Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

System resources not being fully utilized #57

Open
FluffyDiscord opened this issue Dec 12, 2023 · 7 comments
Open

System resources not being fully utilized #57

FluffyDiscord opened this issue Dec 12, 2023 · 7 comments

Comments

@FluffyDiscord
Copy link

FluffyDiscord commented Dec 12, 2023

I have rather high end PC and upscaling videos are slow. Current system usage seems low. Would it be possible to speed up the upscaling for higher end setups? I remember other setups to take less time and use more system resources

image

Video is ~3 minutes long.
image

Setup:
R7 5800X3D
32GB RAM
RTX 4090
Samsung Pro NVME SSD

@Djdefrag
Copy link
Owner

Hi my friend,

This probably happens because the cpu can't send enough images to the gpu to upscale, considering that the gpu is extremely faster than the cpu (in general, not your pc, which has monster components).

At the moment I set it to use only 1 cpu, because even trying with more than 1 the total time to upscale a video remained the same, at least considering my gpu. But yours, which is extremely more powerful, can actually handle many more operations.

This problem will be partly solved with the next version, which will be focused solely on performance.
In particular there will be a new AI engine that is the result of more than a year of work (I'm waiting for Microsoft to release it stably) that will bring:

  • better GPU and CPU utilization
  • better management of VRAM
  • and most importantly a huge jump in performance (~50%+)

However I would also like to add a widget to choose how many cpu to use for the AI engine, this could help a lot especially in your case, that is with a very high performance pc.

@FluffyDiscord
Copy link
Author

The widget would be great - could it be done before the release of the updated version that waits for Microsoft release?

Also I got to the point where ffmpeg was stitching the images back together. It also didnt use fully my CPU and the SSD was barely used, so probably not bottlenecked by IO. Could you also add oltion to specify threads ? Would love to use all of them if possible (there's hard limit of 16 AFAIK for ffmpeg)

@Djdefrag
Copy link
Owner

Yes, my friend

I am actually working on it and it will be released as soon as possible.

For ffmpeg, to set the number of cpus to use you can change the CPU number widget.

Consider that CPU means thread, so if your cpu has 16 threads and you want to use them all you have to enter 16

@FluffyDiscord
Copy link
Author

Just a heads up, the CPU number that I enter to the GUI does not seem to be added to ffmpeg command line:

Screenshot_2

@Djdefrag
Copy link
Owner

Hi,

i do not use ffmpeg directly, but with moviepy library.

despite not being in the commandline, then actually all the cpus are used?

@FluffyDiscord
Copy link
Author

I don't think all threads are used, I entered 8 and I saw the same usage as on the screen, around 15-20%, which would indicate 3-4 threads

@Djdefrag
Copy link
Owner

This is strange, i clearly see all my cpu used.

But i will investigate further, thank you my friend

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants