Media Player .NET (MPDN) - D3D HQ GPU Video Renderer [v2.49.0/v1.31.0 27 Dec 2018]

[huhn]

sorry didn't worked
00000001 0.00000000 [2680] *** Using D3D9 adapter: AMD Radeon R9 200 Series
00000002 0.03274702 [2680] OpenCL is available
00000003 0.22503611 [2680] Platform count: 2
00000004 0.22506173 [2680] *** Platform #1
00000005 0.22509155 [2680] Extensions: cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_event_callback cl_amd_offline_devices
00000006 0.22511385 [2680] Found KHR mem sharing extension
00000007 0.22514789 [2680] clGetDeviceIDsFromDX9MediaAdapterKHR error code: -1
00000008 0.22516958 [2680] Not the GPU that was used to create the DX9 device
00000009 0.22519279 [2680] *** Platform #0
00000010 0.22526719 [2680] Extensions: cl_khr_icd cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_byte_addressable_store cl_khr_spir cl_intel_exec_by_local_thread cl_khr_depth_images cl_khr_3d_image_writes cl_khr_fp64 cl_intel_dx9_media_sharing cl_khr_dx9_media_sharing cl_khr_d3d11_sharing cl_khr_gl_sharing
00000011 0.22528888 [2680] Found KHR mem sharing extension
00000012 0.22532202 [2680] clGetDeviceIDsFromDX9MediaAdapterKHR error code: -1
00000013 0.22534370 [2680] Not the GPU that was used to create the DX9 device
00000014 0.22536540 [2680] The GPU that was used to create the DX9 device does not support OpenCL mem sharing
00000015 0.22539553 [2680] Failed to init OpenCL for Dx9 memory sharing, error 0

[Zachs]

OK thanks again. Looks like I need to get myself an AMD card that supports the extension.

[huhn]

may just wait for someone else. there should be a lot of people out there with a r7 260 or a r9 270x that a is usually the choice of madVR user.

[Garteal]

Got a GTX970 and the application shuts down immediately.
DebugView shows that OpenCL is not available.

Tried DxClTest4 and DxClTest9 only.

[huhn]

Got a GTX970 and the application shuts down immediately.
DebugView shows that OpenCL is not available.

Tried DxClTest4 and DxClTest9 only.

if you have an intel iGPU you may need to run the exe with a 1

[Scyna]

this is with test 9

number 0

00000001 0.00000000 [47120] *** Using D3D9 adapter: AMD Radeon R9 200 Series
00000002 0.01443766 [47120] OpenCL is available
00000003 0.20814273 [47120] Platform count: 2
00000004 0.20835163 [47120] *** Platform #1
00000005 0.20872423 [47120] Extensions: cl_intel_dx9_media_sharing cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_d3d11_sharing cl_khr_depth_images cl_khr_dx9_media_sharing cl_khr_gl_sharing cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_icd cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_spir
00000006 0.20906045 [47120] Found KHR mem sharing extension
00000007 0.20937215 [47120] clGetDeviceIDsFromDX9MediaAdapterKHR error code: -1
00000008 0.20965646 [47120] Not the GPU that was used to create the DX9 device
00000009 0.20988548 [47120] *** Platform #0
00000010 0.21012942 [47120] Extensions: cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_event_callback cl_amd_offline_devices
00000011 0.21036193 [47120] Found KHR mem sharing extension
00000012 0.21120289 [47120] clGetDeviceIDsFromDX9MediaAdapterKHR error code: -1
00000013 0.21187609 [47120] Not the GPU that was used to create the DX9 device
00000014 0.21244131 [47120] The GPU that was used to create the DX9 device does not support OpenCL mem sharing
00000015 0.21311221 [47120] Failed to init OpenCL for Dx9 memory sharing, error 0

number 1

00000001 0.00000000 [47432] *** Using D3D9 adapter: AMD Radeon R9 200 Series
00000002 0.01411890 [47432] OpenCL is available
00000003 0.20938437 [47432] Platform count: 2
00000004 0.20939624 [47432] *** Platform #1
00000005 0.20947592 [47432] Extensions: cl_intel_dx9_media_sharing cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_d3d11_sharing cl_khr_depth_images cl_khr_dx9_media_sharing cl_khr_gl_sharing cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_icd cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_spir
00000006 0.20948598 [47432] Found KHR mem sharing extension
00000007 0.20955338 [47432] clGetDeviceIDsFromDX9MediaAdapterKHR error code: -1
00000008 0.20956211 [47432] Not the GPU that was used to create the DX9 device
00000009 0.20958082 [47432] *** Platform #0
00000010 0.20960639 [47432] Extensions: cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_event_callback cl_amd_offline_devices
00000011 0.20963915 [47432] Found KHR mem sharing extension
00000012 0.20965409 [47432] clGetDeviceIDsFromDX9MediaAdapterKHR error code: -1
00000013 0.20968370 [47432] Not the GPU that was used to create the DX9 device
00000014 0.20969278 [47432] The GPU that was used to create the DX9 device does not support OpenCL mem sharing
00000015 0.20972547 [47432] Failed to init OpenCL for Dx9 memory sharing, error 0

