[clang] 3027688 - [RISCV] Support bf16 vmv.v.v and vmerge.vvm intrinsics with `zvfbfmin` (#101611)

via cfe-commits cfe-commits at lists.llvm.org
Tue Aug 6 02:59:59 PDT 2024


Author: Brandon Wu
Date: 2024-08-06T17:59:53+08:00
New Revision: 3027688a77b5511447b3f060aaecbf30e4b9e63e

URL: https://github.com/llvm/llvm-project/commit/3027688a77b5511447b3f060aaecbf30e4b9e63e
DIFF: https://github.com/llvm/llvm-project/commit/3027688a77b5511447b3f060aaecbf30e4b9e63e.diff

LOG: [RISCV] Support bf16 vmv.v.v and vmerge.vvm intrinsics with `zvfbfmin` (#101611)

These two intrinsics are supported for f16 with `zvfhmin`, also support
them in bf16 to make it aligned to f16.

This resolve:
https://github.com/riscv-non-isa/rvv-intrinsic-doc/issues/349

Added: 
    

Modified: 
    clang/include/clang/Basic/riscv_vector.td
    clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/non-overloaded/vmerge.c
    clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/non-overloaded/vmv.c
    clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/overloaded/vmerge.c
    clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/overloaded/vmv.c
    clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/non-overloaded/vmerge.c
    clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/non-overloaded/vmv.c
    clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/overloaded/vmerge.c
    clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/overloaded/vmv.c
    llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td

Removed: 
    


################################################################################
diff  --git a/clang/include/clang/Basic/riscv_vector.td b/clang/include/clang/Basic/riscv_vector.td
index 0cab4b8067f0d..b838917e39cfe 100644
--- a/clang/include/clang/Basic/riscv_vector.td
+++ b/clang/include/clang/Basic/riscv_vector.td
@@ -1378,6 +1378,9 @@ let HasMasked = false,
     let RequiredFeatures = ["Zvfhmin"] in
       defm vmv_v : RVVOutBuiltinSet<"vmv_v_v", "x",
                                     [["v", "v", "vv"]]>;
+    let RequiredFeatures = ["Zvfbfmin"] in
+      defm vmv_v : RVVOutBuiltinSet<"vmv_v_v", "y",
+                                    [["v", "v", "vv"]]>;
   let SupportOverloading = false in
     defm vmv_v : RVVOutBuiltinSet<"vmv_v_x", "csil",
                                    [["x", "v", "ve"],
@@ -1890,6 +1893,9 @@ let HasMasked = false,
   let RequiredFeatures = ["Zvfhmin"] in
     defm vmerge : RVVOutOp1BuiltinSet<"vmerge", "x",
                                       [["vvm", "v", "vvvm"]]>;
+  let RequiredFeatures = ["Zvfbfmin"] in
+    defm vmerge : RVVOutOp1BuiltinSet<"vmerge", "y",
+                                      [["vvm", "v", "vvvm"]]>;
   defm vfmerge : RVVOutOp1BuiltinSet<"vfmerge", "xfd",
                                      [["vfm", "v", "vvem"]]>;
 }

diff  --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/non-overloaded/vmerge.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/non-overloaded/vmerge.c
index fb41a07cccec0..d2eb01c70950e 100644
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/non-overloaded/vmerge.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/non-overloaded/vmerge.c
@@ -1,7 +1,7 @@
 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 2
 // REQUIRES: riscv-registered-target
 // RUN: %clang_cc1 -triple riscv64 -target-feature +v -target-feature +zfh \
-// RUN:   -target-feature +zvfhmin -disable-O0-optnone  \
+// RUN:   -target-feature +zvfhmin -target-feature +zvfbfmin -disable-O0-optnone  \
 // RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
 // RUN:   FileCheck --check-prefix=CHECK-RV64 %s
 
@@ -1037,3 +1037,62 @@ vfloat64m8_t test_vmerge_vvm_f64m8(vfloat64m8_t op1, vfloat64m8_t op2, vbool8_t
   return __riscv_vmerge_vvm_f64m8(op1, op2, mask, vl);
 }
 
