| 1 |
// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 2 |
| 2 |
// REQUIRES: riscv-registered-target |
| 3 |
// RUN: %clang_cc1 -triple riscv64 -target-feature +v -target-feature +zvfbfmin -target-feature +xsfvfwmaccqqq \ |
| 4 |
// RUN: -disable-O0-optnone -emit-llvm %s -o - | opt -S -passes=mem2reg | \ |
| 5 |
// RUN: FileCheck --check-prefix=CHECK-RV64 %s |
| 6 |
|
| 7 |
#include <sifive_vector.h> |
| 8 |
|
| 9 |
// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_sf_vfwmacc_4x4x4_f32mf2_tu |
| 10 |
// CHECK-RV64-SAME: (<vscale x 1 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] { |
| 11 |
// CHECK-RV64-NEXT: entry: |
| 12 |
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.sf.vfwmacc.4x4x4.nxv1f32.nxv4bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 [[VL]], i64 2) |
| 13 |
// CHECK-RV64-NEXT: ret <vscale x 1 x float> [[TMP0]] |
| 14 |
// |
| 15 |
vfloat32mf2_t test_sf_vfwmacc_4x4x4_f32mf2_tu(vfloat32mf2_t vd, vbfloat16m1_t vs1, vbfloat16mf4_t vs2, size_t vl) { |
| 16 |
return __riscv_sf_vfwmacc_4x4x4_tu(vd, vs1, vs2, vl); |
| 17 |
} |
| 18 |
|
| 19 |
// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_sf_vfwmacc_4x4x4_f32m1 |
| 20 |
// CHECK-RV64-SAME: (<vscale x 2 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] { |
| 21 |
// CHECK-RV64-NEXT: entry: |
| 22 |
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.sf.vfwmacc.4x4x4.nxv2f32.nxv4bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 [[VL]], i64 2) |
| 23 |
// CHECK-RV64-NEXT: ret <vscale x 2 x float> [[TMP0]] |
| 24 |
// |
| 25 |
vfloat32m1_t test_sf_vfwmacc_4x4x4_f32m1(vfloat32m1_t vd, vbfloat16m1_t vs1, vbfloat16mf2_t vs2, size_t vl) { |
| 26 |
return __riscv_sf_vfwmacc_4x4x4_tu(vd, vs1, vs2, vl); |
| 27 |
} |
| 28 |
|
| 29 |
// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_sf_vfwmacc_4x4x4_f32m2 |
| 30 |
// CHECK-RV64-SAME: (<vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] { |
| 31 |
// CHECK-RV64-NEXT: entry: |
| 32 |
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.sf.vfwmacc.4x4x4.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 [[VL]], i64 2) |
| 33 |
// CHECK-RV64-NEXT: ret <vscale x 4 x float> [[TMP0]] |
| 34 |
// |
| 35 |
vfloat32m2_t test_sf_vfwmacc_4x4x4_f32m2(vfloat32m2_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) { |
| 36 |
return __riscv_sf_vfwmacc_4x4x4_tu(vd, vs1, vs2, vl); |
| 37 |
} |
| 38 |
|
| 39 |
// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_sf_vfwmacc_4x4x4_f32m4 |
| 40 |
// CHECK-RV64-SAME: (<vscale x 8 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] { |
| 41 |
// CHECK-RV64-NEXT: entry: |
| 42 |
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.sf.vfwmacc.4x4x4.nxv8f32.nxv4bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 [[VL]], i64 2) |
| 43 |
// CHECK-RV64-NEXT: ret <vscale x 8 x float> [[TMP0]] |
| 44 |
// |
| 45 |
vfloat32m4_t test_sf_vfwmacc_4x4x4_f32m4(vfloat32m4_t vd, vbfloat16m1_t vs1, vbfloat16m2_t vs2, size_t vl) { |
| 46 |
return __riscv_sf_vfwmacc_4x4x4_tu(vd, vs1, vs2, vl); |
| 47 |
} |
| 48 |
|
| 49 |
// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_sf_vfwmacc_4x4x4_f32m8 |
| 50 |
// CHECK-RV64-SAME: (<vscale x 16 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] { |
| 51 |
// CHECK-RV64-NEXT: entry: |
| 52 |
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.sf.vfwmacc.4x4x4.nxv16f32.nxv4bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 [[VL]], i64 2) |
| 53 |
// CHECK-RV64-NEXT: ret <vscale x 16 x float> [[TMP0]] |
| 54 |
// |
| 55 |
vfloat32m8_t test_sf_vfwmacc_4x4x4_f32m8(vfloat32m8_t vd, vbfloat16m1_t vs1, vbfloat16m4_t vs2, size_t vl) { |
| 56 |
return __riscv_sf_vfwmacc_4x4x4_tu(vd, vs1, vs2, vl); |
| 57 |
} |