at this stage that 1) he will make use of tensor cores in madVR/Envy and 2) that even if he does end up using tensor cores in Envy, this will translate to madVR.

These are facts. If you don't like them, there is nothing I can do. I am not posting here for you, I'm posting here for everyone else.

I expect everyone with a brain to do the same :). Tensor Cores are out in the wild for three years now.
Do you know any software outside of DLSS and games leveraging Tensor Cores using a Geforce card ?
So, the fact is that madVR labs even in partnership with nVidia hasn't provided yet a single piece of code demonstrating that is feasible to use Tensor Cores for madVR.
Buy you are still expecting people with brain, as you said, after years of promises to buy RTX cards because of some future, nebulous use of Tensor Cores for madVR ?

OK...

[Manni]

Tensor Cores are out in the wild for three years now.
Do you know any software outside of DLSS and games leveraging Tensor Cores using a Geforce card ?
So, the fact is that madVR labs even in partnership with nVidia hasn't provided yet a single piece of code demonstrating that is feasible to use Tensor Cores for madVR.
Buy you are still expecting people with brain, as you said, after years of promises to buy RTX cards because of some future, nebulous use of Tensor Cores for madVR ?

OK...

Can you please stop putting words in people's mouths?

There is no madVR Labs / nVidia partnership.

madshi has not even STARTED to look into making use of tensor cores.

He has only advised, over the last couple of years, that it was his INTENTION to do so at some point, but never committed to do it.

There has been no work on new algos over the last couple of years, except improvements to HDR to SDR tonemapping. As I said, and as everyone knows, madshi's focus has been on releasing Envy, not on developing new algos with tensor cores...

I'm only making people aware of this INTENTION with no commitment, and you turn it into a FAILURE over THREE YEARS?

Do you realise how twisted your post is? Otherwise, you really need professional help.

Whatever nonsense you decide to post, I will not reply.

[NikosD]

There is no madVR Labs / nVidia partnership. There is no other way for such a difficult task than get in touch with nVidia.
It's that simple and it's not something to be embarrassed of.
madshi has not even STARTED to look into making use of tensor cores.

I'm only making people aware of this INTENTION with no commitment, and you turn it into a FAILURE over THREE YEARS? You must stop putting words in others people mouth.
I've never used the word FAILURE over THREE years for madVR intention of using tensor cores.
I said that tensor cores are out for three years and there are years of promises/rumors/intensions - call it whatever you want - around it.
Do you realise how twisted your post is? Otherwise, you really need professional help. Judging from your first huge post and how you followed up next, I would advise you the same.
Whatever nonsense you decide to post, I will not reply. You have said that before in this thread. Promises, promises, promises...

[huhn]

what is this:
https://github.com/tensorflow/tensorflow/tree/v2.3.1

it links to this:
https://docs.nvidia.com/deeplearning/frameworks/tensorflow-user-guide/index.html

The TensorFlow container includes support Tensor Cores starting in Volta’s architecture, available on Tesla V100 GPUs. Tensor Cores deliver up to 12x higher peak TFLOPs for training. The container enables Tensor Core math by default; therefore, any models containing convolutions or matrix multiplies using the tf.float16 data type will automatically take advantage of Tensor Core hardware whenever possible.

and this also exist:
https://github.com/igv/FSRCNN-TensorFlow

does this prove to be using tensor cores no i have to say to be fair.

[SamuriHL]

There is no real information on the crashing issue at all currently. It might be driver. People might be running their PSUs close to their limit. AIB boards may be OCed a tad too aggressively and a BIOS update will sort that out. The capacitor theory is just that, a theory. I see no conclusive tests in any of that reporting.
Some AIBs are also reporting that they actually changed the capacitor layout in a final board revision to improve the situation, which might not be the one that was sent to reviewers.

The FE on the other hand is reportedly running quite hot and more noisy then good AIB cards. But I suppose if you were happy with previous FEs, this is still an upgrade over that.

And I wouldn't say the release was necessarily rushed by NVIDIA. The problem is the stringent secrecy, which didn't give the AIB much time to do full validation. So AIB cards might be rushed, because they wanted to be part of day 1 sales, but the reason for that being getting drivers, for example, only days before the actual launch.The drivers not getting released until day of launch was weird. It will be interesting to see what the next version fixes / breaks. Lol

As for the capacitor theory, yes it's a theory but it also makes sense given how much power these cards are consuming. You are right that the aib cards are driving higher than the fe. And it's been theorized that a vbios update to drop the max speed even 15 to 20 mhz might do the trick. Yes all still theory. Especially since I can't seem to get my hands on a card yet anyway. Lol I'm sure by the time I can they'll have this figured out.

I would love to get my hands on one of the asus cards and would gladly take one of those over the fe but I've not seen those anywhere yet. I'll keep looking until they become more widely available.

Sent from my SM-G975U using Tapatalk

[NikosD]

I'm reading a lot of rumors regarding ray-tracing performance of Turing/Ampere vs RDNA2 cards and I have to report a big-big-big disclosure regarding RT cores or ray-tracing cores of RTX cards.

Contrary on many beliefs, they do not exist in the form you think they do, meaning that they are not actually what most of the people believe:
An ASIC, fixed-function HW unit fully independent and dedicated on ray-tracing, like video HW decoders/encoders inside the GPU.

RT cores are actually an optimized, low-latency datapath inside the GPU heavily dependent from the well known CUDA cores, not only for hybrid RTX code like most of the RTX games, but even for fully path-tracing games like Minecraft and Quake II RTX.
The dedicated silicon for ray-tracing only calculations is desperately small.

The RT performance follows the classic CUDA core performance in both Turing and Ampere cards, it's not based on a separate block of fixed-function HW doing fully independently and separately ray-tracing calculations.

It's a little, dirty secret of nVidia that doesn't want the world to know, but I decided to disclose it here in my doom9 forum friends.

Don't tell it to the others!

[nevcairiel]

Your "secret" information is part of every official NVIDIA block diagram of both Turing and Ampere.
There is one RT core as part of every SM (Streaming Multiprocessor), which is the unit that contains the CUDA cores (128 per SM on Ampere), the Tensor cores (4 per SM), and also the RT core (1 per SM). Their position inside the SM very closely coupled to all other math functions clearly defines their role. Its not a secret. They don't have to be very huge, similar to Tensor cores, if you make something designed specifically for a certain type of math, it can do this much faster then eg. the CUDA cores.

A 3080 has 68 SMs, which makes for 68 x 128 CUDA cores, 68 x 4 Tensor Cores, and 68 RT cores.

NVIDIA even has block diagrams of what the RT core does internally, its not secret magic. It just does BVH traversal (eg. finding out which object is in a certain space), and ray intersection. Relatively simple math, which is just expensive to run on non-specialized hardware.