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vmerge_vvm_bf16mf4
+// CHECK-RV64-SAME: (<vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], <vscale x 1 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vmerge.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vmerge_vvm_bf16mf4(vbfloat16mf4_t op1, vbfloat16mf4_t op2, vbool64_t mask, size_t vl) {
+  return __riscv_vmerge_vvm_bf16mf4(op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vmerge_vvm_bf16mf2
+// CHECK-RV64-SAME: (<vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], <vscale x 2 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vmerge.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vmerge_vvm_bf16mf2(vbfloat16mf2_t op1, vbfloat16mf2_t op2, vbool32_t mask, size_t vl) {
+  return __riscv_vmerge_vvm_bf16mf2(op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vmerge_vvm_bf16m1
+// CHECK-RV64-SAME: (<vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], <vscale x 4 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vmerge.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vmerge_vvm_bf16m1(vbfloat16m1_t op1, vbfloat16m1_t op2, vbool16_t mask, size_t vl) {
+  return __riscv_vmerge_vvm_bf16m1(op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vmerge_vvm_bf16m2
+// CHECK-RV64-SAME: (<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], <vscale x 8 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vmerge.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vmerge_vvm_bf16m2(vbfloat16m2_t op1, vbfloat16m2_t op2, vbool8_t mask, size_t vl) {
+  return __riscv_vmerge_vvm_bf16m2(op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vmerge_vvm_bf16m4
+// CHECK-RV64-SAME: (<vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], <vscale x 16 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vmerge.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vmerge_vvm_bf16m4(vbfloat16m4_t op1, vbfloat16m4_t op2, vbool4_t mask, size_t vl) {
+  return __riscv_vmerge_vvm_bf16m4(op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vmerge_vvm_bf16m8
+// CHECK-RV64-SAME: (<vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], <vscale x 32 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vmerge.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vmerge_vvm_bf16m8(vbfloat16m8_t op1, vbfloat16m8_t op2, vbool2_t mask, size_t vl) {
+  return __riscv_vmerge_vvm_bf16m8(op1, op2, mask, vl);
+}

diff  --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/non-overloaded/vmv.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/non-overloaded/vmv.c
index c25719a80d4fb..7004c18fac0de 100644
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/non-overloaded/vmv.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/non-overloaded/vmv.c
@@ -1,7 +1,7 @@
 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 2
 // REQUIRES: riscv-registered-target
 // RUN: %clang_cc1 -triple riscv64 -target-feature +v -target-feature +zfh \
-// RUN:   -target-feature +zvfhmin -disable-O0-optnone  \
+// RUN:   -target-feature +zvfhmin -target-feature +zvfbfmin -disable-O0-optnone  \
 // RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
 // RUN:   FileCheck --check-prefix=CHECK-RV64 %s
 
@@ -1917,3 +1917,63 @@ vuint64m8_t test_vmv_s_x_u64m8(uint64_t src, size_t vl) {
   return __riscv_vmv_s_x_u64m8(src, vl);
 }
 
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vmv_v_v_bf16mf4
+// CHECK-RV64-SAME: (<vscale x 1 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vmv.v.v.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vmv_v_v_bf16mf4(vbfloat16mf4_t src, size_t vl) {
+  return __riscv_vmv_v_v_bf16mf4(src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vmv_v_v_bf16mf2
+// CHECK-RV64-SAME: (<vscale x 2 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vmv.v.v.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vmv_v_v_bf16mf2(vbfloat16mf2_t src, size_t vl) {
+  return __riscv_vmv_v_v_bf16mf2(src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vmv_v_v_bf16m1
+// CHECK-RV64-SAME: (<vscale x 4 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vmv.v.v.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vmv_v_v_bf16m1(vbfloat16m1_t src, size_t vl) {
+  return __riscv_vmv_v_v_bf16m1(src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vmv_v_v_bf16m2
+// CHECK-RV64-SAME: (<vscale x 8 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vmv.v.v.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vmv_v_v_bf16m2(vbfloat16m2_t src, size_t vl) {
+  return __riscv_vmv_v_v_bf16m2(src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vmv_v_v_bf16m4
+// CHECK-RV64-SAME: (<vscale x 16 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vmv.v.v.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vmv_v_v_bf16m4(vbfloat16m4_t src, size_t vl) {
+  return __riscv_vmv_v_v_bf16m4(src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vmv_v_v_bf16m8
+// CHECK-RV64-SAME: (<vscale x 32 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vmv.v.v.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vmv_v_v_bf16m8(vbfloat16m8_t src, size_t vl) {
+  return __riscv_vmv_v_v_bf16m8(src, vl);
+}
+