number 2

00000001 0.00000000 [49376] *** Using D3D9 adapter: Intel(R) HD Graphics 4600
00000002 0.00605405 [49376] OpenCL is available
00000003 0.20056851 [49376] Platform count: 2
00000004 0.20058003 [49376] *** Platform #1
00000005 0.20065910 [49376] Extensions: cl_intel_dx9_media_sharing cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_d3d11_sharing cl_khr_depth_images cl_khr_dx9_media_sharing cl_khr_gl_sharing cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_icd cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_spir
00000006 0.20066936 [49376] Found KHR mem sharing extension
00000007 0.20073131 [49376] Found the GPU that was used to create the DX9 device

[Garteal]

if you have an intel iGPU you may need to run the exe with a 1
Technically I do, but I can't use it with this motherboard, so this shouldn't be the problem.
It's trying with the GTX970 anyways.

[2140] *** Using D3D9 adapter: NVIDIA GeForce GTX 970
[2140] OpenCL is *NOT* available

[Zachs]

Hey Garteal, can you try running GPU caps to see what it says on its OpenCL tab?

[ryrynz]

I have had an issue running version 9 on an amd dual graphics system 8500 series + R5 graphics. I disabled each one and also set the preferred card but it didn't work. I'll copy the debugview when I post updated results.

Geforce 820M 90 FPS
Geforce 840M 105 FPS
Geforce GTX 850 187 FPS

HD Graphics (Broadwell) 115 FPS
HD Graphics 4000 360 FPS
HD Graphics 4400 (64 bit) 170 FPS
HD Graphics 5500 346 - 358 FPS

AMD M260 215 FPS
AMD M265X 390 FPS

AMD R5 / HD8570

[11560] *** Using D3D9 adapter: AMD Radeon(TM) R5 Graphics
[11560] OpenCL is available
[11560] Platform count: 1
[11560] *** Platform #0
[11560] Extensions: cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_event_callback cl_amd_offline_devices cl_amd_hsa
[11560] Found KHR mem sharing extension
[11560] clGetDeviceIDsFromDX9MediaAdapterKHR error code: -1
[11560] Not the GPU that was used to create the DX9 device
[11560] The GPU that was used to create the DX9 device does not support OpenCL mem sharing
[11560] Failed to init OpenCL for Dx9 memory sharing, error 0

[Anima123]

Test 9, I believe the 366.7 fps belong to the performance of nVidia 880M this time.

[Zachs]

This is odd! With AMD's mobile products their driver worked fine. This is with the same code.

@ryrynz could you run debugview on the mobile AMD systems that worked? I'd like to see what the log says!

[cyberbeing]

DxClTest9 | GTX 770 2GB | Win7 SP1 x64 | 350.05 Driver | PCI-E 3.0 x8

Stock (1.2Ghz Boost, 224.4GB/s Memory): 1206fps

Overclocked (1.3Ghz Boost, 243.5GB/s Memory): 1301fps

[Zachs]

Very nice! Nvidia's extension is working very well indeed. AMD's KHR extension on the other hand is worse than Intel's! Doesn't even work for all their discrete graphics cards! What gives!?

[huhn]

Very nice! Nvidia's extension is working very well indeed. AMD's KHR extension on the other hand is worse than Intel's! Doesn't even work for all their discrete graphics cards! What gives!?

AMD has a known issue with Dx9 <-> openCL interop is it terrible slow. it is clearly a copyback interop limited by the PCIe speed.

you will never see good results with amd using an dx9 <-> openCL interop.

the openCL performance is still way better on AMD than nvidia only the copyback work like a handbrake.

[Zachs]

OK you guys with AMD cards, can you try this out (DxClTest10.zip (http://zachsaw.com/downloads/mpdn_test/DxClTest10.zip))? I think I've found the problem - silly thing that for some reason didn't bother Intel but does AMD.

