[Mlir-commits] [mlir] ac1e22f - [mlir][vector] Generalize folding of ext-contractionOp to other types. (#96593)

[llvmlistbot at llvm.org]

Author: Stanley Winata
Date: 2024-06-25T09:29:43-07:00
New Revision: ac1e22f3053f761e4e2ef832b92de15876e68335

URL: https://github.com/llvm/llvm-project/commit/ac1e22f3053f761e4e2ef832b92de15876e68335
DIFF: https://github.com/llvm/llvm-project/commit/ac1e22f3053f761e4e2ef832b92de15876e68335.diff

LOG: [mlir][vector] Generalize folding of ext-contractionOp to other types. (#96593)

Many state of the art models and quantization operations are now
directly working on vector.contract on integers.

This commit enables generalizes ext-contraction folding S.T we can emit
more performant vector.contracts on codegen pipelines.

Signed-off-by: Stanley Winata <stanley.winata at amd.com>

Added: 
    

Modified: 
    mlir/lib/Dialect/Vector/Transforms/VectorTransforms.cpp
    mlir/test/Dialect/Vector/fold-arith-extf-into-vector-contract.mlir

Removed: 
    


################################################################################
diff  --git a/mlir/lib/Dialect/Vector/Transforms/VectorTransforms.cpp b/mlir/lib/Dialect/Vector/Transforms/VectorTransforms.cpp
index eac6db585aad7..da3d9648cf283 100644
--- a/mlir/lib/Dialect/Vector/Transforms/VectorTransforms.cpp
+++ b/mlir/lib/Dialect/Vector/Transforms/VectorTransforms.cpp
@@ -1552,6 +1552,7 @@ struct CanonicalizeContractMatmulToMMT final
 /// Cores, i.e, `mma.sync.*.f32.f16.f16.f32` and `mma.sync.*.f32.bf16.bf16.f32`.
 /// This pattern folds the arithmetic extensions into the vector contraction and
 /// enables the usage of native mixed precision Tensor Core instructions.
+template <typename ExtOp>
 struct FoldArithExtIntoContractionOp
     : public OpRewritePattern<vector::ContractionOp> {
   using OpRewritePattern::OpRewritePattern;
@@ -1559,8 +1560,8 @@ struct FoldArithExtIntoContractionOp
   LogicalResult matchAndRewrite(vector::ContractionOp contractOp,
                                 PatternRewriter &rewriter) const override {
 
-    auto lhsDefOp = contractOp.getLhs().getDefiningOp<arith::ExtFOp>();
-    auto rhsDefOp = contractOp.getRhs().getDefiningOp<arith::ExtFOp>();
+    auto lhsDefOp = contractOp.getLhs().getDefiningOp<ExtOp>();
+    auto rhsDefOp = contractOp.getRhs().getDefiningOp<ExtOp>();
 
     if (!lhsDefOp || !rhsDefOp) {
       return rewriter.notifyMatchFailure(contractOp,
@@ -1895,7 +1896,9 @@ struct FoldArithToVectorOuterProduct : public OpRewritePattern<MulOpType> {
 
 void mlir::vector::populateFoldArithExtensionPatterns(
     RewritePatternSet &patterns) {
-  patterns.add<FoldArithExtIntoContractionOp>(patterns.getContext());
+  patterns.add<FoldArithExtIntoContractionOp<arith::ExtFOp>,
+               FoldArithExtIntoContractionOp<arith::ExtSIOp>>(
+      patterns.getContext());
 }
 
 void mlir::vector::populateVectorMaskMaterializationPatterns(

diff  --git a/mlir/test/Dialect/Vector/fold-arith-extf-into-vector-contract.mlir b/mlir/test/Dialect/Vector/fold-arith-extf-into-vector-contract.mlir
index 31ae126906f21..6dbde7afbdd33 100644
--- a/mlir/test/Dialect/Vector/fold-arith-extf-into-vector-contract.mlir
+++ b/mlir/test/Dialect/Vector/fold-arith-extf-into-vector-contract.mlir
@@ -48,3 +48,25 @@ func.func @fold_arith_extf_into_contract_scalable(
       %lhs_f32, %rhs_f32, %arg2 : vector<[64]x64xf32>, vector<64x64xf32> into vector<[64]x64xf32>
     return %result : vector<[64]x64xf32>
 }
+
+// -----
+
+// CHECK-LABEL: func.func @fold_arith_extsi_into_contract
+//  CHECK-SAME: (%[[ARG0:.*]]: vector<64x64xi8>, %[[ARG1:.*]]: vector<64x64xi8>, %[[ARG2:.*]]: vector<64x64xi32>)
+//  CHECK-NEXT:   %[[R:.+]] = vector.contract
+//  CHECK-SAME:   iterator_types = ["parallel", "parallel", "reduction"], kind = #vector.kind<add>}
+//  CHECK-SAME:   %[[ARG0]], %[[ARG1]], %[[ARG2]] : vector<64x64xi8>, vector<64x64xi8> into vector<64x64xi32>
+//  CHECK-NEXT:   return %[[R]] : vector<64x64xi32>
+func.func @fold_arith_extsi_into_contract(
+  %arg0: vector<64x64xi8>,
+  %arg1: vector<64x64xi8>,
+  %arg2: vector<64x64xi32>) -> vector<64x64xi32> {
+    %lhs_i32 = arith.extsi %arg0 : vector<64x64xi8> to vector<64x64xi32>
+    %rhs_i32 = arith.extsi %arg1 : vector<64x64xi8> to vector<64x64xi32>
+    %result = vector.contract {
+      indexing_maps = [affine_map<(d0, d1, d2) -> (d0, d2)>, affine_map<(d0, d1, d2) -> (d2, d1)>, affine_map<(d0, d1, d2) -> (d0, d1)>],
+      iterator_types = ["parallel", "parallel", "reduction"],
+      kind = #vector.kind<add>}
+      %lhs_i32, %rhs_i32, %arg2 : vector<64x64xi32>, vector<64x64xi32> into vector<64x64xi32>
+    return %result : vector<64x64xi32>
+}