maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity

[VectorToGPU] Fix horizontal stride calculation for N-D memref 

Fix a bug in how we calculate the stride of mma load/store ops for N-D
memrefs

Differential Revision: https://reviews.llvm.org/D118378
This commit is contained in:
Thomas Raoux 2022-01-27 08:19:47 -08:00
parent 84fe34a0b7
commit a57ccad5a6
2 changed files with 31 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
mlir/lib/Conversion/VectorToGPU/VectorToGPU.cpp
View File

@ -60,13 +60,17 @@ getMemrefConstantHorizontalStride(ShapedType type) {
auto memrefType = type.dyn_cast<MemRefType>(); auto memrefType = type.dyn_cast<MemRefType>();
if (!memrefType) if (!memrefType)
return false; return false;
// If the memref is 0 or 1D the horizontal stride is 0.
if(memrefType.getRank() < 2)
return 0;
int64_t offset = 0; int64_t offset = 0;
SmallVector<int64_t, 2> strides; SmallVector<int64_t, 2> strides;
if (failed(getStridesAndOffset(memrefType, strides, offset))) if (failed(getStridesAndOffset(memrefType, strides, offset)))
return llvm::None; return llvm::None;
if (strides[0] == ShapedType::kDynamicStrideOrOffset) int64_t stride = strides[strides.size() - 2];
if (stride == ShapedType::kDynamicStrideOrOffset)
return llvm::None; return llvm::None;
return strides[0]; return stride;
} }
// Return true if the transfer op can be converted to a MMA matrix load. // Return true if the transfer op can be converted to a MMA matrix load.

30
mlir/test/Conversion/VectorToGPU/vector-to-mma-ops.mlir
View File

@ -4,11 +4,12 @@
#map1 = affine_map<(d0, d1, d2) -> (d0, d2)> #map1 = affine_map<(d0, d1, d2) -> (d0, d2)>
#map2 = affine_map<(d0, d1, d2) -> (d1, d2)> #map2 = affine_map<(d0, d1, d2) -> (d1, d2)>
#map3 = affine_map<(d0, d1, d2) -> (d0, d1)> #map3 = affine_map<(d0, d1, d2) -> (d0, d1)>
#map4 = affine_map<(d0) -> (d0, 0)>
// CHECK-LABEL: func @matmul // CHECK-LABEL: func @matmul
// CHECK-DAG: %[[A:.+]] = gpu.subgroup_mma_load_matrix %{{.*}}[%{{.*}}, %{{.*}}] {leadDimension = 16 : index} : memref<16x16xf16> -> !gpu.mma_matrix<16x16xf16, "AOp"> // CHECK-DAG: %[[A:.+]] = gpu.subgroup_mma_load_matrix %{{.*}}[%{{.*}}, %{{.*}}] {leadDimension = 16 : index} : memref<16x16xf16> -> !gpu.mma_matrix<16x16xf16, "AOp">
// CHECK-DAG: %[[B:.+]] = gpu.subgroup_mma_load_matrix %{{.*}}[%c0, %c0] {leadDimension = 16 : index} : memref<16x16xf16> -> !gpu.mma_matrix<16x16xf16, "BOp"> // CHECK-DAG: %[[B:.+]] = gpu.subgroup_mma_load_matrix %{{.*}}[%{{.*}}, %{{.*}}] {leadDimension = 16 : index} : memref<16x16xf16> -> !gpu.mma_matrix<16x16xf16, "BOp">
// CHECK-DAG: %[[C:.+]] = gpu.subgroup_mma_load_matrix %{{.*}}[%c0, %c0] {leadDimension = 16 : index} : memref<16x16xf16> -> !gpu.mma_matrix<16x16xf16, "COp"> // CHECK-DAG: %[[C:.+]] = gpu.subgroup_mma_load_matrix %{{.*}}[%{{.*}}, %{{.*}}] {leadDimension = 16 : index} : memref<16x16xf16> -> !gpu.mma_matrix<16x16xf16, "COp">