diff  --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/overloaded/vmerge.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/overloaded/vmerge.c
index d67aa70fc8c8a..2db27fdc512e5 100644
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/overloaded/vmerge.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/overloaded/vmerge.c
@@ -1,7 +1,7 @@
 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 2
 // REQUIRES: riscv-registered-target
 // RUN: %clang_cc1 -triple riscv64 -target-feature +v -target-feature +zfh \
-// RUN:   -target-feature +zvfhmin -disable-O0-optnone  \
+// RUN:   -target-feature +zvfhmin -target-feature +zvfbfmin -disable-O0-optnone  \
 // RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
 // RUN:   FileCheck --check-prefix=CHECK-RV64 %s
 
@@ -1037,3 +1037,63 @@ vfloat64m8_t test_vmerge_vvm_f64m8(vfloat64m8_t op1, vfloat64m8_t op2, vbool8_t
   return __riscv_vmerge(op1, op2, mask, vl);
 }
 
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vmerge_vvm_bf16mf4
+// CHECK-RV64-SAME: (<vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], <vscale x 1 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vmerge.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vmerge_vvm_bf16mf4(vbfloat16mf4_t op1, vbfloat16mf4_t op2, vbool64_t mask, size_t vl) {
+  return __riscv_vmerge(op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vmerge_vvm_bf16mf2
+// CHECK-RV64-SAME: (<vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], <vscale x 2 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vmerge.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vmerge_vvm_bf16mf2(vbfloat16mf2_t op1, vbfloat16mf2_t op2, vbool32_t mask, size_t vl) {
+  return __riscv_vmerge(op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vmerge_vvm_bf16m1
+// CHECK-RV64-SAME: (<vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], <vscale x 4 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vmerge.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vmerge_vvm_bf16m1(vbfloat16m1_t op1, vbfloat16m1_t op2, vbool16_t mask, size_t vl) {
+  return __riscv_vmerge(op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vmerge_vvm_bf16m2
+// CHECK-RV64-SAME: (<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], <vscale x 8 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vmerge.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vmerge_vvm_bf16m2(vbfloat16m2_t op1, vbfloat16m2_t op2, vbool8_t mask, size_t vl) {
+  return __riscv_vmerge(op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vmerge_vvm_bf16m4
+// CHECK-RV64-SAME: (<vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], <vscale x 16 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vmerge.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vmerge_vvm_bf16m4(vbfloat16m4_t op1, vbfloat16m4_t op2, vbool4_t mask, size_t vl) {
+  return __riscv_vmerge(op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vmerge_vvm_bf16m8
+// CHECK-RV64-SAME: (<vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], <vscale x 32 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vmerge.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vmerge_vvm_bf16m8(vbfloat16m8_t op1, vbfloat16m8_t op2, vbool2_t mask, size_t vl) {
+  return __riscv_vmerge(op1, op2, mask, vl);
+}
+

diff  --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/overloaded/vmv.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/overloaded/vmv.c
index 44f34c4fe8101..7b1fc33827e76 100644
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/overloaded/vmv.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/overloaded/vmv.c
@@ -1,7 +1,7 @@
 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 2
 // REQUIRES: riscv-registered-target
 // RUN: %clang_cc1 -triple riscv64 -target-feature +v -target-feature +zfh \
-// RUN:   -target-feature +zvfhmin -disable-O0-optnone  \
+// RUN:   -target-feature +zvfhmin -target-feature +zvfbfmin -disable-O0-optnone  \
 // RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
 // RUN:   FileCheck --check-prefix=CHECK-RV64 %s
 
@@ -1037,3 +1037,63 @@ uint64_t test_vmv_x_s_u64m8_u64(vuint64m8_t src) {
   return __riscv_vmv_x(src);
 }
 
