[Mlir-commits] [mlir] [MLIR][XeGPU] Extend propagation and sg_to_lane distribution pass support broadcast with low rank and scalar source input (PR #170409)

[Charitha Saumya]

================
@@ -1855,9 +1982,9 @@ void xegpu::populateXeGPUSubgroupDistributePatterns(
   patterns.add<CreateNdDescDistribution, StoreNdDistribution,
                LoadNdDistribution, DpasDistribution, PrefetchNdDistribution,
                GpuBarrierDistribution, VectorMultiReductionDistribution,
-               LoadDistribution, StoreDistribution, VectorTransposeDistribution,
-               VectorBitcastDistribution, LoadMatrixDistribution,
-               StoreMatrixDistribution,
+               VectorBroadcastDistribution, LoadDistribution, StoreDistribution,
----------------
charithaintc wrote:

important: this should move to `highPatternBenefit` patterns in line 1993. This makes sure that this pattern is priortized over vector distribution pattern. this is very important for debugging and general functionality. 

https://github.com/llvm/llvm-project/pull/170409