// CHECK: %[[D:.+]] = gpu.subgroup_mma_compute %[[A]], %[[B]], %[[C]] : !gpu.mma_matrix<16x16xf16, "AOp">, !gpu.mma_matrix<16x16xf16, "BOp"> -> !gpu.mma_matrix<16x16xf16, "COp"> // CHECK: %[[D:.+]] = gpu.subgroup_mma_compute %[[A]], %[[B]], %[[C]] : !gpu.mma_matrix<16x16xf16, "AOp">, !gpu.mma_matrix<16x16xf16, "BOp"> -> !gpu.mma_matrix<16x16xf16, "COp">
// CHECK: gpu.subgroup_mma_store_matrix %[[D]], %{{.*}}[%{{.*}}, %{{.*}}] {leadDimension = 16 : index} : !gpu.mma_matrix<16x16xf16, "COp">, memref<16x16xf16> // CHECK: gpu.subgroup_mma_store_matrix %[[D]], %{{.*}}[%{{.*}}, %{{.*}}] {leadDimension = 16 : index} : !gpu.mma_matrix<16x16xf16, "COp">, memref<16x16xf16>
func @matmul(%arg0: memref<16x16xf16>, %arg1: memref<16x16xf16>, %arg2: memref<16x16xf16>) { func @matmul(%arg0: memref<16x16xf16>, %arg1: memref<16x16xf16>, %arg2: memref<16x16xf16>) {
@ -26,7 +27,7 @@ func @matmul(%arg0: memref<16x16xf16>, %arg1: memref<16x16xf16>, %arg2: memref<1
// CHECK-LABEL: func @matmul_cst // CHECK-LABEL: func @matmul_cst
// CHECK-DAG: %[[CST:.+]] = arith.constant 0.000000e+00 : f16 // CHECK-DAG: %[[CST:.+]] = arith.constant 0.000000e+00 : f16
// CHECK-DAG: %[[A:.+]] = gpu.subgroup_mma_load_matrix %{{.*}}[%{{.*}}, %{{.*}}] {leadDimension = 16 : index} : memref<16x16xf16> -> !gpu.mma_matrix<16x16xf16, "AOp"> // CHECK-DAG: %[[A:.+]] = gpu.subgroup_mma_load_matrix %{{.*}}[%{{.*}}, %{{.*}}] {leadDimension = 16 : index} : memref<16x16xf16> -> !gpu.mma_matrix<16x16xf16, "AOp">
// CHECK-DAG: %[[B:.+]] = gpu.subgroup_mma_load_matrix %{{.*}}[%c0, %c0] {leadDimension = 16 : index} : memref<16x16xf16> -> !gpu.mma_matrix<16x16xf16, "BOp"> // CHECK-DAG: %[[B:.+]] = gpu.subgroup_mma_load_matrix %{{.*}}[%{{.*}}, %{{.*}}] {leadDimension = 16 : index} : memref<16x16xf16> -> !gpu.mma_matrix<16x16xf16, "BOp">
// CHECK-DAG: %[[C:.+]] = gpu.subgroup_mma_constant_matrix %[[CST]] : !gpu.mma_matrix<16x16xf16, "COp"> // CHECK-DAG: %[[C:.+]] = gpu.subgroup_mma_constant_matrix %[[CST]] : !gpu.mma_matrix<16x16xf16, "COp">
// CHECK: %[[D:.+]] = gpu.subgroup_mma_compute %[[A]], %[[B]], %[[C]] : !gpu.mma_matrix<16x16xf16, "AOp">, !gpu.mma_matrix<16x16xf16, "BOp"> -> !gpu.mma_matrix<16x16xf16, "COp"> // CHECK: %[[D:.+]] = gpu.subgroup_mma_compute %[[A]], %[[B]], %[[C]] : !gpu.mma_matrix<16x16xf16, "AOp">, !gpu.mma_matrix<16x16xf16, "BOp"> -> !gpu.mma_matrix<16x16xf16, "COp">
// CHECK: gpu.subgroup_mma_store_matrix %[[D]], %{{.*}}[%{{.*}}, %{{.*}}] {leadDimension = 16 : index} : !gpu.mma_matrix<16x16xf16, "COp">, memref<16x16xf16> // CHECK: gpu.subgroup_mma_store_matrix %[[D]], %{{.*}}[%{{.*}}, %{{.*}}] {leadDimension = 16 : index} : !gpu.mma_matrix<16x16xf16, "COp">, memref<16x16xf16>
@ -45,7 +46,7 @@ func @matmul_cst(%arg0: memref<16x16xf16>, %arg1: memref<16x16xf16>, %arg2: memr
// CHECK-SAME: (%{{.*}}: memref<16x16xf16>, %{{.*}}: memref<16x16xf16>, %{{.*}}: memref<16x16xf16>, %[[F:.*]]: f16) // CHECK-SAME: (%{{.*}}: memref<16x16xf16>, %{{.*}}: memref<16x16xf16>, %{{.*}}: memref<16x16xf16>, %[[F:.*]]: f16)
