Recommend a graphics card for video enhancement (Upscaling)

[MalickT]

I currently have Radeon RX 560. It runs 1280MHz on overclocking mode. It has H265 and H264 support.

I use Topaz Video AI Enhance to upscale my 1080p videos to 4K. Im impressed with the result but it takes 7 days for 2h video. About 3 second a frame!

I do not play any games. I only want faster processing so it is the GPU clock frequency that matters, right?

I remember that my card cost only 170$. I have looked for cards that cost around 500-800$ but still the specs say about 1400Mhz GPU. Only 200Mhz faster than my current card. Im confused.

Im willing to pay 500-800$ for a card that can do the job for 1-2 days instead of 7 days.

Any recommendations?

[QBhd]

clock speeds mean nothing in today's processor world. IPC (instructions per clock) is the name of the game today. And there is no real value associated to that. Also there almost is no $500 cards anymore. I could sell my current Sapphire Nitro+ RX 5700XT and probably make a profit on it even now!

QB

[Asmodian]

I remember that my card cost only 170$. I have looked for cards that cost around 500-800$ but still the specs say about 1400Mhz GPU. Only 200Mhz faster than my current card. Im confused.

GPUs are VERY parallel. Most of the difference between models is the number of "cores" not the speed of each one.

The RX 560 has 896 or 1024 shaders (depending on model).
The 6800 XT has 4608 shaders, so it would offer a huge increase in speed without any changes in clock speed.

You cannot blindly compare shaders across architectures, they can be very different, but within the same architecture they are what makes one GPU faster than another.

[RanmaCanada]

Cryptocurrency mining and Covid have made current and past video cards extremely difficult to buy at retail, let alone to buy period. Even if we recommend you a card, would you be willing to pay over twice the retail price of it?

[MalickT]

So pretty much the more expensive, the better/faster?

[benwaggoner]

So pretty much the more expensive, the better/faster?
For current gen cards. There are a lot of prior gen cards around at high prices that may be worse for your application.

It really depends on how the software is leveraging the GPU. Could be VRAM bound. Could have a maximum number of cores it will use. Could have big perf improvements between generations, or not.

Does the documentation or support info make any recommendations?

[ognirats]

VSgan > topaz

[Asmodian]

So pretty much the more expensive, the better/faster?

Pay attention to the number of shaders and clock speed. Also, the memory bandwidth can be very important, but I don't know if that bottlenecks Topaz.

Comparing GPUs based on paper specs is somewhat complicated. You really need benchmarks of your specific application, but shaders and memory bandwidth are the most important things to look at if want to judge if it is worth looking for benchmarks.

[Blue_MiSfit]

This really belongs in the PC Hardware forum, not the HEVC encoding forum :) Moving.

[MalickT]

Yes, hardware topic but involves H265 encoding.

What about this card: https://www.zotac.com/us/product/graphics_card/zotac-gaming-geforce-gtx-1650-oc-gddr6

If I could get average 2 days per video instead of 7 days per video, it would be great...

[Asmodian]

The 1650 is not a significant upgrade over the RX 560. It only has 896 CUDA cores, so it doesn't have any extra cores (not that you can compare directly, but it would need more to reach 100+% faster). I don't know if Topaz is faster on CUDA, it was CUDA originally so the 1650 might offer a more significant boost than expected, but in general a 1650 is only a bit faster than a RX 560.

[MalickT]

How about Palit GeForce RTX 3060 StormX - https://www.palit.com/palit/vgapro.php?id=4107&lang=en&pn=NE63060019K9-190AF&tab=ov

?

[Asmodian]

Yes, that would be a lot faster. 3584 CUDA cores and 360 GB/s of memory bandwidth.

[MalickT]

OR what if I buy a 800$ high end CPU with 8 or more cores? Or CPU will never be a match for GPU with H265 support?

[excellentswordfight]

OR what if I buy a 800$ high end CPU with 8 or more cores? Or CPU will never be a match for GPU with H265 support?
Well, we need to know if the bottleneck is the image processing or the video encoding, and if the encoding is using hw-encoding at all. Cause the encoding cannot be faster then the speed of the image processing (upscaling), so if thats done at 1fps, then the upper limit of the encoding would be 1fps.

I'm not to familiar with Video Enhance AI, I did a bit of googling but Topaz dont specify much at all unfortunately, so making GPU recommendations seems a bit hard; does it scales with tflops? Does it prefer CUDA? Does it utilize tensor cores? Memory bandwith intensive etc. I would direct my question to their forum instead.

If it's the video encoding that is the bottleneck, you would need to know if it does it in sofware (then speed would be CPU-dependant) or in hardware. If its in hardware the price of your GPU wont do much in terms of speed, it would more be a question of AMD vs Nvidia or the generation of the card.

[MalickT]

The bottleneck is both image processing (enhancement) and encoding.

For testing, I used a 20 second 1080p clip of 30MB size. I used no compression (lossless) when encoding. Enhanced the video into 4K

using GPU (AMD Radeon RX 560):
Time - 18 minutes
File size - 2.5GB

using CPU (AMD Ryzen 3 3200G):
Time - 4 hours!
File size - 350MB

There was no difference in video quality.

I have heard that using CPU for encoding will always get you better quality and smaller size but GPU encoding is faster but not so good quality and larger size.

So I am still in between two options: buy CPU (AMD Ryzen 9 5950X) or buy GPU (Palit GeForce RTX 3060 StormX)

[Asmodian]

That does not sound correct to me. Upscaling is very likely the slow step, encoding is relatively quick. Topaz quotes 0.4 sec for SD to HD on a GTX 1080, HD to 4K will a lot slower on a RX560.

There is a very big difference between lossless compression and no compression. What framerate was your test clip? I think you had something else different between those encodes. If you used no compression you cannot have gotten a 350MB file.

Topaz uses the GPU for upscaling, but I have no idea how it encodes the video. If it uses software encoding a faster CPU might speed it up, but it would never be faster than the upscaling, so I don't think a faster CPU is a good idea unless you already had a faster GPU.

The 3200G does have a GPU as well. Are you sure you weren't using the 3200G as the GPU for upscaling during this test? It would be very slow, even compared to a 560.


I have heard that using CPU for encoding will always get you better quality and smaller size but GPU encoding is faster but not so good quality and larger size.

This is the difference between using software video encoding on the CPU or hardware video encoding on the GPU. Hardware encoding is very fast, but generally larger for the same quality. However, this difference in size would be much less that 2.5GB v.s. 350MB you saw.

I wouldn't worry about video encoding until you can do the processing faster. A 3200G would be terrible for CPU video encoding, but the hardware encoder on the 3060 is pretty good, so that would be a great option.

[MalickT]

I got my new CPU today (AMD Ryzen 9 5950X). A little dissapointed, speed is 5 sec/frame. (Previous CPU was 20 sec/frame) So my current GPU is still faster than my new CPU (GPU does it 2 sec/frame)

For some reason the CPU shows about 35% busy when upscaling the video. I was expecting the CPU would be 100% busy. I have to take a look into it. Why isn´t Topaz not taking full advantage of my CPU...

[mastrboy]

I got my new CPU today (AMD Ryzen 9 5950X). A little dissapointed, speed is 5 sec/frame. (Previous CPU was 20 sec/frame) So my current GPU is still faster than my new CPU (GPU does it 2 sec/frame)

For some reason the CPU shows about 35% busy when upscaling the video. I was expecting the CPU would be 100% busy. I have to take a look into it. Why isn´t Topaz not taking full advantage of my CPU...

Unfortunately not all workloads scale linearly with cores, since you are capping out at 35% you could try splitting the video into 3 segments and running 3 instances of the processing software, that should be able to utilize 100% and shorten total rendering time.

[Asmodian]

Even bringing it to 100% usage, with perfect scaling, would only bring it down to 1.75 sec.

You really need a faster GPU.

[RedDwarf1]

I have tried Topaz Video Enhance AI and the more shaders/CUDA cores the better. I have a none Super GTX1660 with 1408 shaders and Hynix DDR5 memory and the card is made by MSI. The memory is utter crap so I would avoid MSI cards if I was buying again. It underperforms in Mining below the speeds for this GPU. The memory quality and possibly the memory timings could be to blame. The memory does overclock to 4950 MHz which can help a little but overall I would not purchase again. I did miss out on a Super by around 6 months because they were just released when I purchased mine but the prices were too high at the time. The GTX1660 Super cards do have quite a lot more memory bandwidth but you might be better off trying to get a higher model nvidia card with more shaders. The video card does help quite a lot with TVEAI. However I was still only getting a little over 1 fps encoding rate but I have forgotten the precise speed because it has been quite some time since I tried it but it is a lot faster than the CPU alone.

I believe even a GTX1660 Super would beat your CPU if encoding with TVEAI. It might be worth encoding to a large lossless file and then software encoding using x265 using the lossless file as input. Then you would do the video processing and enlargement with the video card and then software encode using your CPU. I was not impressed with the actual video encoding done by TVEAI TBH which is why I recommend that you encode to a lossless file and then software encode with your CPU after doing the enlargement. I would stick to nVidia cards and get as large amount of shaders that you can afford. Maybe see if you can get a used RTX2000 series card. Another thing is get one which can switch off the GPU fan(s) because some cannot. Mine can MSI GTX1660 Armour which does have an inbuilt overclock which maxes out at around 2010 MHz which is slightly above the standard boost frequencies of these cards. Getting MSI Afterburner which allows alteration of the GPU speed, memory and fans on most makes of cards but it will not turn off the fans unless the card is capable of doing so. Try and find a card which uses DDR6 or a card which does not use Hynix memory <- MSI.