[huhn]

better i guess:

00000001 0.00000000 [3204] base\appmodel\execmodel\modern\lifetimemanager\appxexecutionutil.cpp(55)\modernexecserver.dll!00007FFDBBFFE8CA: (caller: 00007FFDBC0103BA) ReturnHr[PreRelease](16) tid(b04) 80073D57 The process has no application identity.
00000002 7.16983414 [2904] *** Using D3D9 adapter: AMD Radeon R9 200 Series
00000003 7.20206404 [2904] OpenCL is available
00000004 7.37989759 [2904] Platform count: 2
00000005 7.37991571 [2904] *** Platform #0
00000006 7.38002014 [2904] Extensions: cl_khr_icd cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_byte_addressable_store cl_khr_spir cl_intel_exec_by_local_thread cl_khr_depth_images cl_khr_3d_image_writes cl_khr_fp64 cl_intel_dx9_media_sharing cl_khr_dx9_media_sharing cl_khr_d3d11_sharing cl_khr_gl_sharing
00000007 7.38004827 [2904] Found KHR mem sharing extension
00000008 7.38008165 [2904] clGetDeviceIDsFromDX9MediaAdapterKHR error code: -1
00000009 7.38010693 [2904] Not the GPU that was used to create the DX9 device
00000010 7.38012075 [2904] *** Platform #1
00000011 7.38015461 [2904] Extensions: cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_event_callback cl_amd_offline_devices
00000012 7.38016748 [2904] Found KHR mem sharing extension
00000013 7.38324404 [2904] Found the GPU that was used to create the DX9 device
00000014 7.44802427 [2904] Error Getting CL Surface
00000015 7.44828558 [2904] Error Acquiring Dx9 Surface
00000016 7.44832802 [2904] Error Releasing Dx9 Surface
00000017 7.46745014 [2904] Error Acquiring Dx9 Surface
00000018 7.46755838 [2904] Error Releasing Dx9 Surface
00000019 7.46781921 [2904] Error Acquiring Dx9 Surface
00000020 7.46787310 [2904] Error Releasing Dx9 Surface

there was a lot more from these

00000015 7.44828558 [2904] Error Acquiring Dx9 Surface
00000016 7.44832802 [2904] Error Releasing Dx9 Surface

things

[Zachs]

OK really need an AMD card now!

[Garteal]

Hey Garteal, can you try running GPU caps to see what it says on its OpenCL tab?
Initially it said OpenCL 1.1, but after doing some research I found that NVIDIA had released a hotfix (http://nvidia.custhelp.com/app/answers/detail/a_id/3647) that adds OpenCL 1.2 support amongst other things.

With that I'm getting around 947FPS on a GTX970 in a PCIe2.0x16 slot.

[Zachs]

Nvidia can do true memory sharing without going to system memory and back, otherwise you'd be capped to the system memory bandwidth and there's no way it'll go anywhere near that speed! :)

[Zachs]

better i guess:

there was a lot more from these

00000015 7.44828558 [2904] Error Acquiring Dx9 Surface
00000016 7.44832802 [2904] Error Releasing Dx9 Surface

things

Can all of you with AMD discrete graphics card try this test app (v11) (http://zachsaw.com/downloads/mpdn_test/DxClTest11.zip)? I've written a specific code-path just for AMD's discrete GPUs.

[burfadel]

With test v11, I am getting around 168 fps. It has a quickly flickering big white square in the middle of the window.

With test v10, I am getting 8638.2 fps, give or take. However, the window is black.

AMD R9-280X.

[huhn]

build 11 works.

169 fps good old interop bug i loving it.

[mrcorbo]

With build 11 I'm getting around 37(!) fps on a 270X.

[huhn]

With build 11 I'm getting around 37(!) fps on a 270X.

what pcie version are you using?

[Anime Viewer]

OK guys, here's another OpenCL test app (http://www.zachsaw.com/downloads/mpdn_test/DxClTest4.zip).

Debug log now prints the extensions supported as well as the GPU name of your chosen adapterID.
More importantly, I've added Nvidia GPU support. Could you guys with Nvidia GPU (as well as the Optimus guys) give it a go and report back?
As before, I'm interested in min/max frame rate and your GPU model.

Oh and if you don't see flickering / flashing square then it isn't actually working.

Cheers!

I know other Optimus users have already shared their results, but for the record my system (with the gpu in my signature) is reporting ~338 fps (when run in its default size box ~228 if I maximize the box to the 1920x1080 screen size).

Initially it said OpenCL 1.1, but after doing some research I found that NVIDIA had released a hotfix (http://nvidia.custhelp.com/app/answers/detail/a_id/3647) that adds OpenCL 1.2 support amongst other things.

