[clang] [RISCV][clang] Add Zvfbfwma C intrinsics support (PR #79615)

Brandon Wu via cfe-commits cfe-commits at lists.llvm.org
Sat Feb 3 02:29:57 PST 2024


https://github.com/4vtomat updated https://github.com/llvm/llvm-project/pull/79615

>From 68d30971eea8a1445a131d8f913dbc4852d88591 Mon Sep 17 00:00:00 2001
From: Brandon Wu <brandon.wu at sifive.com>
Date: Fri, 26 Jan 2024 08:44:10 -0800
Subject: [PATCH 1/2] [RISCV][clang] Add Zvfbfwma C intrinsics support

---
 clang/include/clang/Basic/riscv_vector.td     |  14 +
 .../non-policy/non-overloaded/vfwmaccbf16.c   | 479 +++++++++
 .../non-policy/overloaded/vfwmaccbf16.c       | 469 +++++++++
 .../policy/non-overloaded/vfwmaccbf16.c       | 808 +++++++++++++++
 .../policy/overloaded/vfwmaccbf16.c           | 952 ++++++++++++++++++
 5 files changed, 2722 insertions(+)
 create mode 100644 clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/non-overloaded/vfwmaccbf16.c
 create mode 100644 clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/overloaded/vfwmaccbf16.c
 create mode 100644 clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/non-overloaded/vfwmaccbf16.c
 create mode 100644 clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/overloaded/vfwmaccbf16.c