However even Hybrid can do some fairly decent and fast video encoding using the hardware encoder on the GPU. It gives me some fast encoding speeds to H.265 and allows setting the number of reference frames and b frames which most other hardware encoders hide from the user and do not allow alteration. The quality does not match software encoding but the speed really makes up for it if a higher bitrate and file size is an okay trade off.

[EDIT: Additional]
A better graphics card would also enable better settings in MadVR if that takes your fancy. My GTX 1660 does struggle a bit, dropping frames on higher settings when debanding and deblocking is used.

[MalickT]

I split the video into 4 pieces and started 4 instances of Topaz Video Enhance AI assigning each one a piece of the video to process (enhance to 4K). So I had 4 Topaz instances running at the same time, using CPU not GPU.
The CPU still did not get above 40% and every instance of Topaz ran 4 times slower!

I even tried like so - In task manager, I set the 1st Topaz CPU affinity from 1-4. So it would use only 1-4 of CPU cores. Then second Topaz affinity I set to 5-8 CPU, third Topaz 9-12 CPU and 4th instance of Topaz 13-16 CPU. I was hoping that then all my 16 CPU Cores are in use and I get near 100% CPU usage. But that did not help.
Every Topaz instance run 4 times slower than running only one instance and CPU remained around 40%.

I also tried HandBrake encoding software. I just run a test to convert H264 4K video to H265 4K using CPU not GPU. The CPU still did not get passed the 40% mark and the encoding process was only about 2 times faster than my previous CPU.

So where is the bottleneck? Why my PC is not taking full advantage of my CPU? Is it the motherboard?
my motherboard is Gigabyte GA-A320M-H V1.1, not very fancy but supports the latest AMD processors
I updated my motherboard BIOS to the latest version as well. In BIOS I have enabled CPU Virtualization. So I did not see anything else what I could do in BIOS.
I use water cooling block to cool my CPU and the temp is around 65C/2197F

As for memory, I have 16GB DDR4 1600MHz and Task Manager shows its about 40% in use, just like my CPU when Topaz is running.

CPU specs again:
OLD CPU - AMD Ryzen 3 3200G (4 Cores)
NEW CPU AMD AMD Ryzen 9 5950X (16 Cores)

Speed gain only 2x when using CPU for encoding/enhancement :(

[ReinerSchweinlin]

CPU Models in VEAI are different than the GPU ones - and also a CPU will have a very hard time geeping up with even a small GPU... For something like Topaz VEAI it makes no sense to get a high-end CPU and stay on a low-level GPU.

Since VEAI can take advantae of high FP16 speeds - if GPUs offer that - look at cards like VEGA 56 / 64, VEGA7, GEN20 RTX Cards for better speeds... While some GEN30 cards are faster, they cost much more at the moment. A 3090 simply makes no sense, you can get 3 VEGA64 for the same price at the moment and will outperform the 3090 easily.

Stay away from gaming benchmarks, the don´t translate into VEAI Speeds..

If you are lucky, a 2080TI can be bought for around 500 bucks, same for a VEGA64. Both are much faster than what you have now and a VEGA64 even beats some smaller GEN30 cards in VEAI (NOT in games or other benchmarks).

RDNA1 Cards could also be an option.
Stay away from GCN 4 and older or GEN10 and older cards - while they do well in gaming, they really are slower in VEAI..

Of course, it all depends on the price you pay for a card - gaming and mining dictate the prices nowadays...

Or go vor multiple cards - if you are running higher than DVD resolutions, the speedups are good when using two cards.

If you have the possibility to spend a little more, go ask your local distributor for NVIDIA Pro Cards about pro Card Prices.... Cheaper than the gaming counterparts nowadays :)

VRAM is not that important - 8 GB with an FP16 capable Card will do fine in most cases (unless you want to go 4K to 8K or stuff like that - but then you will go for a very big card anyways...)


Encoding: The output section of VEAI is garbage - do single frame exports and encode afterwards with whatever you prefer. In case of a crash you can recover better, too. Don´t bother for audio export, its pretty much broken in VEAI. Mux manually.

[ReinerSchweinlin]

I currently have Radeon RX 560. It runs 1280MHz on overclocking mode. It has H265 and H264 support.
This card is slow for something like VEAI.
h265 and h264 support are worthless in veia - decoding is done via ffmpeg in software - and the encoder quality of your card is not really worth it (ok for streaming, but not for video encoding)....