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vmv_v_v_bf16mf4
+// CHECK-RV64-SAME: (<vscale x 1 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vmv.v.v.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vmv_v_v_bf16mf4(vbfloat16mf4_t src, size_t vl) {
+  return __riscv_vmv_v(src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vmv_v_v_bf16mf2
+// CHECK-RV64-SAME: (<vscale x 2 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vmv.v.v.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vmv_v_v_bf16mf2(vbfloat16mf2_t src, size_t vl) {
+  return __riscv_vmv_v(src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vmv_v_v_bf16m1
+// CHECK-RV64-SAME: (<vscale x 4 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vmv.v.v.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vmv_v_v_bf16m1(vbfloat16m1_t src, size_t vl) {
+  return __riscv_vmv_v(src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vmv_v_v_bf16m2
+// CHECK-RV64-SAME: (<vscale x 8 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vmv.v.v.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vmv_v_v_bf16m2(vbfloat16m2_t src, size_t vl) {
+  return __riscv_vmv_v(src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vmv_v_v_bf16m4
+// CHECK-RV64-SAME: (<vscale x 16 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vmv.v.v.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vmv_v_v_bf16m4(vbfloat16m4_t src, size_t vl) {
+  return __riscv_vmv_v(src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vmv_v_v_bf16m8
+// CHECK-RV64-SAME: (<vscale x 32 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vmv.v.v.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vmv_v_v_bf16m8(vbfloat16m8_t src, size_t vl) {
+  return __riscv_vmv_v(src, vl);
+}
+

diff  --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/non-overloaded/vmerge.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/non-overloaded/vmerge.c
index 4f723c579597e..7667c375937bd 100644
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/non-overloaded/vmerge.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/non-overloaded/vmerge.c
@@ -1,7 +1,7 @@
 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 2
 // REQUIRES: riscv-registered-target
 // RUN: %clang_cc1 -triple riscv64 -target-feature +v -target-feature +zfh \
-// RUN:   -target-feature +zvfhmin -disable-O0-optnone  \
+// RUN:   -target-feature +zvfhmin -target-feature +zvfbfmin -disable-O0-optnone  \
 // RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
 // RUN:   FileCheck --check-prefix=CHECK-RV64 %s
 
@@ -1037,3 +1037,63 @@ vfloat64m8_t test_vmerge_vvm_f64m8_tu(vfloat64m8_t maskedoff, vfloat64m8_t op1,
   return __riscv_vmerge_vvm_f64m8_tu(maskedoff, op1, op2, mask, vl);
 }
 
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vmerge_vvm_bf16mf4_tu
+// CHECK-RV64-SAME: (<vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], <vscale x 1 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vmerge.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vmerge_vvm_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, vbool64_t mask, size_t vl) {
+  return __riscv_vmerge_vvm_bf16mf4_tu(maskedoff, op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vmerge_vvm_bf16mf2_tu
+// CHECK-RV64-SAME: (<vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], <vscale x 2 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vmerge.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vmerge_vvm_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, vbool32_t mask, size_t vl) {
+  return __riscv_vmerge_vvm_bf16mf2_tu(maskedoff, op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vmerge_vvm_bf16m1_tu
+// CHECK-RV64-SAME: (<vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], <vscale x 4 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vmerge.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vmerge_vvm_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, vbool16_t mask, size_t vl) {
+  return __riscv_vmerge_vvm_bf16m1_tu(maskedoff, op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vmerge_vvm_bf16m2_tu
+// CHECK-RV64-SAME: (<vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], <vscale x 8 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vmerge.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vmerge_vvm_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, vbool8_t mask, size_t vl) {
+  return __riscv_vmerge_vvm_bf16m2_tu(maskedoff, op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vmerge_vvm_bf16m4_tu
+// CHECK-RV64-SAME: (<vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], <vscale x 16 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vmerge.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vmerge_vvm_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, vbool4_t mask, size_t vl) {
+  return __riscv_vmerge_vvm_bf16m4_tu(maskedoff, op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vmerge_vvm_bf16m8_tu
+// CHECK-RV64-SAME: (<vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], <vscale x 32 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vmerge.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vmerge_vvm_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, vbool2_t mask, size_t vl) {
+  return __riscv_vmerge_vvm_bf16m8_tu(maskedoff, op1, op2, mask, vl);
+}
+