// CHECK-DAG: %[[C:.+]] = gpu.subgroup_mma_constant_matrix %[[F]] : !gpu.mma_matrix<16x16xf16, "COp"> // CHECK-DAG: %[[C:.+]] = gpu.subgroup_mma_constant_matrix %[[F]] : !gpu.mma_matrix<16x16xf16, "COp">
// CHECK-DAG: %[[A:.+]] = gpu.subgroup_mma_load_matrix %{{.*}}[%{{.*}}, %{{.*}}] {leadDimension = 16 : index} : memref<16x16xf16> -> !gpu.mma_matrix<16x16xf16, "AOp"> // CHECK-DAG: %[[A:.+]] = gpu.subgroup_mma_load_matrix %{{.*}}[%{{.*}}, %{{.*}}] {leadDimension = 16 : index} : memref<16x16xf16> -> !gpu.mma_matrix<16x16xf16, "AOp">
// CHECK-DAG: %[[B:.+]] = gpu.subgroup_mma_load_matrix %{{.*}}[%c0, %c0] {leadDimension = 16 : index} : memref<16x16xf16> -> !gpu.mma_matrix<16x16xf16, "BOp"> // CHECK-DAG: %[[B:.+]] = gpu.subgroup_mma_load_matrix %{{.*}}[%{{.*}}, %{{.*}}] {leadDimension = 16 : index} : memref<16x16xf16> -> !gpu.mma_matrix<16x16xf16, "BOp">
// CHECK: %[[D:.+]] = gpu.subgroup_mma_compute %[[A]], %[[B]], %[[C]] : !gpu.mma_matrix<16x16xf16, "AOp">, !gpu.mma_matrix<16x16xf16, "BOp"> -> !gpu.mma_matrix<16x16xf16, "COp"> // CHECK: %[[D:.+]] = gpu.subgroup_mma_compute %[[A]], %[[B]], %[[C]] : !gpu.mma_matrix<16x16xf16, "AOp">, !gpu.mma_matrix<16x16xf16, "BOp"> -> !gpu.mma_matrix<16x16xf16, "COp">
// CHECK: gpu.subgroup_mma_store_matrix %[[D]], %{{.*}}[%{{.*}}, %{{.*}}] {leadDimension = 16 : index} : !gpu.mma_matrix<16x16xf16, "COp">, memref<16x16xf16> // CHECK: gpu.subgroup_mma_store_matrix %[[D]], %{{.*}}[%{{.*}}, %{{.*}}] {leadDimension = 16 : index} : !gpu.mma_matrix<16x16xf16, "COp">, memref<16x16xf16>
func @matmul_broadcast(%arg0: memref<16x16xf16>, %arg1: memref<16x16xf16>, %arg2: memref<16x16xf16>, %f: f16) { func @matmul_broadcast(%arg0: memref<16x16xf16>, %arg1: memref<16x16xf16>, %arg2: memref<16x16xf16>, %f: f16) {
@ -88,7 +89,7 @@ func @matmul_loop(%arg0: memref<128x128xf16>, %arg1: memref<128x128xf16>, %arg2:
// CHECK-DAG: %[[CST_0:.+]] = arith.constant 0.000000e+00 : f16 // CHECK-DAG: %[[CST_0:.+]] = arith.constant 0.000000e+00 : f16
// CHECK-DAG: %[[CST_1:.+]] = arith.constant 1.000000e+00 : f16 // CHECK-DAG: %[[CST_1:.+]] = arith.constant 1.000000e+00 : f16
// CHECK-DAG: %[[A:.+]] = gpu.subgroup_mma_load_matrix %{{.*}}[%{{.*}}, %{{.*}}] {leadDimension = 16 : index} : memref<16x16xf16> -> !gpu.mma_matrix<16x16xf16, "AOp"> // CHECK-DAG: %[[A:.+]] = gpu.subgroup_mma_load_matrix %{{.*}}[%{{.*}}, %{{.*}}] {leadDimension = 16 : index} : memref<16x16xf16> -> !gpu.mma_matrix<16x16xf16, "AOp">
// CHECK-DAG: %[[B:.+]] = gpu.subgroup_mma_load_matrix %{{.*}}[%c0, %c0] {leadDimension = 16 : index} : memref<16x16xf16> -> !gpu.mma_matrix<16x16xf16, "BOp"> // CHECK-DAG: %[[B:.+]] = gpu.subgroup_mma_load_matrix %{{.*}}[%{{.*}}, %{{.*}}] {leadDimension = 16 : index} : memref<16x16xf16> -> !gpu.mma_matrix<16x16xf16, "BOp">
