[XPU] Conditionally add elementwise optimization pass to the pipeline

Conditionally add `-tritonintelgpu-optimize-elementwise-parallelism` to the
pipeline.

Signed-off-by: victor-eds <[email protected]>

Author: victor-eds

third_party/intel/backend/compiler.py

@@ -251,8 +251,9 @@ def make_ttgir(mod, metadata, opt, properties):
         passes.common.add_cse(pm)
         passes.ttgpuir.add_prefetch(pm)
         passes.ttgpuir.add_optimize_dot_operands(pm, True)
-        if os.getenv("TRITON_INTEL_OPTIMIZE_REDUCTION_LOCALITY", "0") == 1:
+        if os.getenv("TRITON_INTEL_OPTIMIZE_REDUCTION_LOCALITY", "0") == "1":
             intel.passes.ttgpuir.add_optimize_reduction_locality(pm)
+            intel.passes.ttgpuir.add_optimize_elementwise_parallelism(pm)
         intel.passes.ttgpuir.add_remove_layout_conversions(pm)
         intel.passes.ttgpuir.add_reduce_data_duplication(pm)
         passes.ttgpuir.add_reorder_instructions(pm)

third_party/intel/triton_xpu.cc

@@ -100,6 +100,9 @@ void init_triton_intel_passes_ttgpuir(py::module &&m) {
                      gpu::intel::createTritonIntelGPUMaterializeBlockPointer);
   ADD_PASS_WRAPPER_0("add_optimize_reduction_locality",
                      gpu::intel::createTritonIntelGPUOptimizeReductionLocality);
+  ADD_PASS_WRAPPER_0(
+      "add_optimize_elementwise_parallelism",
+      gpu::intel::createTritonIntelGPUOptimizeElementwiseParallelism);
 }
 
 void init_triton_intel(py::module &&m) {