Get a bigger card :)

[MalickT]

I guess I made a mistake buying the new CPU. Now I am having hard time keeping the CPU and GPU temperatures down. I have the ZALMAN RESERATOR 1 Fanless Water Cooling System - https://www.newegg.com/zalman-liquid-cooling-system/p/N82E16835118111
The water tubes going first to CPU and then into GPU so they share the same temp.
I currently have the GPU running VEAI and the temps show 85°C. Google said that the my CPU can handle 90°C so its on the edge...

I guess I will buy better GPU in coming months and just keep the CPU just for show off...

[ReinerSchweinlin]

No need to cool a RX560, the normal air cooler will do. Overclocking such a small GPU is not worth it.
Remove the GPU from the circuit, cool only the CPU with it. Run something like Super-PI or any other extreme stress test, monitor the temps with hw info and if the temps still get above 80°C, there is something wrong in your cooling system. Maybe repaste with a better thermal compound ? My Rxzens with air cooling never go above 60°C (ok, big chunks of metal...)

[MalickT]

My water reservoir is nicely warm about 35°C but CPU temp is currently 72°C. Perhaps the water circulation is not fast enough or CPU waterblock is clogged

[ReinerSchweinlin]

My water reservoir is nicely warm about 35°C but CPU temp is currently 72°C. Perhaps the water circulation is not fast enough or CPU waterblock is clogged
I´d suspekt the thermal storage capacity of your system to be big enough four your CPU. So it seems its a heat transfer problem - die heat doesn´t get transportet away from the CPU fast enough... check for air bubbles, check for thermal paste and if all that is fine - increase flow speeds.

[MalickT]

I specially bought thermal paste before I installed the CPU. I applied a very small layer of paste on the waterblock before mounting it on the CPU. Is there are rule that the more thermal paste the better? I will try to apply more thermal paste.

However, I now have read from the internet that other people with AMD Ryzen 9 5950X are also having heat problems and some say it is a software issue on some of the motherboards. I will try to upgrade the chipsets and look what I can optimize in BIOS.
My MB is Gigabyte GA-A320M-H V1.1

[videoh]

I will try to upgrade the chipsets
What do you mean by that?

[MalickT]

What do you mean by that?

I updated my motherboard chipset drivers and applied more thermal gel between CPU and waterblock. In BIOS I disabled CPU core boost.

CPU Temp now at 54°C while GPU is running Topaz Video Enhance. Much better, but I still need to buy better GPU to increase topaz video enhancement speed.

[videoh]

Gotcha. I never heard of anyone changing the actual chipset on a mobo. :D

[Asmodian]

In BIOS I disabled CPU core boost.

This disables clock speeds above the base clock, which is a huge loss of performance (20%). It does make it generate a lot less heat, but it is a very significant underclock of the CPU. :(

85°C is "fine". I would not want to run much hotter than that but it is not worth losing that much clock speed to keep the temperature below 85°C.

I think people talking about the motherboard causing it to run too hot at stock don't know what they are talking about. There is a lot of misunderstanding of Zen3 on the Internet. :p

I have a 5950X too, and enjoy discussing the details of their voltage and frequency behavior with others. At stock or with PBO overclocking they run AMD's configured voltage/frequency curves, not the motherboards. Don't start looking for BIOS settings or drivers to change it. The reason it runs hot is very unlikely to be the fault of the motherboard at default settings.

The only motherboard option to change the voltage at a given load is LLC, which should be set to "Auto" or the lowest option available (curve optimizer offsets also influence the voltage, but that is getting into PBO overclocking which I don't think you are doing). However, if you haven't changed it, LLC should be set to the correct (non-overclocked) value by default.

My water reservoir is nicely warm about 35°C but CPU temp is currently 72°C. Perhaps the water circulation is not fast enough or CPU waterblock is clogged

A delta of 37°C between water and CPU is great, no sign of any issues there. I see an almost 50°C deltaT between CPU and water (water at 26.1, CPU at 75.3°C) with a custom water loop. My flowrate isn't great with only one pump in the loop, ~150 L/h, but yours looks fine.

[ReinerSchweinlin]

I updated my motherboard chipset drivers and applied more thermal gel between CPU and waterblock. In BIOS I disabled CPU core boost.

CPU Temp now at 54°C while GPU is running Topaz Video Enhance.
to get a real feel of your cooling capacity, don`t disable boost and rund some hevay stress test on the CPU to see if you are safe. Running VEAI may be your workload but doesn´t represent "worst case" for your CPU.

[MalickT]

Im still having some stability issues, random PC crashes. Last night I started Topaz Video Enhance process, it run all night long. This morning all was good, temp was 55°C and Topaz was still running with 12 hours remaining. Then I went to work and when I came back the PC screen was frozen and did not respond. On the frozen screen I saw CPU temp was still 55°C but clock was 4 hours behind meaning the froze took place 4 hours ago.

So far this new CPU has caused me more trouble than benefits! Damn!