derek.holloway
7502bf64a4
CodeQL / Analyze (actions) (push) Canceled after 0s
CodeQL / Analyze (python) (push) Canceled after 0s
pre-commit / pre-commit (push) Canceled after 0s
Multi-Arch CI / setup (push) Canceled after 0s
Publish build_manylinux_x86_64 images / publish_build_manylinux_x86_64 (push) Canceled after 0s
Setup (Multi-Arch) / setup (push) Canceled after 0s
Multi-Arch CI / Linux::${{ fromJSON(needs.setup.outputs.linux_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubi10 image / publish_no_rocm_image_ubi10 (push) Canceled after 0s
Multi-Arch CI / Windows::${{ fromJSON(needs.setup.outputs.windows_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Multi-Arch CI / Manifest Diff (push) Canceled after 0s
TheRock Manifest Diff Report / Generate Manifest Diff Report (push) Canceled after 0s
Multi-Arch CI / CI Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_ocl_rt images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_rocgdb images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Unit Tests / Unit Tests :: ubuntu-24.04 (push) Canceled after 0s
Unit Tests / Unit Tests :: windows-2022 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_media (push) Canceled after 0s
Unit Tests / Unit Tests Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_openmpi (push) Canceled after 0s
117 lines
5.9 KiB
MLIR
117 lines
5.9 KiB
MLIR
// RUN: mlir-opt %s --test-vector-chained-reduction-folding-patterns | FileCheck %s
|
|
|
|
// CHECK-LABEL: func.func @reduce_1x_fp32(
|
|
// CHECK-SAME: %[[ARG0:.+]]: vector<8xf32>) -> f32 {
|
|
// CHECK-DAG: %[[CST:.+]] = arith.constant 0.0
|
|
// CHECK-NEXT: %[[RES:.+]] = vector.reduction <add>, %[[ARG0]], %[[CST]] : vector<8xf32> into f32
|
|
// CHECK-NEXT: return %[[RES]] : f32
|
|
func.func @reduce_1x_fp32(%arg0: vector<8xf32>) -> f32 {
|
|
%cst0 = arith.constant 0.0 : f32
|
|
%0 = vector.reduction <add>, %arg0, %cst0 : vector<8xf32> into f32
|
|
return %0 : f32
|
|
}
|
|
|
|
// CHECK-LABEL: func.func @reduce_2x_fp32(
|
|
// CHECK-SAME: %[[ARG0:.+]]: vector<8xf32>, %[[ARG1:.+]]: vector<8xf32>) -> f32 {
|
|
// CHECK-DAG: %[[CST:.+]] = arith.constant 0.0
|
|
// CHECK-DAG: %[[ADD:.+]] = arith.addf %[[ARG0]], %[[ARG1]] : vector<8xf32>
|
|
// CHECK-NEXT: %[[RES:.+]] = vector.reduction <add>, %[[ADD]], %[[CST]] : vector<8xf32> into f32
|
|
// CHECK-NEXT: return %[[RES]] : f32
|
|
func.func @reduce_2x_fp32(%arg0: vector<8xf32>, %arg1: vector<8xf32>) -> f32 {
|
|
%cst0 = arith.constant 0.0 : f32
|
|
%0 = vector.reduction <add>, %arg0, %cst0 : vector<8xf32> into f32
|
|
%1 = vector.reduction <add>, %arg1, %0 : vector<8xf32> into f32
|
|
return %1 : f32
|
|
}
|
|
|
|
// CHECK-LABEL: func.func @reduce_2x_no_acc_fp32(
|
|
// CHECK-SAME: %[[ARG0:.+]]: vector<8xf32>, %[[ARG1:.+]]: vector<8xf32>) -> f32 {
|
|
// CHECK: %[[ADD:.+]] = arith.addf %[[ARG0]], %[[ARG1]] : vector<8xf32>
|
|
// CHECK-NEXT: %[[RES:.+]] = vector.reduction <add>, %[[ADD]] : vector<8xf32> into f32
|
|
// CHECK-NEXT: return %[[RES]] : f32
|
|
func.func @reduce_2x_no_acc_fp32(%arg0: vector<8xf32>, %arg1: vector<8xf32>) -> f32 {
|
|
%0 = vector.reduction <add>, %arg0 : vector<8xf32> into f32
|
|
%1 = vector.reduction <add>, %arg1, %0 : vector<8xf32> into f32
|
|
return %1 : f32
|
|
}
|
|
|
|
// CHECK-LABEL: func.func @reduce_2x_zero_add_fp32(
|
|
// CHECK-SAME: %[[ARG0:.+]]: vector<8xf32>, %[[ARG1:.+]]: vector<8xf32>) -> f32 {
|
|
// CHECK: %[[ADD:.+]] = arith.addf %[[ARG0]], %[[ARG1]] : vector<8xf32>
|
|
// CHECK-NEXT: %[[RES:.+]] = vector.reduction <add>, %[[ADD]] : vector<8xf32> into f32
|
|
// CHECK-NEXT: return %[[RES]] : f32
|
|
func.func @reduce_2x_zero_add_fp32(%arg0: vector<8xf32>, %arg1: vector<8xf32>) -> f32 {
|
|
%cst0 = arith.constant dense<0.0> : vector<8xf32>
