Skip to content

fix: cuda optimizations#165

Merged
jasongdove merged 2 commits into
mainfrom
fix/maintain-filter-order-cuda-workaround
May 25, 2026
Merged

fix: cuda optimizations#165
jasongdove merged 2 commits into
mainfrom
fix/maintain-filter-order-cuda-workaround

Conversation

@jasongdove

Copy link
Copy Markdown
Contributor

optimizes cuda pipeline to convert to 8-bit early when encoder is 8-bit and hardware overlay can be used with 8-bit, instead of downloading and using software overlay with 10-bit content that will later be converted to 8-bit anyway.

@gemini-code-assist gemini-code-assist Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Code Review

This pull request introduces an eager conversion to 8-bit pixel formats when it enables the use of a hardware overlay filter. It also restructures how decoder filters are prepended to the filter chain to handle CUDA's hardware upload workaround. The review feedback suggests optimizing the eager conversion logic by checking if the input frame's bit depth is actually greater than 8 before performing the analysis, which avoids redundant state cloning and evaluation when the input is already 8-bit.

Comment thread crates/ffpipeline/src/filter_chain.rs
@jasongdove jasongdove merged commit 522e403 into main May 25, 2026
8 checks passed
@jasongdove jasongdove deleted the fix/maintain-filter-order-cuda-workaround branch May 25, 2026 18:01
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