diff --git a/clang/include/clang/Basic/riscv_vector.td b/clang/include/clang/Basic/riscv_vector.td
index a00ca353588ed..40a77fa709b7a 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 0000000000000..75888573d6f34
--- /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 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2_rm(vfloat32mf2_t vd, __bf16 vs1,
+                                            vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32mf2_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1_rm(
+// 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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1_rm(vfloat32m1_t vd, vbfloat16mf2_t vs1,
+                                          vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m1_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1_rm(
+// 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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1_rm(vfloat32m1_t vd, __bf16 vs1,
+                                          vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m1_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2_rm(
+// 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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2_rm(vfloat32m2_t vd, vbfloat16m1_t vs1,
+                                          vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m2_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2_rm(
+// 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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2_rm(vfloat32m2_t vd, __bf16 vs1,
+                                          vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m2_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4_rm(
+// 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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4_rm(vfloat32m4_t vd, vbfloat16m2_t vs1,
+                                          vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m4_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4_rm(
+// 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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4_rm(vfloat32m4_t vd, __bf16 vs1,
+                                          vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m4_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8_rm(
+// 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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8_rm(vfloat32m8_t vd, vbfloat16m4_t vs1,
+                                          vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m8_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8_rm(
+// 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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8_rm(vfloat32m8_t vd, __bf16 vs1,
+                                          vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m8_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vv_f32mf2_rm_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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vv_f32mf2_rm_m(vbool64_t vm, vfloat32mf2_t vd,
+                                              vbfloat16mf4_t vs1,
+                                              vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32mf2_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                            vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vf_f32mf2_rm_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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2_rm_m(vbool64_t vm, vfloat32mf2_t vd,
+                                              __bf16 vs1, vbfloat16mf4_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32mf2_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                            vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1_rm_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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vd,
+                                            vbfloat16mf2_t vs1,
+                                            vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m1_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                           vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1_rm_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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vd,
+                                            __bf16 vs1, vbfloat16mf2_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m1_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                           vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2_rm_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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vd,
+                                            vbfloat16m1_t vs1,
+                                            vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m2_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                           vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2_rm_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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vd,
+                                            __bf16 vs1, vbfloat16m1_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m2_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                           vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4_rm_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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vd,
+                                            vbfloat16m2_t vs1,
+                                            vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m4_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                           vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4_rm_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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vd,
+                                            __bf16 vs1, vbfloat16m2_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m4_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                           vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8_rm_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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vd,
+                                            vbfloat16m4_t vs1,
+                                            vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m8_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                           vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8_rm_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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vd,
+                                            __bf16 vs1, vbfloat16m4_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m8_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                           vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/overloaded/vfwmaccbf16.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/overloaded/vfwmaccbf16.c
new file mode 100644
index 0000000000000..9a66ad264fbe7
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/overloaded/vfwmaccbf16.c
@@ -0,0 +1,469 @@
+// 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(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(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(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(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(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(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(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(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(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(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(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(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(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(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(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(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(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(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(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(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(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 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2_rm(vfloat32mf2_t vd, __bf16 vs1,
+                                            vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1_rm(
+// 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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1_rm(vfloat32m1_t vd, vbfloat16mf2_t vs1,
+                                          vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1_rm(
+// 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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1_rm(vfloat32m1_t vd, __bf16 vs1,
+                                          vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2_rm(
+// 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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2_rm(vfloat32m2_t vd, vbfloat16m1_t vs1,
+                                          vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2_rm(
+// 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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2_rm(vfloat32m2_t vd, __bf16 vs1,
+                                          vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4_rm(
+// 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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4_rm(vfloat32m4_t vd, vbfloat16m2_t vs1,
+                                          vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4_rm(
+// 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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4_rm(vfloat32m4_t vd, __bf16 vs1,
+                                          vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8_rm(
+// 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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8_rm(vfloat32m8_t vd, vbfloat16m4_t vs1,
+                                          vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8_rm(
+// 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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8_rm(vfloat32m8_t vd, __bf16 vs1,
+                                          vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vv_f32mf2_rm_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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vv_f32mf2_rm_m(vbool64_t vm, vfloat32mf2_t vd,
+                                              vbfloat16mf4_t vs1,
+                                              vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vf_f32mf2_rm_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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2_rm_m(vbool64_t vm, vfloat32mf2_t vd,
+                                              __bf16 vs1, vbfloat16mf4_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmaccbf16(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1_rm_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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vd,
+                                            vbfloat16mf2_t vs1,
+                                            vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1_rm_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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vd,
+                                            __bf16 vs1, vbfloat16mf2_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmaccbf16(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2_rm_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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vd,
+                                            vbfloat16m1_t vs1,
+                                            vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2_rm_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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vd,
+                                            __bf16 vs1, vbfloat16m1_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmaccbf16(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4_rm_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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vd,
+                                            vbfloat16m2_t vs1,
+                                            vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4_rm_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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vd,
+                                            __bf16 vs1, vbfloat16m2_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmaccbf16(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8_rm_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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vd,
+                                            vbfloat16m4_t vs1,
+                                            vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8_rm_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 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vd,
+                                            __bf16 vs1, vbfloat16m4_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmaccbf16(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/non-overloaded/vfwmaccbf16.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/non-overloaded/vfwmaccbf16.c
new file mode 100644
index 0000000000000..1e1aaf52a30af
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/non-overloaded/vfwmaccbf16.c
@@ -0,0 +1,808 @@
+// 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_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vv_f32mf2_tu(vfloat32mf2_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32mf2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vf_f32mf2_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2_tu(vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32mf2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1_tu(vfloat32m1_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m1_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1_tu(vfloat32m1_t vd, __bf16 vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m1_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2_tu(vfloat32m2_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2_tu(vfloat32m2_t vd, __bf16 vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4_tu(vfloat32m4_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4_tu(vfloat32m4_t vd, __bf16 vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8_tu(vfloat32m8_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m8_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8_tu(vfloat32m8_t vd, __bf16 vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m8_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vv_f32mf2_tum(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vv_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32mf2_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vf_f32mf2_tum(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32mf2_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1_tum(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1_tum(vbool32_t vm, vfloat32m1_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m1_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1_tum(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1_tum(vbool32_t vm, vfloat32m1_t vd, __bf16 vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m1_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2_tum(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2_tum(vbool16_t vm, vfloat32m2_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m2_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2_tum(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2_tum(vbool16_t vm, vfloat32m2_t vd, __bf16 vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m2_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4_tum(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4_tum(vbool8_t vm, vfloat32m4_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m4_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4_tum(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4_tum(vbool8_t vm, vfloat32m4_t vd, __bf16 vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m4_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8_tum(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8_tum(vbool4_t vm, vfloat32m8_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m8_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8_tum(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8_tum(vbool4_t vm, vfloat32m8_t vd, __bf16 vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m8_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vv_f32mf2_tumu(
+// 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 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vv_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32mf2_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vf_f32mf2_tumu(
+// 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 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32mf2_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1_tumu(
+// 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 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m1_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1_tumu(
+// 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 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd, __bf16 vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m1_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2_tumu(
+// 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 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m2_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2_tumu(
+// 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 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd, __bf16 vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m2_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4_tumu(
+// 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 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m4_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4_tumu(
+// 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 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd, __bf16 vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m4_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8_tumu(
+// 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 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m8_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8_tumu(
+// 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 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd, __bf16 vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m8_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vv_f32mf2_mu(
+// 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 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vv_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32mf2_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vf_f32mf2_mu(
+// 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 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32mf2_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1_mu(
+// 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 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1_mu(vbool32_t vm, vfloat32m1_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m1_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1_mu(
+// 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 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1_mu(vbool32_t vm, vfloat32m1_t vd, __bf16 vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m1_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2_mu(
+// 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 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2_mu(vbool16_t vm, vfloat32m2_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m2_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2_mu(
+// 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 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2_mu(vbool16_t vm, vfloat32m2_t vd, __bf16 vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m2_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4_mu(
+// 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 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4_mu(vbool8_t vm, vfloat32m4_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m4_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4_mu(
+// 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 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4_mu(vbool8_t vm, vfloat32m4_t vd, __bf16 vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m4_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8_mu(
+// 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 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8_mu(vbool4_t vm, vfloat32m8_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m8_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8_mu(
+// 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 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8_mu(vbool4_t vm, vfloat32m8_t vd, __bf16 vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m8_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vv_f32mf2_rm_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vv_f32mf2_rm_tu(vfloat32mf2_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32mf2_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vf_f32mf2_rm_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2_rm_tu(vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32mf2_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1_rm_tu(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1_rm_tu(vfloat32m1_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m1_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1_rm_tu(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1_rm_tu(vfloat32m1_t vd, __bf16 vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m1_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2_rm_tu(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2_rm_tu(vfloat32m2_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m2_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2_rm_tu(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2_rm_tu(vfloat32m2_t vd, __bf16 vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m2_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4_rm_tu(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4_rm_tu(vfloat32m4_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m4_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4_rm_tu(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4_rm_tu(vfloat32m4_t vd, __bf16 vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m4_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8_rm_tu(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8_rm_tu(vfloat32m8_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m8_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8_rm_tu(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8_rm_tu(vfloat32m8_t vd, __bf16 vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m8_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vv_f32mf2_rm_tum(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vv_f32mf2_rm_tum(vbool64_t vm, vfloat32mf2_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32mf2_rm_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vf_f32mf2_rm_tum(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2_rm_tum(vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32mf2_rm_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1_rm_tum(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m1_rm_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1_rm_tum(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd, __bf16 vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m1_rm_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2_rm_tum(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m2_rm_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2_rm_tum(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd, __bf16 vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m2_rm_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4_rm_tum(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m4_rm_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4_rm_tum(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd, __bf16 vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m4_rm_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8_rm_tum(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m8_rm_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8_rm_tum(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd, __bf16 vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m8_rm_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vv_f32mf2_rm_tumu(
+// 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 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vv_f32mf2_rm_tumu(vbool64_t vm, vfloat32mf2_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32mf2_rm_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vf_f32mf2_rm_tumu(
+// 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 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2_rm_tumu(vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32mf2_rm_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1_rm_tumu(
+// 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 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m1_rm_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1_rm_tumu(
+// 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 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd, __bf16 vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m1_rm_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2_rm_tumu(
+// 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 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m2_rm_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2_rm_tumu(
+// 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 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd, __bf16 vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m2_rm_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4_rm_tumu(
+// 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 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m4_rm_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4_rm_tumu(
+// 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 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd, __bf16 vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m4_rm_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8_rm_tumu(
+// 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 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m8_rm_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8_rm_tumu(
+// 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 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd, __bf16 vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m8_rm_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vv_f32mf2_rm_mu(
+// 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 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vv_f32mf2_rm_mu(vbool64_t vm, vfloat32mf2_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32mf2_rm_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vf_f32mf2_rm_mu(
+// 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 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2_rm_mu(vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32mf2_rm_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1_rm_mu(
+// 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 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m1_rm_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1_rm_mu(
+// 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 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd, __bf16 vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m1_rm_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2_rm_mu(
+// 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 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m2_rm_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2_rm_mu(
+// 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 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd, __bf16 vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m2_rm_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4_rm_mu(
+// 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 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m4_rm_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4_rm_mu(
+// 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 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd, __bf16 vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m4_rm_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8_rm_mu(
+// 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 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vv_f32m8_rm_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8_rm_mu(
+// 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 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd, __bf16 vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_vf_f32m8_rm_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/overloaded/vfwmaccbf16.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/overloaded/vfwmaccbf16.c
new file mode 100644
index 0000000000000..96ea2e943a8de
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/overloaded/vfwmaccbf16.c
@@ -0,0 +1,952 @@
+// 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_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vv_f32mf2_tu(vfloat32mf2_t vd,
+                                            vbfloat16mf4_t vs1,
+                                            vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vf_f32mf2_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2_tu(vfloat32mf2_t vd, __bf16 vs1,
+                                            vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1_tu(vfloat32m1_t vd, vbfloat16mf2_t vs1,
+                                          vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1_tu(vfloat32m1_t vd, __bf16 vs1,
+                                          vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2_tu(vfloat32m2_t vd, vbfloat16m1_t vs1,
+                                          vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2_tu(vfloat32m2_t vd, __bf16 vs1,
+                                          vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4_tu(vfloat32m4_t vd, vbfloat16m2_t vs1,
+                                          vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4_tu(vfloat32m4_t vd, __bf16 vs1,
+                                          vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8_tu(vfloat32m8_t vd, vbfloat16m4_t vs1,
+                                          vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8_tu(vfloat32m8_t vd, __bf16 vs1,
+                                          vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vv_f32mf2_tum(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vv_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                             vbfloat16mf4_t vs1,
+                                             vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vf_f32mf2_tum(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                             __bf16 vs1, vbfloat16mf4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwmaccbf16_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1_tum(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                           vbfloat16mf2_t vs1,
+                                           vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1_tum(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                           __bf16 vs1, vbfloat16mf2_t vs2,
+                                           size_t vl) {
+  return __riscv_vfwmaccbf16_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2_tum(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                           vbfloat16m1_t vs1, vbfloat16m1_t vs2,
+                                           size_t vl) {
+  return __riscv_vfwmaccbf16_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2_tum(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                           __bf16 vs1, vbfloat16m1_t vs2,
+                                           size_t vl) {
+  return __riscv_vfwmaccbf16_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4_tum(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                           vbfloat16m2_t vs1, vbfloat16m2_t vs2,
+                                           size_t vl) {
+  return __riscv_vfwmaccbf16_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4_tum(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                           __bf16 vs1, vbfloat16m2_t vs2,
+                                           size_t vl) {
+  return __riscv_vfwmaccbf16_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8_tum(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                           vbfloat16m4_t vs1, vbfloat16m4_t vs2,
+                                           size_t vl) {
+  return __riscv_vfwmaccbf16_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8_tum(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                           __bf16 vs1, vbfloat16m4_t vs2,
+                                           size_t vl) {
+  return __riscv_vfwmaccbf16_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vv_f32mf2_tumu(
+// 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 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vv_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                              vbfloat16mf4_t vs1,
+                                              vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vf_f32mf2_tumu(
+// 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 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                              __bf16 vs1, vbfloat16mf4_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmaccbf16_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1_tumu(
+// 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 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                            vbfloat16mf2_t vs1,
+                                            vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1_tumu(
+// 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 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                            __bf16 vs1, vbfloat16mf2_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmaccbf16_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2_tumu(
+// 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 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                            vbfloat16m1_t vs1,
+                                            vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2_tumu(
+// 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 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                            __bf16 vs1, vbfloat16m1_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmaccbf16_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4_tumu(
+// 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 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                            vbfloat16m2_t vs1,
+                                            vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4_tumu(
+// 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 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                            __bf16 vs1, vbfloat16m2_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmaccbf16_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8_tumu(
+// 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 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                            vbfloat16m4_t vs1,
+                                            vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8_tumu(
+// 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 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                            __bf16 vs1, vbfloat16m4_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmaccbf16_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vv_f32mf2_mu(
+// 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 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vv_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                            vbfloat16mf4_t vs1,
+                                            vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vf_f32mf2_mu(
+// 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 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                            __bf16 vs1, vbfloat16mf4_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmaccbf16_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1_mu(
+// 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 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                          vbfloat16mf2_t vs1,
+                                          vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1_mu(
+// 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 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                          __bf16 vs1, vbfloat16mf2_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwmaccbf16_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2_mu(
+// 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 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                          vbfloat16m1_t vs1, vbfloat16m1_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwmaccbf16_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2_mu(
+// 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 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                          __bf16 vs1, vbfloat16m1_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwmaccbf16_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4_mu(
+// 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 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                          vbfloat16m2_t vs1, vbfloat16m2_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwmaccbf16_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4_mu(
+// 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 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                          __bf16 vs1, vbfloat16m2_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwmaccbf16_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8_mu(
+// 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 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                          vbfloat16m4_t vs1, vbfloat16m4_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwmaccbf16_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8_mu(
+// 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 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                          __bf16 vs1, vbfloat16m4_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwmaccbf16_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vv_f32mf2_rm_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vv_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs1,
+                                               vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vf_f32mf2_rm_tu(
+// 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 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2_rm_tu(vfloat32mf2_t vd, __bf16 vs1,
+                                               vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1_rm_tu(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1_rm_tu(vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs1,
+                                             vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1_rm_tu(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1_rm_tu(vfloat32m1_t vd, __bf16 vs1,
+                                             vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2_rm_tu(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2_rm_tu(vfloat32m2_t vd, vbfloat16m1_t vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2_rm_tu(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2_rm_tu(vfloat32m2_t vd, __bf16 vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4_rm_tu(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4_rm_tu(vfloat32m4_t vd, vbfloat16m2_t vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4_rm_tu(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4_rm_tu(vfloat32m4_t vd, __bf16 vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8_rm_tu(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8_rm_tu(vfloat32m8_t vd, vbfloat16m4_t vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8_rm_tu(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8_rm_tu(vfloat32m8_t vd, __bf16 vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vv_f32mf2_rm_tum(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vv_f32mf2_rm_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vf_f32mf2_rm_tum(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2_rm_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                __bf16 vs1, vbfloat16mf4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmaccbf16_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1_rm_tum(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1_rm_tum(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              __bf16 vs1, vbfloat16mf2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmaccbf16_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2_rm_tum(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                              vbfloat16m1_t vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2_rm_tum(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                              __bf16 vs1, vbfloat16m1_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmaccbf16_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4_rm_tum(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                              vbfloat16m2_t vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4_rm_tum(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                              __bf16 vs1, vbfloat16m2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmaccbf16_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8_rm_tum(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                              vbfloat16m4_t vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8_rm_tum(
+// 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 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                              __bf16 vs1, vbfloat16m4_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmaccbf16_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vv_f32mf2_rm_tumu(
+// 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 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vv_f32mf2_rm_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmaccbf16_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vf_f32mf2_rm_tumu(
+// 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 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2_rm_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 __bf16 vs1, vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmaccbf16_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1_rm_tumu(
+// 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 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1_rm_tumu(
+// 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 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               __bf16 vs1, vbfloat16mf2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmaccbf16_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2_rm_tumu(
+// 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 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs1,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2_rm_tumu(
+// 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 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                               __bf16 vs1, vbfloat16m1_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmaccbf16_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4_rm_tumu(
+// 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 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs1,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4_rm_tumu(
+// 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 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                               __bf16 vs1, vbfloat16m2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmaccbf16_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8_rm_tumu(
+// 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 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs1,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8_rm_tumu(
+// 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 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                               __bf16 vs1, vbfloat16m4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmaccbf16_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vv_f32mf2_rm_mu(
+// 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 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vv_f32mf2_rm_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs1,
+                                               vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmaccbf16_vf_f32mf2_rm_mu(
+// 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 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmaccbf16_vf_f32mf2_rm_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               __bf16 vs1, vbfloat16mf4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmaccbf16_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vv_f32m1_rm_mu(
+// 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 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vv_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs1,
+                                             vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmaccbf16_vf_f32m1_rm_mu(
+// 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 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmaccbf16_vf_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             __bf16 vs1, vbfloat16mf2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwmaccbf16_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vv_f32m2_rm_mu(
+// 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 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vv_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                             vbfloat16m1_t vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmaccbf16_vf_f32m2_rm_mu(
+// 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 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmaccbf16_vf_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                             __bf16 vs1, vbfloat16m1_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwmaccbf16_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vv_f32m4_rm_mu(
+// 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 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vv_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                             vbfloat16m2_t vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmaccbf16_vf_f32m4_rm_mu(
+// 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 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmaccbf16_vf_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                             __bf16 vs1, vbfloat16m2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwmaccbf16_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vv_f32m8_rm_mu(
+// 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 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vv_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                             vbfloat16m4_t vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmaccbf16_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmaccbf16_vf_f32m8_rm_mu(
+// 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 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmaccbf16_vf_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                             __bf16 vs1, vbfloat16m4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwmaccbf16_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}

>From cd6782e524e76a0eb7b4b7797ecb025eca31de4f Mon Sep 17 00:00:00 2001
From: Brandon Wu <brandon.wu at sifive.com>
Date: Sat, 3 Feb 2024 02:29:44 -0800
Subject: [PATCH 2/2] fixup! [RISCV][clang] Add Zvfbfwma C intrinsics support

---
 clang/include/clang/Basic/riscv_vector.td                     | 2 ++
 clang/include/clang/Support/RISCVVIntrinsicUtils.h            | 3 ++-
 clang/lib/Sema/SemaRISCVVectorLookup.cpp                      | 1 +
 .../non-policy/non-overloaded/vfwmaccbf16.c                   | 4 ++--
 .../non-policy/overloaded/vfwmaccbf16.c                       | 4 ++--
 .../policy/non-overloaded/vfwmaccbf16.c                       | 4 ++--
 .../policy/overloaded/vfwmaccbf16.c                           | 4 ++--
 clang/test/Sema/rvv-required-features-invalid.c               | 4 ++++
 clang/utils/TableGen/RISCVVEmitter.cpp                        | 1 +
 9 files changed, 18 insertions(+), 9 deletions(-)

diff --git a/clang/include/clang/Basic/riscv_vector.td b/clang/include/clang/Basic/riscv_vector.td
index 40a77fa709b7a..93532fc4fc940 100644
--- a/clang/include/clang/Basic/riscv_vector.td
+++ b/clang/include/clang/Basic/riscv_vector.td
@@ -1733,6 +1733,7 @@ let ManualCodegen = [{
 
     // Vector BF16 widening multiply-accumulate
     let Log2LMUL = [-2, -1, 0, 1, 2],
+        RequiredFeatures = ["Zvfbfwma"],
         HasMaskedOffOperand = false in
     defm vfwmaccbf16 : RVVOutOp1Op2BuiltinSet<"vfwmaccbf16", "y",
                                               [["vv", "Fw", "FwFwvvu"],
@@ -1746,6 +1747,7 @@ let ManualCodegen = [{
 
   // Vector BF16 widening multiply-accumulate
   let Log2LMUL = [-2, -1, 0, 1, 2],
+      RequiredFeatures = ["Zvfbfwma"],
       HasMaskedOffOperand = false in
   defm vfwmaccbf16 : RVVOutOp1Op2BuiltinSet<"vfwmaccbf16", "y",
                                             [["vv", "Fw", "FwFwvv"],
diff --git a/clang/include/clang/Support/RISCVVIntrinsicUtils.h b/clang/include/clang/Support/RISCVVIntrinsicUtils.h
index 15416a3393441..30bf36edb7bfc 100644
--- a/clang/include/clang/Support/RISCVVIntrinsicUtils.h
+++ b/clang/include/clang/Support/RISCVVIntrinsicUtils.h
@@ -501,7 +501,8 @@ enum RVVRequire : uint32_t {
   RVV_REQ_Zvknhb = 1 << 13,
   RVV_REQ_Zvksed = 1 << 14,
   RVV_REQ_Zvksh = 1 << 15,
-  RVV_REQ_Experimental = 1 << 16,
+  RVV_REQ_Zvfbfwma = 1 << 16,
+  RVV_REQ_Experimental = 1 << 17,
 
   LLVM_MARK_AS_BITMASK_ENUM(RVV_REQ_Experimental)
 };
diff --git a/clang/lib/Sema/SemaRISCVVectorLookup.cpp b/clang/lib/Sema/SemaRISCVVectorLookup.cpp
index 25d5d20be7c36..bf89a4ac51afd 100644
--- a/clang/lib/Sema/SemaRISCVVectorLookup.cpp
+++ b/clang/lib/Sema/SemaRISCVVectorLookup.cpp
@@ -215,6 +215,7 @@ void RISCVIntrinsicManagerImpl::ConstructRVVIntrinsics(
       {"zvknhb", RVV_REQ_Zvknhb},
       {"zvksed", RVV_REQ_Zvksed},
       {"zvksh", RVV_REQ_Zvksh},
+      {"zvfbfwma", RVV_REQ_Zvfbfwma},
       {"experimental", RVV_REQ_Experimental}};
 
   // Construction of RVVIntrinsicRecords need to sync with createRVVIntrinsics
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
index 75888573d6f34..b02d5f0861af4 100644
--- 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
@@ -1,7 +1,7 @@
 // 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: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfwma -disable-O0-optnone  \
 // RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
 // RUN:   FileCheck --check-prefix=CHECK-RV64 %s
 
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/overloaded/vfwmaccbf16.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/overloaded/vfwmaccbf16.c
index 9a66ad264fbe7..2583fccb77ae0 100644
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/overloaded/vfwmaccbf16.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/overloaded/vfwmaccbf16.c
@@ -1,7 +1,7 @@
 // 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: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfwma -disable-O0-optnone  \
 // RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
 // RUN:   FileCheck --check-prefix=CHECK-RV64 %s
 
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/non-overloaded/vfwmaccbf16.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/non-overloaded/vfwmaccbf16.c
index 1e1aaf52a30af..0174eba06d22b 100644
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/non-overloaded/vfwmaccbf16.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/non-overloaded/vfwmaccbf16.c
@@ -1,7 +1,7 @@
 // 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: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfwma -disable-O0-optnone  \
 // RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
 // RUN:   FileCheck --check-prefix=CHECK-RV64 %s
 
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/overloaded/vfwmaccbf16.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/overloaded/vfwmaccbf16.c
index 96ea2e943a8de..14eb333fb82cd 100644
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/overloaded/vfwmaccbf16.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/overloaded/vfwmaccbf16.c
@@ -1,7 +1,7 @@
 // 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: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfwma -disable-O0-optnone  \
 // RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
 // RUN:   FileCheck --check-prefix=CHECK-RV64 %s
 
diff --git a/clang/test/Sema/rvv-required-features-invalid.c b/clang/test/Sema/rvv-required-features-invalid.c
index 933348eae5507..e73ffda95ab1a 100644
--- a/clang/test/Sema/rvv-required-features-invalid.c
+++ b/clang/test/Sema/rvv-required-features-invalid.c
@@ -35,3 +35,7 @@ void test_xsfvfnrclipxfqf() {
 void test_xsfvfnrclipxufqf() {
   __riscv_sf_vfnrclip_xu_f_qf(); // expected-error {{call to undeclared function '__riscv_sf_vfnrclip_xu_f_qf'}}
 }
+
+void test_zvfbfwma_vfwmaccbf16() {
+  __riscv_vfwmaccbf16(); // expected-error {{call to undeclared function '__riscv_vfwmaccbf16'}}
+}
diff --git a/clang/utils/TableGen/RISCVVEmitter.cpp b/clang/utils/TableGen/RISCVVEmitter.cpp
index 2ca47f1ba59f3..9f6ed39f01309 100644
--- a/clang/utils/TableGen/RISCVVEmitter.cpp
+++ b/clang/utils/TableGen/RISCVVEmitter.cpp
@@ -671,6 +671,7 @@ void RVVEmitter::createRVVIntrinsics(
               .Case("Zvknhb", RVV_REQ_Zvknhb)
               .Case("Zvksed", RVV_REQ_Zvksed)
               .Case("Zvksh", RVV_REQ_Zvksh)
+              .Case("Zvfbfwma", RVV_REQ_Zvfbfwma)
               .Case("Experimental", RVV_REQ_Experimental)
               .Default(RVV_REQ_None);
       assert(RequireExt != RVV_REQ_None && "Unrecognized required feature?");



More information about the cfe-commits mailing list