[tyguy]

I was told that nvidia changed their hdr implementation and madvr needed to be updated for it to not have a washed out picture. So today I loaded up a movie and forgot to change to os hdr and noticed the picture looks great again. Did nvidia change their hdr implementation once again?

[Klaus1189]

Look on post 2 ;)

[nevcairiel]

I was told that nvidia changed their hdr implementation and madvr needed to be updated for it to not have a washed out picture. So today I loaded up a movie and forgot to change to os hdr and noticed the picture looks great again. Did nvidia change their hdr implementation once again?

The latest driver has a workaround that allows it to work again.

[Manni]

As for the capacitor theory, yes it's a theory but it also makes sense given how much power these cards are consuming. You are right that the aib cards are driving higher than the fe. And it's been theorized that a vbios update to drop the max speed even 15 to 20 mhz might do the trick. Yes all still theory. Especially since I can't seem to get my hands on a card yet anyway. Lol I'm sure by the time I can they'll have this figured out.

I would love to get my hands on one of the asus cards and would gladly take one of those over the fe but I've not seen those anywhere yet. I'll keep looking until they become more widely available.


Yes if the theory is true, Asus seems to have built their cards above the specs.

Re EVGA, I'm glad I went for the 3090 (2 MLCCs out of 6 (https://i.imgur.com/20VTCrx.jpg), same as the FE) vs the 3080 (1 MLCC out of 6). See here for a full table per AIB (https://new.reddit.com/r/nvidia/comments/izhpvs/the_possible_reason_for_crashes_and_instabilities/g6jivta/?utm_source=reddit&utm_medium=web2x&context=3).

That's if Igor's theory for the desktop crash is true of course. It has been confirmed by EVGA (https://imgur.com/1l3PTBN), so it's most likely true, though it looks like only reviewers got cards with 6 POSCAPs.

I never overclock, but I'll report how my 3090 XC3 does when tested with stock settings (if it fits in my case!).

Hopefully by the time they release the 3080 20Gb the reason for this will be known and the issue will be addressed.

[claw]

Yes if the theory is true, Asus seems to have built their cards above the specs.

Re EVGA, I'm glad I went for the 3090 (2 MLCCs out of 6 (https://i.imgur.com/20VTCrx.jpg), same as the FE) vs the 3080 (1 MLCC out of 6). See here for a full table per AIB (https://new.reddit.com/r/nvidia/comments/izhpvs/the_possible_reason_for_crashes_and_instabilities/g6jivta/?utm_source=reddit&utm_medium=web2x&context=3).

That's if Igor's theory for the desktop crash is true of course. It has been confirmed by EVGA (https://imgur.com/1l3PTBN), so it's most likely true, though it looks like only reviewers got cards with 6 POSCAPs.

I never overclock, but I'll report how my 3090 XC3 does when tested with stock settings (if it fits in my case!).

Hopefully by the time they release the 3080 20Gb the reason for this will be known and the issue will be addressed.

You might want to check this thread on the EVGA forum concerning performance in games. Don't know if it would affect MadVR performance as well. Thread title is 3080 but there is one poster who said it affects 3090 as well. Maybe an issue with the new driver and a workaround.

https://forums.evga.com/EVGA-XC3-ULTRA-3080-UNDERPERFORMING-m3096945.aspx

I have an EVGA 2080 Ti so I don't intend to upgrade unless a 3080 20G is helpful with any new algorithms Madshi might introduce.

[SamuriHL]

Yea, I'm inclined to believe there's at least some truth to the capacitor issue. And yes, Asus tends to engineer very good stuff. I'm leaning more and more towards an Asus. The FE, as Nev pointed out, is not able to boost as much because of thermal issues. Not that I OC, either, but if a card is CAPABLE of doing so then running at normal clock rates means it'll be quite a bit cooler (and maybe quieter?). My 680 was an ASUS and it lasted a *LONG* time until I replaced it with a 1060. The 2070 FE has been a great card for me which is why I was considering the 3080 FE. However, I've been convinced that Asus has their stuff together as usual. And with these cards pushing this much power, engineering matters more than ever.

[SamuriHL]

madshi has said memory won't be a problem which is why the Envy is going with 3080 10gb. The memory will be an issue for gaming, but, not madvr.

[Manni]

You might want to check this thread on the EVGA forum concerning performance in games. Don't know if it would affect MadVR performance as well. Thread title is 3080 but there is one poster who said it affects 3090 as well. Maybe an issue with the new driver and a workaround.

https://forums.evga.com/EVGA-XC3-ULTRA-3080-UNDERPERFORMING-m3096945.aspx

I have an EVGA 2080 Ti so I don't intend to upgrade unless a 3080 20G is helpful with any new algorithms Madshi might introduce.

Thanks, I use Studio drivers anyway, so it looks like the Precision X1 settings resolves this when/if it happens.

Note that I didn't go for the ULTRA because I don't want any factory overclock, or the FTW3 because of the form factor. I went for the XC3 gaming (no OC, but better binning than the black).

I'll report back any issue. I upgraded my PSU to 1000W a while back to prepare for my GPU upgrade. I'm glad I decided to overspec it, at least I should be fine on that front.

20GB is a waste of RAM/Money for video playback. There is zero need to go above the 3080 10Gb for madVR use, and even the 3080 power headroom over the 2080 super in Envy won't be used for a while, unless you stack multiple algos at 4K60p. So if you decide to upgrade, save some $ and get a 3080 10Gb.

If you don't need HDMI 2.1, you're good to go with your 2080ti for a good while, at least until prices stabilize and availability becomes a non-isssue. You might even skip a gen entirely. :)

[ryrynz]

Imagine buying a 30 series for potential tensor core functions with MadVR... Nobody is doing this. It's all about the raw performance and improved thermals/noise levels for the workload they require. I believe everything ever stated as being in MadVR 1.0 is now frozen, in the distant future we may see some more but given what I've see happening I don't expect much. The goals appear quite clear.

[NikosD]

Your "secret" information is part of every official NVIDIA block diagram of both Turing and Ampere. Can you share an official link showing the block diagram with RT cores which are placed inside SMs ?
I have seen and know very well the BVH diagram.
I want to see the building block diagrams describing RT cores as one per SM inside it.

[nevcairiel]

Techpowerup, for example, has it in their architecture rundown. The diagrams shown are officially from NVIDIA.
https://www.techpowerup.com/review/nvidia-geforce-ampere-architecture-board-design-gaming-tech-software/3.html

Full block diagram (red parts are disabled in a 3080, in comparison to a full Die):
https://tpucdn.com/review/nvidia-geforce-ampere-architecture-board-design-gaming-tech-software/images/rtx-3080-block-diagram.jpg

Couldn't find a proper high-res version in this second, but the yellow RT cores can be seen in every SM.