diff  --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/non-overloaded/vmv.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/non-overloaded/vmv.c
index 2a5a0f4f892d3..123aace607264 100644
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/non-overloaded/vmv.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/non-overloaded/vmv.c
@@ -1,7 +1,7 @@
 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 2
 // REQUIRES: riscv-registered-target
 // RUN: %clang_cc1 -triple riscv64 -target-feature +v -target-feature +zfh \
-// RUN:   -target-feature +zvfhmin -disable-O0-optnone  \
+// RUN:   -target-feature +zvfhmin -target-feature +zvfbfmin -disable-O0-optnone  \
 // RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
 // RUN:   FileCheck --check-prefix=CHECK-RV64 %s
 
@@ -1477,3 +1477,63 @@ vuint64m8_t test_vmv_s_x_u64m8_tu(vuint64m8_t maskedoff, uint64_t src, size_t vl
   return __riscv_vmv_s_x_u64m8_tu(maskedoff, src, vl);
 }
 
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vmv_v_v_bf16mf4_tu
+// CHECK-RV64-SAME: (<vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vmv.v.v.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vmv_v_v_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t src, size_t vl) {
+  return __riscv_vmv_v_v_bf16mf4_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vmv_v_v_bf16mf2_tu
+// CHECK-RV64-SAME: (<vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vmv.v.v.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vmv_v_v_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t src, size_t vl) {
+  return __riscv_vmv_v_v_bf16mf2_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vmv_v_v_bf16m1_tu
+// CHECK-RV64-SAME: (<vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vmv.v.v.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vmv_v_v_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t src, size_t vl) {
+  return __riscv_vmv_v_v_bf16m1_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vmv_v_v_bf16m2_tu
+// CHECK-RV64-SAME: (<vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vmv.v.v.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vmv_v_v_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t src, size_t vl) {
+  return __riscv_vmv_v_v_bf16m2_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vmv_v_v_bf16m4_tu
+// CHECK-RV64-SAME: (<vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vmv.v.v.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vmv_v_v_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t src, size_t vl) {
+  return __riscv_vmv_v_v_bf16m4_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vmv_v_v_bf16m8_tu
+// CHECK-RV64-SAME: (<vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vmv.v.v.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vmv_v_v_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t src, size_t vl) {
+  return __riscv_vmv_v_v_bf16m8_tu(maskedoff, src, vl);
+}
+

diff  --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/overloaded/vmerge.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/overloaded/vmerge.c
index 8149be4cb2e71..ba1838b3fc097 100644
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/overloaded/vmerge.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/overloaded/vmerge.c
@@ -1,7 +1,7 @@
 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 2
 // REQUIRES: riscv-registered-target
 // RUN: %clang_cc1 -triple riscv64 -target-feature +v -target-feature +zfh \
-// RUN:   -target-feature +zvfhmin -disable-O0-optnone  \
+// RUN:   -target-feature +zvfhmin -target-feature +zvfbfmin -disable-O0-optnone  \
 // RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
 // RUN:   FileCheck --check-prefix=CHECK-RV64 %s
 
@@ -1037,3 +1037,63 @@ vfloat64m8_t test_vmerge_vvm_f64m8_tu(vfloat64m8_t maskedoff, vfloat64m8_t op1,
   return __riscv_vmerge_tu(maskedoff, op1, op2, mask, vl);
 }
 
