3 Pull requests merged by 3 people

• plot_layout.py Refactoring

  #775 merged Apr 15, 2025

• Fix pid remapping logic when GRID_MN cannot divide NUM_XCDS

  #722 merged Apr 11, 2025

• Fix the bwd Mode in flash-attention.py

  #772 merged Apr 11, 2025

3 Pull requests opened by 3 people

• update scale dot assertion in plot_layout.py

  #774 opened Apr 10, 2025

• [AMD] Added bufferOps refinement

  #776 opened Apr 14, 2025

• Cherry-pick more performance improvmenet commits

  #778 opened Apr 17, 2025

1 Issue opened by 1 person

• [Issue]: Performance degradation on mamba2's triton kernel

  #777 opened Apr 15, 2025

4 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• [Issue]: Compiling with amdclang fails due to `-Werror`

  #708 commented on Apr 11, 2025 • 0 new comments

• Rebasing upstream changes into refine-ops-pass

  #753 commented on Apr 17, 2025 • 0 new comments

• [Draft] Adding support for refining ElementWise, ExpandDims and Broadcast

  #763 commented on Apr 17, 2025 • 0 new comments

• Optimize RMSNorm backward pass

  #769 commented on Apr 16, 2025 • 0 new comments