[Mlir-commits] [mlir] [MLIR][XeGPU] Improve deinterleave/interleave/dpas_mx ops handling (PR #197223)

[Igor Zamyatin]

================
@@ -96,6 +96,8 @@ void buildGPUPassPipeline(OpPassManager &pm,
     pm.addNestedPass<gpu::GPUModuleOp>(createLoopInvariantCodeMotionPass());
     pm.addNestedPass<gpu::GPUModuleOp>(createCSEPass());
     pm.addNestedPass<gpu::GPUModuleOp>(xegpu::createXeGPUVectorLinearize());
+    pm.addNestedPass<gpu::GPUModuleOp>(createCanonicalizerPass());
----------------
Garra1980 wrote:

do we really need so many CSE and canonicalizer passes here - one after SG to WI distribution, one after linearize. Can we combine those perhaps and call in just one place?

https://github.com/llvm/llvm-project/pull/197223