Skip to content

Feat/fine grained int8 transpose load#19

Open
he-sheng-jie wants to merge 4 commits into
masterfrom
feat/fine-grained-int8-transpose-load
Open

Feat/fine grained int8 transpose load#19
he-sheng-jie wants to merge 4 commits into
masterfrom
feat/fine-grained-int8-transpose-load

Conversation

@he-sheng-jie

Copy link
Copy Markdown

Summary

  • Implement fine-grained load support for matrix memory loaders.
  • Add int8 transpose load path for A/B matrix loads.
  • Add byte-mask propagation through loaders, MatrixReg writes, LocalMMU requests, and difftest finish events.
  • Wire TaskController load metadata for tail handling, beat count, and transpose mode.

@he-sheng-jie he-sheng-jie requested a review from Wonicon June 5, 2026 01:53
@Wonicon

Wonicon commented Jun 9, 2026

Copy link
Copy Markdown
Contributor

The previous merge commit still contains compilation errors. The conflict zone was not resolved correctly, leaving behind duplicate code and incorrectly deleted sections. The latest commit fixes this.

The difftest updates (OpenXiangShan/difftest#887) is built on feat-matrix, which is required by the ongoing OpenXiangShan/XSAI#68. The current XSAI master is incompatible with the difftest version needed for this PR.

Should we wait for OpenXiangShan/XSAI#68 to merge first?

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