|
|
%x = arith.addf %arg0, %cst0 : vector<8xf32>
|
|
%0 = vector.reduction <add>, %x : vector<8xf32> into f32
|
|
%1 = vector.reduction <add>, %arg1, %0 : vector<8xf32> into f32
|
|
return %1 : f32
|
|
}
|
|
|
|
// CHECK-LABEL: func.func @reduce_3x_fp32(
|
|
// CHECK-SAME: %[[ARG0:.+]]: vector<8xf32>, %[[ARG1:.+]]: vector<8xf32>,
|
|
// CHECK-SAME: %[[ARG2:.+]]: vector<8xf32>) -> f32 {
|
|
// CHECK-DAG: %[[CST:.+]] = arith.constant 0.0
|
|
// CHECK-DAG: %[[ADD0:.+]] = arith.addf %[[ARG1]], %[[ARG2]] : vector<8xf32>
|
|
// CHECK-DAG: %[[ADD1:.+]] = arith.addf %[[ARG0]], %[[ADD0]] : vector<8xf32>
|
|
// CHECK-NEXT: %[[RES:.+]] = vector.reduction <add>, %[[ADD1]], %[[CST]] : vector<8xf32> into f32
|
|
// CHECK-NEXT: return %[[RES]] : f32
|
|
func.func @reduce_3x_fp32(%arg0: vector<8xf32>, %arg1: vector<8xf32>,
|
|
%arg2: vector<8xf32>) -> f32 {
|
|
%cst0 = arith.constant 0.0 : f32
|
|
%0 = vector.reduction <add>, %arg0, %cst0 : vector<8xf32> into f32
|
|
%1 = vector.reduction <add>, %arg1, %0 : vector<8xf32> into f32
|
|
%2 = vector.reduction <add>, %arg2, %1 : vector<8xf32> into f32
|
|
return %2 : f32
|
|
}
|
|
|
|
// CHECK-LABEL: func.func @reduce_1x_i32(
|
|
// CHECK-SAME: %[[ARG0:.+]]: vector<8xi32>) -> i32 {
|
|
// CHECK-DAG: %[[CST:.+]] = arith.constant 0
|
|
// CHECK-NEXT: %[[RES:.+]] = vector.reduction <add>, %[[ARG0]], %[[CST]] : vector<8xi32> into i32
|
|
// CHECK-NEXT: return %[[RES]] : i32
|
|
func.func @reduce_1x_i32(%arg0: vector<8xi32>) -> i32 {
|
|
%cst0 = arith.constant 0 : i32
|
|
%0 = vector.reduction <add>, %arg0, %cst0 : vector<8xi32> into i32
|
|
return %0 : i32
|
|
}
|
|
|
|
// CHECK-LABEL: func.func @reduce_2x_i32(
|
|
// CHECK-SAME: %[[ARG0:.+]]: vector<8xi32>, %[[ARG1:.+]]: vector<8xi32>) -> i32 {
|
|
// CHECK-DAG: %[[CST:.+]] = arith.constant 0
|
|
// CHECK-DAG: %[[ADD:.+]] = arith.addi %[[ARG0]], %[[ARG1]] : vector<8xi32>
|
|
// CHECK-NEXT: %[[RES:.+]] = vector.reduction <add>, %[[ADD]], %[[CST]] : vector<8xi32> into i32
|
|
// CHECK-NEXT: return %[[RES]] : i32
|
|
func.func @reduce_2x_i32(%arg0: vector<8xi32>, %arg1: vector<8xi32>) -> i32 {
|
|
%cst0 = arith.constant 0 : i32
|
|
%0 = vector.reduction <add>, %arg0, %cst0 : vector<8xi32> into i32
|
|
%1 = vector.reduction <add>, %arg1, %0 : vector<8xi32> into i32
|
|
return %1 : i32
|
|
}
|
|
|
|
// CHECK-LABEL: func.func @reduce_2x_no_acc_i32(
|
|
// CHECK-SAME: %[[ARG0:.+]]: vector<8xi32>, %[[ARG1:.+]]: vector<8xi32>) -> i32 {
|
|
// CHECK: %[[ADD:.+]] = arith.addi %[[ARG0]], %[[ARG1]] : vector<8xi32>
|
|
// CHECK-NEXT: %[[RES:.+]] = vector.reduction <add>, %[[ADD]] : vector<8xi32> into i32
|
|
// CHECK-NEXT: return %[[RES]] : i32
|
|
func.func @reduce_2x_no_acc_i32(%arg0: vector<8xi32>, %arg1: vector<8xi32>) -> i32 {
|
|
%0 = vector.reduction <add>, %arg0 : vector<8xi32> into i32
|
|
%1 = vector.reduction <add>, %arg1, %0 : vector<8xi32> into i32
|
|
return %1 : i32
|
|
}
|
|
|
|
// CHECK-LABEL: func.func @reduce_2x_zero_add_i32(
|
|
// CHECK-SAME: %[[ARG0:.+]]: vector<8xi32>, %[[ARG1:.+]]: vector<8xi32>) -> i32 {
|
|
// CHECK-DAG: %[[CST:.+]] = arith.constant 0
|
|
// CHECK-DAG: %[[ADD:.+]] = arith.addi %[[ARG0]], %[[ARG1]] : vector<8xi32>
|
|
// CHECK-NEXT: %[[RES:.+]] = vector.reduction <add>, %[[ADD]], %[[CST]] : vector<8xi32> into i32
|
|
// CHECK-NEXT: return %[[RES]] : i32
|
|
func.func @reduce_2x_zero_add_i32(%arg0: vector<8xi32>, %arg1: vector<8xi32>) -> i32 {
|
|
%cst0 = arith.constant 0 : i32
|
|
%cstV = arith.constant dense<0> : vector<8xi32>
|
|
%x = arith.addi %arg0, %cstV : vector<8xi32>
|
|
%0 = vector.reduction <add>, %x, %cst0 : vector<8xi32> into i32
|
|
%1 = vector.reduction <add>, %arg1, %0 : vector<8xi32> into i32
|
|
return %1 : i32
|
|
}
|