Single SM, build up of 4 warps with 32 CUDA Cores and 1 Tensor core each, and the single RT core.
https://tpucdn.com/review/nvidia-geforce-ampere-architecture-board-design-gaming-tech-software/images/ampere-sm.jpg

If you want even more official, this Whitepaper about Turing contains a rather similar block diagram, just for Turing instead of Ampere, but the location of the RT core is the same:
https://www.nvidia.com/content/dam/en-zz/Solutions/design-visualization/technologies/turing-architecture/NVIDIA-Turing-Architecture-Whitepaper.pdf

[huhn]

is there a diagram showing them outside?

https://www.nvidia.com/content/dam/en-zz/Solutions/design-visualization/technologies/turing-architecture/NVIDIA-Turing-Architecture-Whitepaper.pdf

Ray tracing acceleration is performed by a new RT Core processing engine within each SM (RT
Core and ray tracing features are discussed in more depth in Turing Ray Tracing Technology
starting on page 25).

more details are on 30/31.

[NikosD]

The message was never communicated like the post above by me.

The message from nVidia and tech journal at Turing release was as they managed to accelerate in HW and in real-time the ray-tracing procedure .

And the message still remains the same, even today.

In terms of video decoding acceleration, the implementation of RTX ray-tracing is what we call a very partial ray-tracing acceleration, a HW assisted implementation.

It implements in HW a tiny, little part of the ray-tracing algo, resulting a performance level, deeply dependant and proportional by the shader performance.

This is not a HW accelerated ray-tracing implementation.

It looks like the good old days of GPU assisted video acceleration, by shaders or small dedicated HW not able to accelerate the whole video decoding pipeline to make an analogy, not the nowadays of fully hardware accelerated video decoding/ encoding.

It was and still is a rushed move just to justify the price of the whole RTX 2000 series and especially asking 1200€ for 2080 Ti with a big performance hit due to the partial acceleration.

Take a look at the size of RT cores in the photos of the block diagram of SMs
It looks huge compared to CUDA cores, while we know it's a tiny little silicon inside.

Ampere improved the ray-tracing acceleration mainly because of faster shader performance (~70% for games at 4K vs 2080) and only around 10% for ray-tracing HW acceleration by itself.

This is the message that I want to make it clear.

[butterw2]

When any discrete gpu can handle 1080p60 gaming, but you still want to sell expensive gpus, what are you to do ? :)

RT Cores, Nvidia RTX ON, always sounded like marketing BS to me.

I wish they would at least bring some efficient/affordable gpus to the market.

For GPU accelerated machine learning training, however, my understanding is that Tensor Cores and VRAM quantity are actually useful.

[Klaus1189]

https://www.youtube.com/watch?v=RlDC7u7GpR4
and following link also in the description of the video - the english "review"
https://www.igorslab.de/en/nvidia-geforce-rtx-3080-und-rtx-3090-and-the-crash-why-the-capacitors-are-so-important-and-what-are-the-object-behind/

[Klaus1189]

https://www.youtube.com/watch?v=THMukcOzB8g

[NikosD]

It is now reported by many HW review sites, that nVidia decided to rush in an extreme way the release of 3080/3090 cards by pushing the AIBs to their limit.

They got the drivers in the last minute and they had time to test only power on/off and basic stability issues.

The AIB cards used the owners/ buyers of 3080 cards like guinea pigs for real workloads and games, mainly because of nVidia's aggressive behavior against them.

But after the last leaks it's time to forget all these and leave them behind, in the past.

We have to speak about the elephant in the room - Big Navi21

It seems, according to the latest leaks, that AMD is going to release two flavors - 21a and 21b - both of them maxed out at 80CUs.
The first flavor 21a has a huge clock of 2.0GHz at only 200W (!) TDP and 21b has even faster clock at 2.2GHz at only 240W (!!))
According to the latest leaks, the 21b is already faster than 3080 and is closing the small gap of 3090.
This is HUGE news.
A card that is faster than 3080 with a 240W TDP is simply unbelievable, it's like that RDNA2 architecture is the 8th Miracle of the World.
Of course we have to see the real performance, not leaks, but it seems confirmed by many sources.
The 28th of October could be a day that nVidia will remember as a nightmare, if the leaks proved to be valid but for all the others could be a day for celebration.

[SamuriHL]

OMG your "conspiracies" and "inside information" is beyond laughable. None of what you posted is "new information". We already know they didn't get the drivers until a couple days before release. We already knew they didn't have time to fully test before cards were sent to reviewers. They used REVIEWERS as advanced testers, yes, but they did NOT as you seem to be stating use owners as such. And those reviewers that got pre-release boards? Oh noes! What are they to do? Oh, right, wait until the AIB replaces the board for them with production versions...which they are doing case by case. Your attempts to make it seem SO MUCH WORSE than actual reality is literally hilarious. I'm certainly not apologizing for nVidia here, far from it as I still can't buy a card, but, you go the EXTRA mile to make it seem like the biggest conspiracy in history. LOL

All hail the almighty AMD to rescue us from the TRAVESTY of the evil nVidia. LMAO

[oldpainlesskodi]

Me, still waiting for 16k....heard it's going to be mind blowing......

[NikosD]

@SamuriHL Are you OK my friend ?

The cards have problems right here right now, they are in the hands of BUYERS and OWNERS not REVIEWERS.

What conspiracy theories and inside information are you talking about ?

GAMERS not reviewers, mentioned problems with these cards, although if you read the videocardz article you will see that one AIB mentioned some problems before anyone else.

The guinea pigs are the frustrated GAMERS, OWNERS not REVIEWERS.

I made a summary of more than five articles that I read on this subject for users of this forum that didn't have the time to read them.

AIBs released the cards without testing them in games and real-world workloads.

It's just the simple truth and nothing else.

I can't feel pity for your frustration reading my posts, but the truth is that I don't remember such a problematic release of super high-end cards in the last many years.

28th of October is coming....

[SamuriHL]

Me, still waiting for 16k....heard it's going to be mind blowing......

LMAO :D Imagine 8K in each eye for 3D?! WOA LOL

[SamuriHL]

@SamuriHL Are you OK my friend ?


Yup. Perfectly fine. And grounded in reality. Thanks for asking.

[chros]

AMD is going to release two flavors - 21a and 21b - both of them maxed out at 80CUs.
The first flavor 21a has a huge clock of 2.0GHz at only 200W (!) TDP and 21b has even faster clock at 2.2GHz at only 240W (!!))
According to the latest leaks, the 21b is already faster than 3080 and is closing the small gap of 3090.
This is HUGE news.
A card that is faster than 3080 with a 240W TDP is simply unbelievable
Indeed, if it will be true then this will be a huge slap in the face :)

[el Filou]

When any discrete gpu can handle 1080p60 gaming, but you still want to sell expensive gpus, what are you to do ? :)
RT Cores, Nvidia RTX ON, always sounded like marketing BS to me.Probably like 32-bit colour rendering sounded like marketing BS to some people (mostly fans of 3dfx) in 1998. I mean, you're just describing the normal evolution of tech, the first implementations of a new tech are always lower performance than the older standard but after a few generations nobody would want to go back to the old standard. Would you have have wanted to go back to 16-bit rendering in the mid-2000s?

