[llvm-branch-commits] [mlir] [mlir][linalg] Enable scalable vectorization of linalg.unpack (PR #149293)

https://github.com/hanhanW edited https://github.com/llvm/llvm-project/pull/149293 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits

[llvm-branch-commits] [mlir] [mlir][linalg] Enable scalable vectorization of linalg.unpack (PR #149293)

https://github.com/hanhanW approved this pull request. LGTM, thanks! Just 3 nits. https://github.com/llvm/llvm-project/pull/149293 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/l

[llvm-branch-commits] [mlir] [mlir][linalg] Enable scalable vectorization of linalg.unpack (PR #149293)

@@ -1871,115 +1872,98 @@ static VectorType getCollapsedVecType(VectorType type, return VectorType::get(newShape, type.getElementType(), newScalableFlags); } -/// Vectorize a `linalg::UnPackOp` to these 4 Ops: -/// Vector::TransferReadOp - Reads a vector from the source ten

[llvm-branch-commits] [mlir] [mlir][linalg] Enable scalable vectorization of linalg.unpack (PR #149293)

@@ -1876,22 +1872,35 @@ static VectorType getCollapsedVecType(VectorType type, return VectorType::get(newShape, type.getElementType(), newScalableFlags); } -/// Vectorize a `linalg::UnPackOp` to these 4 Ops: -/// Vector::TransferReadOp - Reads a vector from the source tens

[llvm-branch-commits] [mlir] [mlir][linalg] Enable scalable vectorization of linalg.unpack (PR #149293)

@@ -1871,115 +1872,98 @@ static VectorType getCollapsedVecType(VectorType type, return VectorType::get(newShape, type.getElementType(), newScalableFlags); } -/// Vectorize a `linalg::UnPackOp` to these 4 Ops: -/// Vector::TransferReadOp - Reads a vector from the source ten

[llvm-branch-commits] [mlir] [mlir][linalg] Restrict linalg.pack to not have artificial padding. (PR #149624)

https://github.com/hanhanW closed https://github.com/llvm/llvm-project/pull/149624 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits

[llvm-branch-commits] [mlir] [mlir][linalg] Restrict linalg.pack to not have artificial padding. (PR #149624)

@@ -4717,6 +4697,12 @@ static LogicalResult commonVerifierPackAndUnPackOp(OpTy packOrUnPack) { return op->emitError("mismatch in inner tile sizes specified and shaped of " "tiled dimension in the packed type"); } + if (failed(verifyCompatibleSh

[llvm-branch-commits] [mlir] [mlir][linalg] Restrict linalg.pack to not have artificial padding. (PR #149624)

@@ -1824,27 +1825,47 @@ func.func @unpack_invalid_outer_dims_perm(%source: tensor<128x256xf32>, %dest: t // - +func.func @pack_with_artificial_padding(%input: tensor<9xf32>, %output: tensor<3x8xf32>) -> tensor<3x8xf32> { + %cst = arith.constant 0.0 : f32 + // expected

[llvm-branch-commits] [mlir] [mlir][linalg] Restrict linalg.pack to not have artificial padding. (PR #149624)

@@ -150,9 +150,15 @@ def Linalg_PackOp : Linalg_RelayoutOp<"pack", [ `padding_value` specifies a padding value at the boundary on non-perfectly divisible dimensions. Padding is optional: -- If absent, it is UB if the tile does not perfectly divide the dimension. +

[llvm-branch-commits] [mlir] [mlir][linalg] Restrict linalg.pack to not have artificial padding. (PR #149624)

hanhanW wrote: Yes, updated. Thanks for pointing it out! https://github.com/llvm/llvm-project/pull/149624 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/

[llvm-branch-commits] [mlir] [mlir][linalg] Restrict linalg.pack to not have artificial padding. (PR #149624)

https://github.com/hanhanW updated https://github.com/llvm/llvm-project/pull/149624 >From 6de929abea2ad4fac10605e333050b64eedbb904 Mon Sep 17 00:00:00 2001 From: hanhanW Date: Tue, 22 Jul 2025 15:15:08 -0700 Subject: [PATCH 1/4] [mlir][linalg] Restrict linalg.pack to not have extra padding siz

[llvm-branch-commits] [mlir] [mlir][linalg] Restrict linalg.pack to not have artificial padding. (PR #149624)

@@ -150,9 +150,15 @@ def Linalg_PackOp : Linalg_RelayoutOp<"pack", [ `padding_value` specifies a padding value at the boundary on non-perfectly divisible dimensions. Padding is optional: -- If absent, it is UB if the tile does not perfectly divide the dimension. +

[llvm-branch-commits] [mlir] [mlir][linalg] Restrict linalg.pack to not have artificial padding. (PR #149624)

@@ -150,9 +150,13 @@ def Linalg_PackOp : Linalg_RelayoutOp<"pack", [ `padding_value` specifies a padding value at the boundary on non-perfectly divisible dimensions. Padding is optional: -- If absent, it is UB if the tile does not perfectly divide the dimension. +

[llvm-branch-commits] [mlir] [mlir][linalg] Restrict linalg.pack to not have artificial padding. (PR #149624)

https://github.com/hanhanW edited https://github.com/llvm/llvm-project/pull/149624 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits

[llvm-branch-commits] [mlir] [mlir][linalg] Restrict linalg.pack to not have artificial padding. (PR #149624)

hanhanW wrote: > think you should drop @2548809 from this PR, so we could handle #150127 > separately :) Or target that branch from this one if you're planning to get > that one in first. Otherwise LGTM. I need to land the other one first. Otherwise, the CI is not happy. I recreate the other

[llvm-branch-commits] [mlir] [mlir][linalg] Restrict linalg.pack to not have artificial padding. (PR #149624)

https://github.com/hanhanW edited https://github.com/llvm/llvm-project/pull/149624 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits