[clang] [RISCV][clang] Add Zvfbfwma C intrinsics support (PR #79615)
via cfe-commits
cfe-commits at lists.llvm.org
Fri Jan 26 08:48:05 PST 2024
llvmbot wrote:
<!--LLVM PR SUMMARY COMMENT-->
@llvm/pr-subscribers-clang
Author: Brandon Wu (4vtomat)
<details>
<summary>Changes</summary>
---
Patch is 204.88 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/79615.diff
5 Files Affected:
- (modified) clang/include/clang/Basic/riscv_vector.td (+14)
- (added) clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/non-overloaded/vfwmaccbf16.c (+479)
- (added) clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/overloaded/vfwmaccbf16.c (+469)
- (added) clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/non-overloaded/vfwmaccbf16.c (+808)
- (added) clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/overloaded/vfwmaccbf16.c (+952)
``````````diff
diff --git a/clang/include/clang/Basic/riscv_vector.td b/clang/include/clang/Basic/riscv_vector.td
index a00ca353588ed5..40a77fa709b7ae 100644
--- a/clang/include/clang/Basic/riscv_vector.td
+++ b/clang/include/clang/Basic/riscv_vector.td
@@ -1730,12 +1730,26 @@ let ManualCodegen = [{
defm vfwnmacc : RVVFloatingWidenTerBuiltinSetRoundingMode;
defm vfwmsac : RVVFloatingWidenTerBuiltinSetRoundingMode;
defm vfwnmsac : RVVFloatingWidenTerBuiltinSetRoundingMode;
+
+ // Vector BF16 widening multiply-accumulate
+ let Log2LMUL = [-2, -1, 0, 1, 2],
+ HasMaskedOffOperand = false in
+ defm vfwmaccbf16 : RVVOutOp1Op2BuiltinSet<"vfwmaccbf16", "y",
+ [["vv", "Fw", "FwFwvvu"],
+ ["vf", "Fw", "FwFwevu"]]>;
}
// 13.7. Vector Widening Floating-Point Fused Multiply-Add Instructions
defm vfwmacc : RVVFloatingWidenTerBuiltinSet;
defm vfwnmacc : RVVFloatingWidenTerBuiltinSet;
defm vfwmsac : RVVFloatingWidenTerBuiltinSet;
defm vfwnmsac : RVVFloatingWidenTerBuiltinSet;
+
+ // Vector BF16 widening multiply-accumulate
+ let Log2LMUL = [-2, -1, 0, 1, 2],
+ HasMaskedOffOperand = false in
+ defm vfwmaccbf16 : RVVOutOp1Op2BuiltinSet<"vfwmaccbf16", "y",
+ [["vv", "Fw", "FwFwvv"],
+ ["vf", "Fw", "FwFwev"]]>;
}
}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/non-overloaded/vfwmaccbf16.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/non-overloaded/vfwmaccbf16.c
new file mode 100644
index 00000000000000..75888573d6f344
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/non-overloaded/vfwmaccbf16.c
@@ -0,0 +1,479 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 4
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v -target-feature +zfh \
+// RUN: -target-feature +zvfh -target-feature +experimental-zvfbfwma -disable-O0-optnone \
+// RUN: -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN: FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vv_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmaccbf16.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vv_f32mf2(vfloat32mf2_t vd, vbfloat16mf4_t vs1,
+ vbfloat16mf4_t vs2, size_t vl) {
+ return __riscv_vfwmaccbf16_vv_f32mf2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vf_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmaccbf16.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2(vfloat32mf2_t vd, __bf16 vs1,
+ vbfloat16mf4_t vs2, size_t vl) {
+ return __riscv_vfwmaccbf16_vf_f32mf2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmaccbf16.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1(vfloat32m1_t vd, vbfloat16mf2_t vs1,
+ vbfloat16mf2_t vs2, size_t vl) {
+ return __riscv_vfwmaccbf16_vv_f32m1(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmaccbf16.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1(vfloat32m1_t vd, __bf16 vs1,
+ vbfloat16mf2_t vs2, size_t vl) {
+ return __riscv_vfwmaccbf16_vf_f32m1(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2(
+// 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]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmaccbf16.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2(vfloat32m2_t vd, vbfloat16m1_t vs1,
+ vbfloat16m1_t vs2, size_t vl) {
+ return __riscv_vfwmaccbf16_vv_f32m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmaccbf16.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2(vfloat32m2_t vd, __bf16 vs1,
+ vbfloat16m1_t vs2, size_t vl) {
+ return __riscv_vfwmaccbf16_vf_f32m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmaccbf16.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4(vfloat32m4_t vd, vbfloat16m2_t vs1,
+ vbfloat16m2_t vs2, size_t vl) {
+ return __riscv_vfwmaccbf16_vv_f32m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmaccbf16.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4(vfloat32m4_t vd, __bf16 vs1,
+ vbfloat16m2_t vs2, size_t vl) {
+ return __riscv_vfwmaccbf16_vf_f32m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmaccbf16.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8(vfloat32m8_t vd, vbfloat16m4_t vs1,
+ vbfloat16m4_t vs2, size_t vl) {
+ return __riscv_vfwmaccbf16_vv_f32m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmaccbf16.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8(vfloat32m8_t vd, __bf16 vs1,
+ vbfloat16m4_t vs2, size_t vl) {
+ return __riscv_vfwmaccbf16_vf_f32m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vv_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmaccbf16.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vv_f32mf2_m(vbool64_t vm, vfloat32mf2_t vd,
+ vbfloat16mf4_t vs1,
+ vbfloat16mf4_t vs2, size_t vl) {
+ return __riscv_vfwmaccbf16_vv_f32mf2_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vf_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmaccbf16.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2_m(vbool64_t vm, vfloat32mf2_t vd,
+ __bf16 vs1, vbfloat16mf4_t vs2,
+ size_t vl) {
+ return __riscv_vfwmaccbf16_vf_f32mf2_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmaccbf16.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1_m(vbool32_t vm, vfloat32m1_t vd,
+ vbfloat16mf2_t vs1, vbfloat16mf2_t vs2,
+ size_t vl) {
+ return __riscv_vfwmaccbf16_vv_f32m1_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmaccbf16.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1_m(vbool32_t vm, vfloat32m1_t vd,
+ __bf16 vs1, vbfloat16mf2_t vs2,
+ size_t vl) {
+ return __riscv_vfwmaccbf16_vf_f32m1_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmaccbf16.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2_m(vbool16_t vm, vfloat32m2_t vd,
+ vbfloat16m1_t vs1, vbfloat16m1_t vs2,
+ size_t vl) {
+ return __riscv_vfwmaccbf16_vv_f32m2_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmaccbf16.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2_m(vbool16_t vm, vfloat32m2_t vd,
+ __bf16 vs1, vbfloat16m1_t vs2,
+ size_t vl) {
+ return __riscv_vfwmaccbf16_vf_f32m2_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmaccbf16.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4_m(vbool8_t vm, vfloat32m4_t vd,
+ vbfloat16m2_t vs1, vbfloat16m2_t vs2,
+ size_t vl) {
+ return __riscv_vfwmaccbf16_vv_f32m4_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmaccbf16.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4_m(vbool8_t vm, vfloat32m4_t vd,
+ __bf16 vs1, vbfloat16m2_t vs2,
+ size_t vl) {
+ return __riscv_vfwmaccbf16_vf_f32m4_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmaccbf16.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8_m(vbool4_t vm, vfloat32m8_t vd,
+ vbfloat16m4_t vs1, vbfloat16m4_t vs2,
+ size_t vl) {
+ return __riscv_vfwmaccbf16_vv_f32m8_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmaccbf16.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8_m(vbool4_t vm, vfloat32m8_t vd,
+ __bf16 vs1, vbfloat16m4_t vs2,
+ size_t vl) {
+ return __riscv_vfwmaccbf16_vf_f32m8_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vv_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmaccbf16.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vv_f32mf2_rm(vfloat32mf2_t vd,
+ vbfloat16mf4_t vs1,
+ vbfloat16mf4_t vs2, size_t vl) {
+ return __riscv_vfwmaccbf16_vv_f32mf2_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vf_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT: entry:
+// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmaccbf16.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT: ret <vscale x ...
[truncated]
``````````
</details>
https://github.com/llvm/llvm-project/pull/79615
More information about the cfe-commits
mailing list