[SamuriHL]

nVidia 456.55

https://www.guru3d.com/files-get/geforce-456-55-whql-driver-download,5.html

[huhn]

looks liek my repoprted bug is fixed: [Notebook]: Performance Power Mode cannot be set from the NVIDIA Control Panel.
[200657525]

the notebook part is odd but what ever.

[NikosD]

He could easily be the fourth or fifth tech journalist writing the same garbage regarding Ampere's performance versus any other card.

I have really lost counting the nVidia pro journalists.

I'm talking about Anthony Garreffa and his hilarious article: https://www.tweaktown.com/news/75367/if-these-new-amd-big-navi-leaks-are-true-the-rtx-3080-will-destroy-it/index.html

I really have no idea if Big Navi21b exists and if it will be faster than 3080 eventually.

All that I'm saying - regarding the same mistake or "mistake" of tech journalists - is that according to the article if Big Navi21b has the specs that have already leaked - we all agree that the performance in TFLOPS metric will be 22.5 TFLOPS just like he says while 3080 we already know is 29.8 TFLOPS like he says too.

Then the result of this comparison of Big Navi21b at 22.5 TFLOPS vs 3080 at 29.8 TFLOPS means that Big Navi21b will be FASTER, clearly FASTER in GAMES than 3080!

Not the opposite!

What an idiot!

[butterw2]

I mean, you're just describing the normal evolution of tech, the first implementations of a new tech are always lower performance than the older standard but after a few generations nobody would want to go back to the old standard. Would you have have wanted to go back to 16-bit rendering in the mid-2000s?

Many shills for Nvidia marketing were adamant that Raytracing was a game changer when RTX released. I'm just saying there has been very little evidence of it so far. Also the cards are expensive, power hungry and low on VRAM.

Even VR might come good at some stage. That doesn't make those who bought into the early hype right.

[SamuriHL]

That is why you always buy a card for what it can do at the time you buy it rather than for some future promise. I never bought the RTX cards for RTX in gaming, although I do enjoy it where I've seen it. I bought them for madvr and don't regret that purchase one bit.

[glc650]

looks liek my repoprted bug is fixed: [Notebook]: Performance Power Mode cannot be set from the NVIDIA Control Panel.
[200657525]

the notebook part is odd but what ever.Have you been able to confirm if that's the case?

[huhn]

not yet

edit: the entire behaviour is totally different on my system. even with optimal is very aggressive but still different from the rest but it's for my system now totally save to use it's very aggressive compared to the old adaptive.

i even trigger a new bug never seen by me before:
https://abload.de/img/bugwdjiz.png
it even was at over 50% clock deviation.

[SamuriHL]

Holy hell that doesn't seem good. LOL I haven't had a chance to play with the new driver yet myself. On the list for tomorrow night hopefully.

[ryrynz]

I'm interested to know if with the 20/30 series and 20H2+ builds and latest driver I can get zero frame drops and repeats over the course of a couple of hours or so.

[VBB]

Latest Nvidia driver is good. Fixes at least one issue I had with the previous one (RGB@60Hz HDR bug). Timings are the same for me, power profiles as well (as far as I can tell).

[NikosD]

Latest nVidia driver tries to fix the "2.0 GHz Frequency Wall" issue of 3080/3090 cards but without success.
They managed only to cap GPU frequency at 2.0GHz, dialing back some performance of GPU boost, in order to avoid the Crash-To-Desktop result of a still unknown root cause.
Some AIBs like Colorful which was the first one to report officially the problem, have already released new BIOS to cap frequency at 2.0GHz.
After the low availability of 3080 cards and the extreme low availability of 3090 cards, this "Frequency Wall" issue was the last thing needed to make this release even worse.
Tough times for 3080/3090 buyers or potential buyers...

[ryrynz]

Buy a manufacturer that cheaps out on caps n that's what you get.

[huhn]

while it is true that there are now 2 revision out for card with different cap setup but that doesn't change that the 3080 FE is crashing like other cards too.

these kind of things sadly happen with most releases.
NAVI was black screening for 9 month this issue was just removed from the known issue list in the new driver.
polaris was breaking the PCIe spec.

we still don't know if the caps and the crashing issue are the same issue EVGA claims the "bad" revision where only send to reviewer.

and bad AIB cards are sadly pretty normal asus TUF 5700 XT does someone really want revision 1 or the strix where asus wasn't able to screw correctly?

https://youtu.be/OJU8jKIYtS4?t=256
don't blame this on AMD.

[nevcairiel]

It appears that at least part of it is that the boost algorithm is too aggressive under certain workloads, which makes the clock jump up far higher then it usually does, and causing the crash in the worst case. Limiting those jumps should not impact performance noticeably - in fact some people are reporting higher speeds with the new drivers since the clock is more stable and thus a bit higher in average, instead of jumping up and then being corrected down for power limiting. But more thorough testing will surely follow.

[el Filou]

I'm afraid we're headed to more frequent issues like this as chip manufacturers rely on dynamic boosting to the absolute limit like GPUs have been doing for a few gens, and AMD started doing with Zen and Intel with TVB in a less extreme way. Margins are thin, feels like the very first P3 1,13 GHz. Thankfully clock controls are now very finely grained and they can fix it like this.

[brazen1]

I can confirm 456.55 modded with 388.59 audio driver is working pretty good using a GTX 960 4GB. I understand this is a well seasoned GPU but I honestly have no reason to upgrade because I passthrough to a Q90R. I know others still use this model as well. I notice great improvements in clock deviations making frame repeats a thing of the past now as has been reported by others. SDR and HDR. I did notice a 2160p 60fps title (Gemini Man) slowly improves but is improved nonetheless.

I do see when the player is windowed or when the seek bar is in view, HDR goes away. The switch is immediate back and forth. The panel does not flash in or out of mode. I tried setting nVidia and madVr to 8bit instead of 12bit. I also toggled FSE on/off in madVr. This did not help fwiw.

Despite this small problem, 456.55 will now become my daily driver because the trade-off of gaining a near perfect clock deviation is worth more to me. Good job nVidia!

On a side note, Gemini Man (2019) has been my go to video for displaying my system. The way video presents is superb imo. I recently obtained Aircraft Carrier Guardian of the Seas (2016). This too is a very high quality presenter imo.