// CHECK-DAG: %[[C0:.+]] = gpu.subgroup_mma_constant_matrix %[[CST_0]] : !gpu.mma_matrix<16x16xf16, "COp"> // CHECK-DAG: %[[C0:.+]] = gpu.subgroup_mma_constant_matrix %[[CST_0]] : !gpu.mma_matrix<16x16xf16, "COp">
// CHECK-DAG: %[[C1:.+]] = gpu.subgroup_mma_constant_matrix %[[CST_1]] : !gpu.mma_matrix<16x16xf16, "COp"> // CHECK-DAG: %[[C1:.+]] = gpu.subgroup_mma_constant_matrix %[[CST_1]] : !gpu.mma_matrix<16x16xf16, "COp">
// CHECK: %[[D:.+]] = gpu.subgroup_mma_compute %[[A]], %[[B]], %[[C0]] : !gpu.mma_matrix<16x16xf16, "AOp">, !gpu.mma_matrix<16x16xf16, "BOp"> -> !gpu.mma_matrix<16x16xf16, "COp"> // CHECK: %[[D:.+]] = gpu.subgroup_mma_compute %[[A]], %[[B]], %[[C0]] : !gpu.mma_matrix<16x16xf16, "AOp">, !gpu.mma_matrix<16x16xf16, "BOp"> -> !gpu.mma_matrix<16x16xf16, "COp">
@ -131,3 +132,22 @@ func @matmul_fused_broadcast(%arg0: memref<16x16xf16>, %arg1: memref<16x16xf16>,
vector.transfer_write %F, %arg2[%c0, %c0] {in_bounds = [true, true]} : vector<16x16xf16>, memref<16x16xf16> vector.transfer_write %F, %arg2[%c0, %c0] {in_bounds = [true, true]} : vector<16x16xf16>, memref<16x16xf16>
return return
} }
// CHECK-LABEL: func @matmul_3Dmemref
// CHECK-DAG: %[[C0:.+]] = arith.constant 0 : index
// CHECK-DAG: %[[A:.+]] = gpu.subgroup_mma_load_matrix %{{.*}}[%[[C0]], %[[C0]], %[[C0]]] {leadDimension = 16 : index} : memref<2x16x16xf16> -> !gpu.mma_matrix<16x16xf16, "AOp">
// CHECK-DAG: %[[B:.+]] = gpu.subgroup_mma_load_matrix %{{.*}}[%[[C0]]] {leadDimension = 0 : index} : memref<16xf16> -> !gpu.mma_matrix<16x16xf16, "BOp">
// CHECK-DAG: %[[C:.+]] = gpu.subgroup_mma_load_matrix %{{.*}}[%[[C0]], %[[C0]], %[[C0]]] {leadDimension = 16 : index} : memref<2x16x16xf16> -> !gpu.mma_matrix<16x16xf16, "COp">
// CHECK: %[[D:.+]] = gpu.subgroup_mma_compute %[[A]], %[[B]], %[[C]] : !gpu.mma_matrix<16x16xf16, "AOp">, !gpu.mma_matrix<16x16xf16, "BOp"> -> !gpu.mma_matrix<16x16xf16, "COp">
// CHECK: gpu.subgroup_mma_store_matrix %[[D]], %{{.*}}[%[[C0]], %[[C0]], %[[C0]]] {leadDimension = 16 : index} : !gpu.mma_matrix<16x16xf16, "COp">, memref<2x16x16xf16>
func @matmul_3Dmemref(%arg0: memref<2x16x16xf16>, %arg1: memref<16xf16>, %arg2: memref<2x16x16xf16>) {
%cst_0 = arith.constant dense<0.000000e+00> : vector<16x16xf16>
%c0 = arith.constant 0 : index
%cst = arith.constant 0.000000e+00 : f16
%A = vector.transfer_read %arg0[%c0, %c0, %c0], %cst {in_bounds = [true, true]} : memref<2x16x16xf16>, vector<16x16xf16>
%B = vector.transfer_read %arg1[%c0], %cst {permutation_map = #map4, in_bounds = [true, true]} : memref<16xf16>, vector<16x16xf16>
%C = vector.transfer_read %arg2[%c0, %c0, %c0], %cst {in_bounds = [true, true]} : memref<2x16x16xf16>, vector<16x16xf16>
%D = vector.contract {indexing_maps = [#map1, #map2, #map3], iterator_types = ["parallel", "parallel", "reduction"], kind = #vector.kind<add>} %A, %B, %C : vector<16x16xf16>, vector<16x16xf16> into vector<16x16xf16>
vector.transfer_write %D, %arg2[%c0, %c0, %c0] {in_bounds = [true, true]} : vector<16x16xf16>, memref<2x16x16xf16>
return
}