Can someone explain what affects performance and how?

Topaz Video AI General
1

Back in the day when I was working on Unix machines there are various things that affect performance and can bottleneck a system. How does one configure or even purchase a new system for performance. I have a Ryzan 7 Windows machine with 32GB and an NVIDIA 1660 card. Not the biggest card but a good one. Why us the performance 6fps. What can I do to boost performance?

Topaz Video AI  v4.0.4
System Information
OS: Windows v10.22
CPU: AMD Ryzen 7 2700X Eight-Core Processor           31.924 GB
GPU: NVIDIA GeForce GTX 1660  5.8408 GB
Processing Settings
device: 0 vram: 1 instances: 1
Input Resolution: 720x480
Benchmark Results
Artemis		1X: 	06.87 fps 	2X: 	05.32 fps 	4X: 	02.17 fps 	
Iris		1X: 	09.31 fps 	2X: 	05.52 fps 	4X: 	01.96 fps 	
Proteus		1X: 	06.65 fps 	2X: 	05.31 fps 	4X: 	02.10 fps 	
Gaia		1X: 	02.27 fps 	2X: 	01.54 fps 	4X: 	01.00 fps 	
Nyx		1X: 	02.52 fps 	2X: 	02.43 fps 	
4X Slowmo		Apollo: 	10.20 fps 	APFast: 	43.16 fps 	Chronos: 	05.24 fps 	CHFast: 	07.95 fps
2
3

I understand the benchmark process. The issue is activity monitor shows almost no use of resources during benchmark 14% CPU and 2% GPU with a spike to 100% for under 1 second.

4

https://community.topazlabs.com/t/video-ai-4-0-x-user-benchmarking-results/54396

:point_up_2: Click the Link to view benchmark from others :point_up_2:

5

I am asking what affects the performance. Seeing other peoples performance numbers is not a help. Is performance strictly in the GPU?

6

Both GPU and CPU are important.
It depends on which AI model you use, upscale ratio (1x, 2x, 4x) and input resolution.

You can compare benchmark scores from users with the same GPU but different CPUs. This way, you can determine the effect of the CPU.
Similarly, you can compare benchmark scores from users with the same CPU but different GPUs. This way, you can determine the effect of the GPU.

1 Like
7

Because your GPU is relatively week, your CPU is not the bottleneck…
I have a weaker CPU then you (10 years old) but a stronger GPU (not that strong on the market, for sure) and you can see that my results are higher then what you posted. so it is pretty conclusive that your bottleneck is the GPU.

Topaz Video AI  v4.0.5
System Information
OS: Windows v11.22
CPU:        Intel(R) Core(TM) i7-4930K CPU @ 3.40GHz  31.941 GB
GPU: NVIDIA GeForce GTX 1080 Ti  10.859 GB
Processing Settings
device: 0 vram: 1 instances: 1
Input Resolution: 768x576
Benchmark Results
Artemis		1X: 	18.08 fps 	2X: 	13.77 fps 	4X: 	05.24 fps 	
Iris		1X: 	15.30 fps 	2X: 	12.39 fps 	4X: 	04.33 fps 	
Proteus		1X: 	14.48 fps 	2X: 	12.64 fps 	4X: 	05.73 fps 	
Gaia		1X: 	06.92 fps 	2X: 	04.55 fps 	4X: 	02.32 fps 	
Nyx		1X: 	04.53 fps 	2X: 	04.21 fps 	
4X Slowmo		Apollo: 	16.55 fps 	APFast: 	65.84 fps 	Chronos: 	10.69 fps 	CHFast: 	15.78 fps
8

For most models it goes something like this: RTX GPU gives the most boost. If it’s super powerful, you’ll need more cores in your CPU to load and unload the GPU (and do other processes that the GPU cannot do). For an RTX 3080, at least 10 cores are needed. RAM frequency can increase speed. Lastly, since you can be dealing with large files, SSD abilities can be something that will hurt or help speed.

So, if you put in an RTX 3080, your CPU would most likely be the new bottleneck.

(AMDs 3DVcache does nothing as far as I know. New AMD GPUs are getting better, but don’t match the speeds of the Tensor enabled Nvidia cards.)