[NikosD]

After reviewing some overclocking data regarding 3080/3090 cards, it is more than obvious that nVidia pushed the overclocking envelope to its limit.
They didn't leave almost any performance on the table, out of the box.
It's a very aggressively clocked product, extremely high-powered and rushed.
But why ?
Because as I have said before, nVidia is not simply afraid of Big Navi.
nVidia is terrified.
So, after the first leaks of Zen 3 performance in games, with the 5800X using 8C/16T being faster than the fastest Intel CPU for desktop the 10900K of 10C/20T, we can now be sure for one thing:

AMD is going to be the champion in both CPUs and GPUs at the same time for gaming workloads, probably for the first time in history.

It will be an epic, monumental milestone for tech.

The next month coming in a few days, will be described in history tech books as Red October

Historic moments.

[chros]

The next month coming in a few days, will be described in history tech books as Red October
:D Hope so!
Btw I'm curious whether they will have 45W 6C 12T CPU (for HTPC without fans).

[el Filou]

Do HTPCs really need 6 cores? Or are you thinking of resale value?
I'd be happy with a much lower TDP 4C/8T, I'm still not limited by my 12-year old actual 25-30W overclocked 2C/2T except for hypothetical 4K copyback which I don't need yet.

[huhn]

maybe there will be a mainboard with a 4800U/4700U that should be your CPU have your eyes on ali express you can find these things there from time to time.

current zen 2 desktops have the relative inefficient io die.

[pirlouy]

AMD is going to be the champion in both CPUs and GPUs at the same time for gaming workloads, probably for the first time in history.

It will be an epic, monumental milestone for tech.

The next month coming in a few days, will be described in history tech books as Red October

Historic moments.
:D

I quote it for the future (I know you don't say it seriously).

[butterw2]

:D

I quote it for the future (I know you don't say it seriously).

Aunt Lisa is coming for you Jensen !


Zen3 5000 series should be released in October.

4600G, 4600GE Renoir zen2 apu might become more widely available:
https://www.amd.com/en/products/apu/amd-ryzen-5-4600g

For low power use, power limits would be an option also, and ClockTuner for Ryzen (CTR) has just been released.

[NikosD]

Another day, another incompetent (shills) tech journalist(s) come up.

They blame in their article the drivers/gaming engines of not using the full FP32 (TFLOPS) potential of Ampere architecture, like VRAY, just as gaming engines have to be CUDA rendering apps.

Instead of "blaming" the Compute-oriented (not gaming) Ampere's architecture and its maximum TFLOPS number advertised by the-leather-jacket man, completely unrelated with gaming workloads which execute 36 INT commands after 100 FP commands on average.

Gaming workloads are not interested in maximum advertised TFLOPS, achieved under specific non gaming conditions.

https://wccftech.com/nvidia-rtx-30-fine-wine-investigating-the-curious-case-of-missing-gaming-performance/

[Klaus1189]

I added it, but I must admit, it is still hard for me to follow the intel update and naming sheme. I also don't understand the supported grafics, like older integrated grafics will require older or other drivers, isn't it.

Please let me know of future releases and I add it, same for issues of it :devil:

[el Filou]

There you go: https://www.intel.com/content/www/us/en/support/articles/000005654/graphics.html ;)

iGPUs before Gen9 are out of performance/quality support now, they just release new drivers for older hardware when there's a security issue.

[SamuriHL]

Geeze Intel. Gen 9?? Really?

[el Filou]

That's what I'm understanding from the driver I'm using for my HD 4600 that covers Gen7 & Gen8, the latest "quality update" was in september 2015, all the ones released since then just mention "security updates" in the release notes. They're only adding features/correcting bugs in the newer 100.x driver that only supports Gen9+
(Edit: just to be sure to avoid any confusion, I mean Gen9 graphics (Skylake), not 9th gen Core CPUs.)

[VBB]

I have been using the 100-series drivers with my 6th gen CPU from the beginning. There is a cut-off with all Intel drivers, but in the case of the 100-series, the cut-off is 6th gen, not 9th gen. What's always been confusing about Intel drivers is overlap between their major revisions. With my 6th gen, I could also use the 15-series driver, for example.

Just saw your edit. That clears things up.

[SamuriHL]

Yea that makes more sense.

[NikosD]

According to the latest speculative video of MLID, the leaks of reverse-engineered MacOS drivers of Big Navi Radeon Pro card were right, although he says that AMD is very cryptic and intensionally is misleading specs, names inside drivers and performance even to AIBs!

The prosumer Big Navi21 will be 80CUs@2.05-2.2GHz with a TDP of around 250W and 32GB GDDR6 using 256bit bus, according to the video.

BUT the Top GAMING Big Navi21 will be even faster, if it sticks to 80CUs and not less.
So, the top gaming Big Navi21 will be a 72CU-80CU at 2.15-2.3GHz with a TDP of 250W to 300W and 16GB of GDDR6 using 256bit bus.
IF AMD actually releases such a monster of 80CUs at 2.3GHz, it is impossible to not be faster than 3090 in games.

Yes, AMD can kill nVidia this time.

His opinion is that Top Gaming Big Navi21 just trade-blows 3080, which doesn't make sense with such a configuration.

[NikosD]

I just saw a message from Tapatalk that @videoh quoted a post of me.

But where is his post ?

[chros]

I saw that video as well, interesting times, similar performance with 100W less?
The price will be the real question.

[Klaus1189]

He/She asked for the link of the video of MLID, but it seems that the post is gone...

[NikosD]

I was thinking about it the last few minutes...

Why He/She ?

[Klaus1189]

I don‘t know if @videoh is a boy or a girl :) The post was a quote of your first sentence of post 1934. He (assuming now) is asking for a link to the video of MLID. Way to less girls around ...

[NikosD]

I think videoh is a very well known member of this forum and has been in the past a member of the moderation group of doom9, if I'm not mistaken.

[Klaus1189]

Probably, I don‘t know. What is MLID? Maybe obvious but I can not think of anything that matches.

I am really looking forward for Zen3 and RDNA2. Very exciting October.

[NikosD]

MLID is just the acronym of the youtube channel Moore's Law Is Dead.

According to the latest leaks, Zen 3 is ~25% faster in single-threaded performance than Zen 2.

Yes, all signs point to a Red October

[Klaus1189]

My cute RX 5700 XT needs a bigger younger brother, since I wanted a 3080, also to get new information by testing it by myself and I didn‘t get one I am kinda pissed but hey AMD is going strong, I just wait for them.

[huhn]

after your death prediction for nvidia by vega i can't wait for how this will turn out.

[ryrynz]

You and me both. There's more hype on this thread than just about anywhere else on the internet.

[NikosD]

No hype...Just leaks of the performance of Zen 3 CPUs and leaks of the RDNA2 specs.

And much of debanking tech journalists and users of this forum laughing of nVidia's fear regarding RDNA2...