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vmerge_vvm_bf16mf4_tu
+// CHECK-RV64-SAME: (<vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], <vscale x 1 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vmerge.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vmerge_vvm_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, vbool64_t mask, size_t vl) {
+  return __riscv_vmerge_tu(maskedoff, op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vmerge_vvm_bf16mf2_tu
+// CHECK-RV64-SAME: (<vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], <vscale x 2 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vmerge.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vmerge_vvm_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, vbool32_t mask, size_t vl) {
+  return __riscv_vmerge_tu(maskedoff, op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vmerge_vvm_bf16m1_tu
+// CHECK-RV64-SAME: (<vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], <vscale x 4 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vmerge.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vmerge_vvm_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, vbool16_t mask, size_t vl) {
+  return __riscv_vmerge_tu(maskedoff, op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vmerge_vvm_bf16m2_tu
+// CHECK-RV64-SAME: (<vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], <vscale x 8 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vmerge.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vmerge_vvm_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, vbool8_t mask, size_t vl) {
+  return __riscv_vmerge_tu(maskedoff, op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vmerge_vvm_bf16m4_tu
+// CHECK-RV64-SAME: (<vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], <vscale x 16 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vmerge.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vmerge_vvm_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, vbool4_t mask, size_t vl) {
+  return __riscv_vmerge_tu(maskedoff, op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vmerge_vvm_bf16m8_tu
+// CHECK-RV64-SAME: (<vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], <vscale x 32 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vmerge.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vmerge_vvm_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, vbool2_t mask, size_t vl) {
+  return __riscv_vmerge_tu(maskedoff, op1, op2, mask, vl);
+}
+

diff  --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/overloaded/vmv.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/overloaded/vmv.c
index ac95c77340202..2f15b66a43789 100644
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/overloaded/vmv.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/policy/overloaded/vmv.c
@@ -1,7 +1,7 @@
 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 2
 // REQUIRES: riscv-registered-target
 // RUN: %clang_cc1 -triple riscv64 -target-feature +v -target-feature +zfh \
-// RUN:   -target-feature +zvfhmin -disable-O0-optnone  \
+// RUN:   -target-feature +zvfhmin -target-feature +zvfbfmin -disable-O0-optnone  \
 // RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
 // RUN:   FileCheck --check-prefix=CHECK-RV64 %s
 
@@ -1477,3 +1477,63 @@ vuint64m8_t test_vmv_s_x_u64m8_tu(vuint64m8_t maskedoff, uint64_t src, size_t vl
   return __riscv_vmv_s_tu(maskedoff, src, vl);
 }
 
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vmv_v_v_bf16mf4_tu
+// CHECK-RV64-SAME: (<vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vmv.v.v.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vmv_v_v_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t src, size_t vl) {
+  return __riscv_vmv_v_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vmv_v_v_bf16mf2_tu
+// CHECK-RV64-SAME: (<vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vmv.v.v.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vmv_v_v_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t src, size_t vl) {
+  return __riscv_vmv_v_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vmv_v_v_bf16m1_tu
+// CHECK-RV64-SAME: (<vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vmv.v.v.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vmv_v_v_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t src, size_t vl) {
+  return __riscv_vmv_v_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vmv_v_v_bf16m2_tu
+// CHECK-RV64-SAME: (<vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vmv.v.v.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vmv_v_v_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t src, size_t vl) {
+  return __riscv_vmv_v_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vmv_v_v_bf16m4_tu
+// CHECK-RV64-SAME: (<vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vmv.v.v.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vmv_v_v_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t src, size_t vl) {
+  return __riscv_vmv_v_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vmv_v_v_bf16m8_tu
+// CHECK-RV64-SAME: (<vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vmv.v.v.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vmv_v_v_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t src, size_t vl) {
+  return __riscv_vmv_v_tu(maskedoff, src, vl);
+}
+

diff  --git a/llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td b/llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td
index b5dbc055e03dd..f78e404bbafd1 100644
--- a/llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td
+++ b/llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td
@@ -7245,6 +7245,14 @@ foreach vti = AllFloatVectors in {
                                  vti.RegClass, vti.ScalarRegClass>;
 }
 
+foreach vti = AllBFloatVectors in
+  let Predicates = [HasVInstructionsBF16Minimal] in
+    defm : VPatBinaryCarryInTAIL<"int_riscv_vmerge", "PseudoVMERGE", "VVM",
+                                 vti.Vector,
+                                 vti.Vector, vti.Vector, vti.Mask,
+                                 vti.Log2SEW, vti.LMul, vti.RegClass,
+                                 vti.RegClass, vti.RegClass>;
+
 foreach fvti = AllFloatVectors in {
   defvar instr = !cast<Instruction>("PseudoVMERGE_VIM_"#fvti.LMul.MX);
   let Predicates = GetVTypePredicates<fvti>.Predicates in


        


More information about the cfe-commits mailing list