intel / intel-xpu-backend-for-triton

OpenAI Triton backend for Intel® GPUs
MIT License
143 stars 44 forks source link

Fix coalescing pass (#2760) #2764

Closed alexbaden closed 2 days ago

alexbaden commented 2 days ago

Fix Intel coalescing pass for cases where the result of a SCF loop (containing a coalescable block ptr load) is used by an operation with operands that do not have block ptr type (e.g. tt.reduce)


Signed-off-by: Tiotto, Ettore ettore.tiotto@intel.com (cherry picked from commit a8ca9e558026ff49c3bb74c6471c112b04f63d2d)