Skip to content

CUDA feature extraction#1152

Merged
kylophone merged 1 commit intomasterfrom
cuda_feature_extraction_rebased_squashed
Jan 11, 2023
Merged

CUDA feature extraction#1152
kylophone merged 1 commit intomasterfrom
cuda_feature_extraction_rebased_squashed

Conversation

@kylophone
Copy link
Copy Markdown
Collaborator

@kylophone kylophone commented Jan 7, 2023

Replaces #1080.

@kylophone kylophone mentioned this pull request Jan 7, 2023
Closed
@kylophone kylophone changed the title initial import: cuda feature extraction CUDA feature extraction Jan 7, 2023
@kylophone kylophone force-pushed the cuda_feature_extraction_rebased_squashed branch 2 times, most recently from 8fdd66c to ce4f563 Compare January 9, 2023 22:12
@kylophone @rarzumanyan
@nv-smiddelberg @MorkTheOrk @gedoensmax @mtavenrath
libvmaf: CUDA feature extraction
81c5cbf 
Co-authored-by: Roman Arzumanyan <rarzumanyan@nvidia.com>
Co-authored-by: Sven Middelberg <smiddelberg@nvidia.com>
Co-authored-by: Cem Moluluo <cmoluluo@nvidia.com>
Co-authored-by: Maximilian Mueller <maximilianm@nvidia.com>
Co-authored-by: Kyle Swanson <kswanson@netflix.com>
Co-authored-by: Markus Tavenrath <matavenrath@nvidia.com>
@kylophone kylophone force-pushed the cuda_feature_extraction_rebased_squashed branch from 7193979 to 81c5cbf Compare January 11, 2023 18:16
@kylophone kylophone merged commit 3071d7e into master Jan 11, 2023
@HunterAP23
Copy link
Copy Markdown

HunterAP23 commented Jan 11, 2023

I know this was merged, but is there any guide on how to build libvmaf for Windows under MINGW64 MYSYS2 with CUDA? I'm not able to find any method of putting CUDA on MINGW64 MYSYS2, so building this with CUDA enabled for Windows doesn't seem possible at the moment, unless there's some alternative method of doing so that isn't documented.

For reference, I've also tried Ubuntu through WSL2 but that seems to have a lot of errors, and it's not possible to cross-compile from there anyways.

@kylophone
Copy link
Copy Markdown
Collaborator Author

kylophone commented Jan 12, 2023

I have not personally tried this with Windows yet, I have been developing on Linux only. Could you please make an issue, we should be able to support it.

@HunterAP23 HunterAP23 mentioned this pull request Jan 12, 2023
Open
@ehmkhay
Copy link
Copy Markdown

ehmkhay commented Jan 25, 2023

Does this PR implement CUDA pipeline to ffmpeg -lavfi for in-GPU nvdec>vmaf (in linux)?
(or is this still in development?)
Thanks!

@kylophone
Copy link
Copy Markdown
Collaborator Author

kylophone commented Jan 25, 2023

Does this PR implement CUDA pipeline to ffmpeg -lavfi for in-GPU nvdec>vmaf (in linux)? (or is this still in development?) Thanks!

That work is in progress, you should see it on ffmpeg-devel soon.

@ehmkhay
Copy link
Copy Markdown

ehmkhay commented Jan 25, 2023

Brilliant, thanks!

@maxpain
Copy link
Copy Markdown

maxpain commented Jul 29, 2023

How to use cuda acceleration in vmaf today?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

4 participants

@kylophone @HunterAP23 @ehmkhay @maxpain
⚡