[SamuriHL]

I'm not sure what's happening with my system but I THINK it's related to them messing with the power options and I'm *NOT* happy. I had my machine tuned very well and all of a sudden yesterday I'm getting massive dropped frames when upscaling 1080p->2160p. The decoder queue is all over the place whereas usually it's quite stable. I haven't yet tried a clean driver install but this completely blows. My machine was working fine previously.

[NikosD]

Please, take a look of the comments of this "tech journalist" that I previously posted with his hilarious assumptions regarding Big Navi performance versus Ampere 3080 and the TFLOPS numbers in gaming.
I just read the official review of 3080 from Anthony Garreffa an Aussie gamer who thinks he is a tech journalist at TweakTown.com.

Just quotes of the official review of 3080, if you want to see what hype means in official "tech journalism" of well respected sites.
Now, a sensible person would wait until AMD drops Big Navi -- but there are two types of people here, so let's break it down.

Long time NVIDIA fans: Just like a certain website said back with the GeForce RTX 2080 Ti launch: Just Buy It. If you were happy to spend $1199 on the GeForce RTX 2080 Ti at launch (which was fine, because it was a beast at the time and still was, until today) then you'll be ecstatic to spend considerably less than that and get a huge chunk in performance with the RTX 3080.
Performance enthusiasts: You too, guys -- Just Buy It. Seriously, the GeForce RTX 3080 is a graphics card that enthusiasts like me having been waiting for, for a very long time. Almost no restraints total grunt, 4K 60FPS easy and 4K 120FPS in most games.

And another, even more disturbing:
At this point I don't know why you would wait for AMD... I don't see Big Navi being able to topple the GeForce RTX 3080 although I think it will come close. It should launch with 16GB of RAM (GDDR6 or HBM2, rumors of both are floating around) and see NVIDIA retaliate with 20GB versions of the RTX 3080.

You don't need 20GB of VRAM, so if you wanted to bite the bullet now it's probably your absolute best time to get a graphics card before they sell out and the price increases. If you can buy a GeForce RTX 3080 Founders Edition (especially the FE card, it is a monster) for $699, do it.

**a flaming animated gif inserted in his official "review" so to speak**

https://media1.giphy.com/media/qDPg6HNz2NfAk/giphy.gif

[chros]

You don't have to quote idiots/mercenaries, the internet is full of them.

[NikosD]

Probably you didn't read carefully my post. It happens sometimes, especially in this forum.
It's not just an idiot.
He is a tech journalist, officially the main GPU card reviewer of TweakTown.com
Can you see the difference ?

[el Filou]

Could we at least wait until stuff is released and reviewed before going on and on about performance ?
These last few weeks this thread had more off-topic speculations and rants than actual information on driver issues.

[Klaus1189]

@huhn: Were you successfully in finding the issue"s" which you have with 456.55? I am thinking of the clock deviation, but you wrote that there was more (https://forum.doom9.org/showpost.php?p=1924776&postcount=60266)?
@SamuriHL: Were you able to find the root of the issue? DDUed already?

[DMU]

Recently I noticed in the OS HDR mode the metadata became correct.
Pic (https://drive.google.com/file/d/16lAiIuyXnsS338Y7PSftadfIyB_Ffhen/view?usp=sharing)

[SamuriHL]

No not yet. Will work on it tonight.

Sent from my SM-G975U using Tapatalk

[NikosD]

nVidia officially cancelled the launch of 3070 on 15th of October and pushed it back, just one day after the D-Day, the announcement of Big Navi.

The Earth has already started shaking for the upcoming release...

[huhn]

@huhn: Were you successfully in finding the issue"s" which you have with 456.55? I am thinking of the clock deviation, but you wrote that there was more (https://forum.doom9.org/showpost.php?p=1924776&postcount=60266)?
@SamuriHL: Were you able to find the root of the issue? DDUed already?

can't reproduce it and even if i have a problem it doesn't mean it's the driver.

456.55 makes HDR stop working for me i have no clue why it's not the only thing i updated.

[Klaus1189]

@SamuriHL and @huhn: Thank you :)
Please keep us updated here when you too know more.

[chros]

These last few weeks this thread had more off-topic speculations and rants than actual information on driver issues.
I created a new thread for Hardware discussion (https://forum.doom9.org/showthread.php?t=181881).

[chros]

Actually I found a bug, not a small one, with madvr hdr output and 442.92 SD:
- start a HDR10 clip
- don't close the player
- start an SDR clip
- TV stays in HDR10 mode, it doesn't switch back to SDR! :mad:

It works the other way around: SDR -> HDR10
This was fine with 385.28.
Let me know if someone can reproduce it ...

I had/have this issue 3 times now since I reported:
- I thought the bug appears after sleep, but no, it just comes randomly (sometimes it's all good for 4-5 sleep session)
- once it happens, it always happens afterwards until reboot
- only reboot solves the issue

And this just popped into my mind about this (strictly about video playback including audio): they say that AMD drivers are bad in general.
But do we know which was the last nvidia driver that didn't screw up something?! :) I don't remember any ...

[VBB]

New hotfix driver: https://nvidia.custhelp.com/app/answers/detail/a_id/5091/kw/456.71

I'll check it out tonight.

[SamuriHL]

I guess I'll try clean installing that one later and see if it helps.

[Klaus1189]

Recently I noticed in the OS HDR mode the metadata became correct.
Pic (https://drive.google.com/file/d/16lAiIuyXnsS338Y7PSftadfIyB_Ffhen/view?usp=sharing)

Is it the same on Nvidia? I mean working.

[SamuriHL]

We can all stand down on my video issue. That was unpleasant to track down but it turns out JRiver has seemingly broken something in the last 2 builds. Which now that I think about it makes sense cause I hadn't upgraded the nVidia driver yet when the issue started, but, JRiver did update. Going back a couple builds brings stability back to freaking amazing. I'm currently using the hotfix driver without issue.

[SamuriHL]

Next jriver build will be fixed. Love those guys.

Sent from my SM-G975U using Tapatalk

[Annisman]

Can somebody please confirm the latest Nvidia driver that plays well with MadVR AND games for a 2070 super ?

I see 456.71 is listed with no additional notes next to it in the first post, can I assume that the one to use ?

[VBB]

Yes, 456.71 is fine.

[SamuriHL]

Yup, very good driver.

[ryrynz]

What are u guys getting as far as frame skips/repeats with that with 24fps content at 60hz without smooth motion ofc.

[NikosD]

It's almost official....Even the most skeptical guys are talking about a new 290X card for Big Navi.

It seems that AMD went ALL-IN this time.

Yes, AMD has the new King of GPUs:

https://youtu.be/Fe1xycgVVyM

[VBB]

@ryrynz - The same old 41.71 seconds, but why would you want to do 24fps at 60Hz without SM?

