Skip to content

[Exp. Overhaul] Faster rendering (more GPU acceleration [CUDA/OpenCV/NVDEC/NVENC], Numba, optimization patches)#2552

Open
nycalexander wants to merge 5 commits intoZulko:masterfrom
nycalexander:master
Open

[Exp. Overhaul] Faster rendering (more GPU acceleration [CUDA/OpenCV/NVDEC/NVENC], Numba, optimization patches)#2552
nycalexander wants to merge 5 commits intoZulko:masterfrom
nycalexander:master

Conversation

@nycalexander
Copy link
Copy Markdown

Warning: this is completely experimental (not production ready, initially only meant to be used for myself), and NVIDIA focused. I have not tested pure CPU rendering, AMD cards, or Intel QSV.

Version tested on RTX 3000 series card successfully with no issues among multiple different workflows with notably less RAM usage, much faster rendering, less CPU overhead and more GPU utilization.

It isn't the best and fastest potential MoviePy can get to, all of these changes were done under current flow, safety and the least need to update API usage for the end user.

Users who want to try out this experimental version can use pip install -U --force-reinstall "moviepy[accelerate,gpu] @ git+https://github.com/nycalexander/moviefast.git" (virtual environment and CUDA Toolkit install recommended)

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

Successfully merging this pull request may close these issues.

1 participant