Ivan Butygin 91f3cdbd4f [mlir][gpu] Pattern to promote gpu.shuffle to specialized AMDGPU ops (#137109) ...

Only swizzle promotion for now, may add DPP ops support later.

2025-05-13 13:26:46 +03:00

..

all-reduce-add.mlir

…

all-reduce-maxf.mlir

…

async-region.mlir

…

barrier-elimination.mlir

…

bufferization-buffer-deallocation.mlir

…

canonicalize.mlir

…

decompose-memrefs.mlir

…

dynamic-shared-memory.mlir

…

globalId-rewrite.mlir

…

indirect-device-func-call.mlir

…

int-range-interface.mlir

…

invalid.mlir

[mlir] GPUToROCDL: Add support for non-i32/f32 shuffle types (#136320)

2025-04-18 19:53:24 +03:00

mapping.mlir

…

module-to-binary-invalid.mlir

…

module-to-binary-nvvm.mlir

…

module-to-binary-rocdl.mlir

…

module-to-binary-spirv.mlir

…

multiple-all-reduce.mlir

…

nvvm-attach-target.mlir

…

ops.mlir

…

outlining.mlir

…

promote-shuffle-amdgpu.mlir

[mlir][gpu] Pattern to promote gpu.shuffle to specialized AMDGPU ops (#137109)

2025-05-13 13:26:46 +03:00

promotion.mlir

…

shuffle-rewrite.mlir

[mlir] GPUToROCDL: Fix crashes with unsupported shuffle datatypes (#135504)

2025-04-13 20:26:19 +02:00

sink-ops.mlir

…

sparse-roundtrip.mlir

…

spirv-attach-targets.mlir

…

subgroup-mma-vector-unroll.mlir

…

subgroup-reduce-lowering.mlir

[mlir][AMDGPU] Implement gpu.subgroup_reduce with DPP intrinsics on AMD GPUs (#133204)

2025-04-23 17:37:32 -07:00

subgroupId-rewrite.mlir

[MLIR][GPU] Use arith instead of index for subgroup_id (#137843)

2025-04-30 09:03:24 -04:00

test-nvvm-pipeline.mlir

…

transform-gpu-failing.mlir

…

transform-gpu.mlir

…

value-bounds-op-interface-impl.mlir

…