Skip to content

feat: new syrk passes + lowering#1908

Merged
avik-pal merged 7 commits intomainfrom
ap/to_syrk
Nov 30, 2025
Merged

feat: new syrk passes + lowering#1908
avik-pal merged 7 commits intomainfrom
ap/to_syrk

Conversation

@avik-pal
Copy link
Copy Markdown
Collaborator

No description provided.

Comment thread src/Compiler.jl Outdated

if !is_sharded
# these passes don't have optimized sharding implementations
append!(transform_passes_list, ["dot_general_to_syrk"])
Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

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

can't we just (inefficiently) lower syrk back to a dot general at the end, if not supported

Copy link
Copy Markdown
Collaborator Author

@avik-pal avik-pal Nov 25, 2025

Choose a reason for hiding this comment

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

we do that as fallback.

But shardy passes are run post this transform, so a transform here would mean an implicit all-gather on CPU.

Comment thread test/integration/linear_algebra.jl Outdated
@avik-pal avik-pal marked this pull request as ready for review November 30, 2025 18:08
@avik-pal avik-pal merged commit 409fdf7 into main Nov 30, 2025
70 checks passed
@avik-pal avik-pal deleted the ap/to_syrk branch November 30, 2025 18:08
@avik-pal avik-pal restored the ap/to_syrk branch April 18, 2026 12:46
@giordano giordano deleted the ap/to_syrk branch April 18, 2026 18:21
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.

2 participants