[SamuriHL]

I was thinking the same question. What kind of sadistic person plays 23/24fps @ 60 without SM??? LMAO :D I realize it's broken in 113 but isn't the point of having a new awesome driver that we can use proper refresh rates??? :)

[ryrynz]

@ryrynz - The same old 41.71 seconds, but why would you want to do 24fps at 60Hz without SM?

Was curious if that changed any, not every screen can have it's refresh rate changed. How's 24hz? Is it still like about 1hr 30 before drops/repeats?


@ryrynz for example didn't read my post that RTX IO isn't supported by Turing GTX and he replied that it is supported by Turing RTX which is obviously a different card.


Granted I didn't make the distinction between GTX/RTX with regards to Turing when I posted a reply to that but something labelled RTX-anything is going to be obviously only supported only by an RTX card, what the hell do you expect? I honestly didn't expect someone to criticize something so obvious.

[VBB]

At 24Hz it's a drop about every hour and five minutes, unchanged.

[Magik Mark]

Is it possible for you to try CRU? Are there any problems with other refresh rates?

Sent from my SM-G988B using Tapatalk

[NikosD]

...what the hell do you expect? I honestly didn't expect someone to criticize something so obvious. Dude, watch your tone.
You replied today 4th of October to a post of September 9th with that language.
Are you on something ?

[VBB]

Is it possible for you to try CRU? Are there any problems with other refresh rates?

Mark, I don't use CRU, but with these new drivers, 24Hz with one drop every 1.05 hours is actually the worst out of all of them. So, that's much, much better than before, but there really wasn't an issue with anything outside of 24Hz to begin with, if I remember correctly.

[ryrynz]

Are you on something ?

You got asked the same thing. Imagine not realizing product naming is directly related to a feature set.

[nevcairiel]

isnt that the hotfix driver from last week? Guess they were happy with how it worked.

[brazen1]

WoW! 456.71 is performing better than I anticipated.

The audio problem no longer exists! (Stereo vs Multichannel dependent on AVR On/Off state). This was problematic for years. I no longer have to modify the installation with an older HDAudio driver.

2160p 60fps frame drops are pretty much nonexistent. (Gemini Man) The same for other resolutions @ 24, 30, etc. madVr reports some fluctuations (not perfectly steady) but no drops are recorded except for a couple when titles first start. I did actually look for any drops (I'm sensitive to them) to see if madVr just didn't report them. I saw none. Tested opening scenes of The Wrestler and Cowboys and Aliens. Both of these have long panning shots that would drop one frame in the past. Now they drop none and are smooth as butter.

I continue to select 12bit in NCP and 10bit or higher in madVr using passthrough for my 10bit 82Q90R. I get no banding.
When using any player via nVidia API - DVDFab 3, MPC, etc., if the GUI or seek bar is accessed with mouse movement or the player is windowed, HDR switches to SDR as has been reported in many of the newer drivers. This persists. This does not happen when using the Windows API with any of the various players I use. Changing to an 8bit chain makes no difference using the nVidia API fwiw.

This is a minor annoyance for me since I seldom if ever use a mouse or window the player except when multitasking but I do hope the issue gets resolved. 456.71 with zero modifications is now my new daily driver.

[Klaus1189]

@brazen1: Thank you for the feedback, I added it to the second post ;)

@all: 23 Hz still out-of-the-box better of about 1 hour and some minutes? That would be really nice. This looks like Nvidia is finally taking action related to HTPC stuff.

[VBB]

Klaus - Yes, 23Hz has been at about 1h5mins with all the new drivers.

[SamuriHL]

That's all fantastic news. I gotta hand it to nVidia for cleaning this stuff up. The audio issue especially. Impressive!

[ryrynz]

Does CRU or anything else allow better timings or are there still problems with that?

[KrazyLurt]

Does CRU or anything else allow better timings or are there still problems with that?
No not for me anyway not after 451.61. I can still create a custom 23hz with my own tweaked timings that pre 451.61 showed up in windows and as 24hz in ncp under “PC” resolution. But post 451.61 it will revert back to nvidias 23hz if my timings is too close to Nv and as Nv’s 24hz if it’s too close to 24hz.

I’ve also tried using an extension block / DisplayID instead of Detailed Resolution in CRU but it made no difference.

EDIT: Also tried solving the issue with dc2 , nice application overall but didn’t make a difference.

Hopefully someone will come up with a solution.

[Manni]

Unfortunately 456.71 only seems to be available as DCH, not standard. Hopefully that will come soon.

[nevcairiel]

There is no difference in the actual driver between DCH or other releases. But the Standard is also available, NVIDIA just doesn't advertise it big anymore, other sites will list it just fine though, eg. https://www.guru3d.com/files-details/geforce-456-71-whql-driver-download.html

[Manni]

You can certainly download standard drivers.

Well, you can but I can't. I'm using the same link, I use the same options (I know where/how to download standard drivers), and I get the same results for standard drivers except 456.71. No idea why if it's just me.

It does show up if I select DCH though.

EDIT: I changed the language from English US to English UK, and 456.71 does show up now at the top of the list. Weird.

Which language were you using? It doesn't show up on your screenshot.

[Klaus1189]

Here you go:
DE (https://www.nvidia.de/download/driverResults.aspx/16576/de)
EN-US (https://www.nvidia.com/download/driverResults.aspx/165683/en-us)

[Manni]

Here you go:
DE (https://www.nvidia.de/download/driverResults.aspx/16576/de)
EN-US (https://www.nvidia.com/download/driverResults.aspx/165683/en-us)

Thanks, I'm fine with the UK version but these might be useful to others. I'm more interested in finding out if GTPVHD has the same issue if he selects US drivers, which I believe he will.

By the way, has anyone reported any downside for nVidia moving to 2004? I'm still on 1909, but considering the move if there are no known issues with the latest.

[el Filou]

When using any player via nVidia API - DVDFab 3, MPC, etc., if the GUI or seek bar is accessed with mouse movement or the player is windowed, HDR switches to SDR as has been reported in many of the newer drivers. This persists. This does not happen when using the Windows API with any of the various players I use. This is a minor annoyance for me since I seldom if ever use a mouse or window the player except when multitasking but I do hope the issue gets resolved.I wouldn't count on it, at least not as long as there are no changes on the application side. Judging from what is explained on this KB article: https://nvidia.custhelp.com/app/answers/detail/a_id/5072 (they say 'games' but this applies to madVR), the way it's working now is already a workaround applied by the driver "behind Windows' back" to avoid needing users to manually enable HDR in Windows (which, as you described, does not give the problem). The driver is detecting when the apps are displaying SDR or HDR, and dynamically switches depending on that. My understanding is that the driver detects the player GUI as SDR, which it is, so I believe what you're describing is actually by design.
What would be ideal is HDR-aware player apps that display a HDR GUI when playing such content, like the D3D Exclusive GUIs that were developed a long time ago to avoid constantly getting in and out of exclusive mode, but this could be some big work for the players devs.