With that I'm getting around 947FPS on a GTX970 in a PCIe2.0x16 slot.

My original test (above) was with OpenCL 1.1 also, after installing 1.2 which was bundled with driver 350.05 Garteal linked I was getting 345 in the default window and 269 with the window maximized. OpenCL and the new driver were an improvement for me too based on that test.

Edit: 350.12 is now out...

[Scyna]

on build 11 i get 170 fps and build 10 like 8.8k on amd 290

[mrcorbo]

what pcie version are you using?

2.0/x16

[huhn]

2.0/x16

maybe the reason but pcie 2.0 should be able to do better.

[Zachs]

Build 10 wasn't actually working so it wasn't doing anything at all which explains why it's so fast :)

Oh BTW don't maximize the window as you'd then be causing windows desktop composition to do a bigger memory copy which would affect the result obviously.

[Zachs]

There's actually no requirement for any vendors to do memory sharing on the GPU itself, so technically it isn't actually a bug on AMD's part. But yeah it's painfully slow.

[huhn]

There's actually no requirement for any vendors to do memory sharing on the GPU itself, so technically it isn't actually a bug on AMD's part. But yeah it's painfully slow.

it's not doing a dx9 <-> openCL interop if you think about it.

[Zachs]

You should blame Khronous for defining it as a proper way of sharing surfaces then. It says it's up to the driver to avoid memory copy but stopped short of saying it should when possible. My feeling is it's a hardware design flaw in AMD's discrete cards that it's the actual problem and it can't be fixed unless they redesign the hardware. Just a guess.

[huhn]

dx9 <-> direct compute was fine as far as i remember.

[Zachs]

Yeah but a different set of API. OpenCL works on cards that don't support DirectCompute.

[huhn]

this is just my guess why this maybe isn't hardware related.

http://abload.de/img/directcomputegtugt.png

at least i remembered correctly.

edit: yeah i totally forgot. some people can't reach 24 FPS so they can't use dx9 <-> openCL for any realtime program. i call this a bug and i'm talking about GPU like r7 or r9 series.

[Zachs]

When I said hardware related, I meant its inability to share memory on the GPU without doing a copyback.

[ryrynz]

Can confirm build 11 working on the R5 system.

[Zachs]

What frame rate did you get with it? Around 170 too?

[ryrynz]

LOL. No. R5 is weak. 15 FPS. I can't get the 8570 detected debugview only says there's one platform.

[huhn]

When I said hardware related, I meant its inability to share memory on the GPU without doing a copyback.

but directcompute share memory on the GPU. yes it's a different APi no question but this still gives hopes doesn't it?

but speculating doesn't fix this issue. the issue is that some people can't reach 100 FPS and all AMD card even r5 should be able to do that easily. even which this copyback madness. even intel iGPU can do this and these GPU are running on the CPU ram in the first place...

[aufkrawall]

this is just my guess why this maybe isn't hardware related.

http://abload.de/img/directcomputegtugt.png

at least i remembered correctly.

edit: yeah i totally forgot. some people can't reach 24 FPS so they can't use dx9 <-> openCL for any realtime program. i call this a bug and i'm talking about GPU like r7 or r9 series.
Could you please relink that benchmark? :)

[huhn]

Could you please relink that benchmark? :)

doesn't work with nvidia.

if i remember correctly someone has written a version for nvidia(was it cyberbeing? sorry i can't remember) but yeah this is the original version from madshi from his server for AMD GPUs (works fine for intel).

i use it with every new AMD driver version just to see nothing has changed...

[aufkrawall]

Thanks. Since I'm with NV, I hope that the author of the compatible version will read this or maybe someone else still has it.
:thanks:

[Zachs]

It's a synthetic benchmark, so don't read too much into the result.

[huhn]

and it's a copy benchmark the really important part is the processing that's done on the frame later at openCL.

but numbers like 37 fps slow some real problem...

[Zachs]

I meant the Nvidia benchmark.

[aufkrawall]

With your benchmark 11, I'm getting 1278fps (GTX 980). Seems to be CPU limited. Cuda wrapper overhead?

[Zachs]

I think it's gpu memory bandwidth limited.

[huhn]

just run it with gpu-z.

[aufkrawall]

I get lower fps when I have Process Hacker opened, which consumes CPU power. But probably it could also be due to its way of GPU polling?
It's definitely bandwidth limited. When I OC VRAM from 3.6 to 4.0Ghz, I get an increase to 1380fps.