[el Filou]

It's due to the domain (at least on my connection).
https://www.nvidia.com/Download/Find.aspx?lang=en-us does not list 456.71 when choosing Standard driver and English US
https://www.nvidia.co.uk/Download/Find.aspx?lang=en-us does

(also their site is incredibly slow right now)

[SamuriHL]

By the way, has anyone reported any downside for nVidia moving to 2004? I'm still on 1909, but considering the move if there are no known issues with the latest.

Nope. Been running the fall update for a few weeks now with no problems. Had 2004 since it went release as well.

Sent from my SM-G975U using Tapatalk

[huhn]

you can force the download of the desirable standard driver by simply downloading the DCH driver and changing the URL to the standard driver by removing the "DCH-"/"-DCH" part

[Manni]

Nope. Been running the fall update for a few weeks now with no problems. Had 2004 since it went release as well.

Sent from my SM-G975U using Tapatalk

Thanks, let's hope it's the same with the 3000 series. :)

My plan is to upgrade my 1080ti to the 3090 with 1909, and if everything works well, attempt the upgrade to 2004, with a saved image to go back to if necessary.

I still haven't received my 3090, it's been postponed for a few days. I'm hoping to get it by the end of the week or early next week.

I'll report back when it's installed and I have checked driver's behaviour with madVR.

[Manni]

It's due to the domain (at least on my connection).


Thanks for the explanation, for some reason it shows on both here now. Problem soved!

you can force the download of the desirable standard driver by simply downloading the DCH driver and changing the URL to the standard driver by removing the "DCH-"/"-DCH" part

Thanks for the workaround, I'll try to remember this next time, but I'll try changing the domain first.

[NikosD]

For my wallet, I want to see the performance per price battle of:
3600 (180€) vs 3700X (270€) vs 5600X (300€ ?) on games, single-threaded and multi-threaded performance.
(Obviously the first battle is already known and I don't think Intel is competitive in this price segment with 10600K (250€)

Unfortunately, 5800X is too far away at 450$ but 5900X is a bargain for those who offer 550$ for the fastest 12C/24T on everything (games, 1T, MT) vs 10900K

And one more thing...

[nussman]

Can you stop this off topic nonsense please? This thread is for gpu drivers issues with madVR and/or mediaplayers (read first post). You can open your own thread for amd vs intel vs nvidia flame wars ... :rolleyes:

[NikosD]

The same users of this forum, the same bunch of guys "replying" to me as I remember to "my" threads.
You have discovered a new way to cover your fanboyism under the "off topic" label.
You wasted a completely off topic post just to "reply" to me with language, as the bunch of guys use to do here. A certain bunch of guys.

You offered nothing with your post to this thread, completely useless.
It's the last time replying to completely useless and nonsense posts.

[nussman]

If my post stops your useless posts here it offers a lot to this thread.

[chros]

:D NikosD, why don't you use the thread I created exactly for that purpose???

[NikosD]

My last comment regarding today's spicy presentation.

The "one last thing" I mentioned earlier is obviously for RX 6000 series official benchmarks in three games at 4K Ultra settings by AMD.

AMD decided to simplify things for good this time.

It is definitely more than clear to me that Dr. Lisa Su didn't present the fastest Big Navi card today, but the second fastest.

So, on the contrary with all tech journalists beliefs, if the second and not the first - as they believe - fastest Big Navi card is clearly faster than 3070 and a little slower than 3080 in 4K Ultra, then the fastest Big Navi will be faster than 3080, at least.

Unless AMD and Dr. Lisa Su are so happy that after all these years that managed to be competitive in a place between the second fastest 3080 and the third fastest 3070 nVidia card of Ampere architecture.

But I definitely don't think so this is the case.

Anyway, take your pills in order to sleep tonight after my post and wait for about 20 days to learn for sure.

Interesting start for Red October

[el Filou]

I don't think anybody's losing any sleep over your posts. I'm even following that thread created by chros even though I was anticipating that it would be full of them. We would just like that when we see this thread has new posts, that it isn't stuff about the war between hardware manufacturers and their respective fans instead of useful information about GPU drivers.

[nussman]

We would just like that when we see this thread has new posts, that it isn't stuff about the war between hardware manufacturers and their respective fans instead of useful information about GPU drivers.
:goodpost:

[SamuriHL]

Thanks, let's hope it's the same with the 3000 series. :)

My plan is to upgrade my 1080ti to the 3090 with 1909, and if everything works well, attempt the upgrade to 2004, with a saved image to go back to if necessary.

I still haven't received my 3090, it's been postponed for a few days. I'm hoping to get it by the end of the week or early next week.

I'll report back when it's installed and I have checked driver's behaviour with madVR.

I'm extremely curious what that looks like. Looks like I have no choice but to wait until sometime next year to upgrade. Rumors are they're unleashing the 20gb model in December of the 3080. I'll believe it when I see it. I'm hoping I can grab a card in February for my b-day. But I'd like to see what madvr can do with that much power.

[Manni]

I'm extremely curious what that looks like. Looks like I have no choice but to wait until sometime next year to upgrade. Rumors are they're unleashing the 20gb model in December of the 3080. I'll believe it when I see it. I'm hoping I can grab a card in February for my b-day. But I'd like to see what madvr can do with that much power.

My thinking, given that the 3080 10gb is ruled out for my needs, is that the 20gb street price will end up being closer to 3090 msrp than to 3080 10gb MSRP, as the 3090 will likely sell above msrp for a while.

So given that I was able to secure the EVGA 3090 model I wanted at msrp price, I went for it.

I just hope that at some point they will unlock 10bits on HDMI in RGB for the 3090 for studio drivers. I’m not sure this will happen, but if it does, the 3090 (which isn’t a gaming card) stands a better chance of getting it than a 3080 20gb.

Also, I think I’ll be able to get better cooling/noise/performance with the 3090 than with a 3080 due to the better cooling. I really don’t plan to push that 3090 to the max, just to have it more efficient than the 3080 for a given load with factory clocks. Apart from the HTPC case friendly form factor, that’s why I chose the XC3 with only 2 x 8-pin. Cheaper due to minimal factory OC and little to no OC headroom.

We’ll see!

[NikosD]

There is no room left for 3080/3090 after the official benchmarks of RX 6000 series from AMD yesterday.

We have to wait and see of course third party tech journalists to benchmark the whole line of RDNA2 cards that will be announced officially on 28th of October, but if the availability of RX 6000 series is right, the RDNA2 cards will annihilate Ampere cards.

Of course, for the rest of us, we wait for the prices too.