[llvm] 9fe6b9e - [TargetLowering][RISCV] Fixed a scalable vector issue when lowering [s|u]mul.overflow intrinsics

Thu Nov 18 02:25:14 PST 2021

Author: Eric Tang
Date: 2021-11-18T10:16:08Z
New Revision: 9fe6b9e802543c75dbb1d1c12ade64b7734f55aa

URL: https://github.com/llvm/llvm-project/commit/9fe6b9e802543c75dbb1d1c12ade64b7734f55aa
DIFF: https://github.com/llvm/llvm-project/commit/9fe6b9e802543c75dbb1d1c12ade64b7734f55aa.diff

LOG: [TargetLowering][RISCV] Fixed a scalable vector issue when lowering [s|u]mul.overflow intrinsics

    Fixed the vector type issue that where we used getVectorNumElements()
    should be replaced by getVectorElementCount() when lowering these
    intrinsics.

    This is similar to D94149

Signed-off-by: Eric Tang <tangxingxin1008 at gmail.com>

Reviewed By: craig.topper, frasercrmck

Differential Revision: https://reviews.llvm.org/D109809

Added: 
    llvm/test/CodeGen/AArch64/sve-smulo-sdnode.ll
    llvm/test/CodeGen/AArch64/sve-umulo-sdnode.ll
    llvm/test/CodeGen/RISCV/rvv/smulo-sdnode.ll
    llvm/test/CodeGen/RISCV/rvv/umulo-sdnode.ll

Modified: 
    llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp

Removed: 
    


################################################################################
diff  --git a/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp b/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
index 1c1712c63fd61..e4a69adff05bd 100644

--- a/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
@@ -8464,8 +8464,8 @@ bool TargetLowering::expandMULO(SDNode *Node, SDValue &Result,
 
   EVT WideVT = EVT::getIntegerVT(*DAG.getContext(), VT.getScalarSizeInBits() * 2);
   if (VT.isVector())
-    WideVT = EVT::getVectorVT(*DAG.getContext(), WideVT,
-                              VT.getVectorNumElements());
+    WideVT =
+        EVT::getVectorVT(*DAG.getContext(), WideVT, VT.getVectorElementCount());
 
   SDValue BottomHalf;
   SDValue TopHalf;

diff  --git a/llvm/test/CodeGen/AArch64/sve-smulo-sdnode.ll b/llvm/test/CodeGen/AArch64/sve-smulo-sdnode.ll
new file mode 100644
index 0000000000000..60d72118426c8
--- /dev/null
+++ b/llvm/test/CodeGen/AArch64/sve-smulo-sdnode.ll
@@ -0,0 +1,494 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sve -verify-machineinstrs < %s | FileCheck %s
+
+declare { <vscale x 2 x i8>, <vscale x 2 x i1> } @llvm.smul.with.overflow.nxv2i8(<vscale x 2 x i8>, <vscale x 2 x i8>)
+
+define <vscale x 2 x i8> @smulo_nxv2i8(<vscale x 2 x i8> %x, <vscale x 2 x i8> %y) {
+; CHECK-LABEL: smulo_nxv2i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.d
+; CHECK-NEXT:    sxtb z1.d, p0/m, z1.d
+; CHECK-NEXT:    sxtb z0.d, p0/m, z0.d
+; CHECK-NEXT:    movprfx z2, z0
+; CHECK-NEXT:    smulh z2.d, p0/m, z2.d, z1.d
+; CHECK-NEXT:    mul z0.d, p0/m, z0.d, z1.d
+; CHECK-NEXT:    asr z1.d, z0.d, #63
+; CHECK-NEXT:    movprfx z3, z0
+; CHECK-NEXT:    sxtb z3.d, p0/m, z0.d
+; CHECK-NEXT:    cmpne p1.d, p0/z, z2.d, z1.d
+; CHECK-NEXT:    cmpne p2.d, p0/z, z3.d, z0.d
+; CHECK-NEXT:    orr p0.b, p0/z, p2.b, p1.b
+; CHECK-NEXT:    mov z0.d, p0/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 2 x i8>, <vscale x 2 x i1> } @llvm.smul.with.overflow.nxv2i8(<vscale x 2 x i8> %x, <vscale x 2 x i8> %y)
+  %b = extractvalue { <vscale x 2 x i8>, <vscale x 2 x i1> } %a, 0
+  %c = extractvalue { <vscale x 2 x i8>, <vscale x 2 x i1> } %a, 1
+  %d = select <vscale x 2 x i1> %c, <vscale x 2 x i8> zeroinitializer, <vscale x 2 x i8> %b
+  ret <vscale x 2 x i8> %d
+}
+
+declare { <vscale x 4 x i8>, <vscale x 4 x i1> } @llvm.smul.with.overflow.nxv4i8(<vscale x 4 x i8>, <vscale x 4 x i8>)
+
+define <vscale x 4 x i8> @smulo_nxv4i8(<vscale x 4 x i8> %x, <vscale x 4 x i8> %y) {
+; CHECK-LABEL: smulo_nxv4i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.s
+; CHECK-NEXT:    sxtb z1.s, p0/m, z1.s
+; CHECK-NEXT:    sxtb z0.s, p0/m, z0.s
+; CHECK-NEXT:    movprfx z2, z0
+; CHECK-NEXT:    smulh z2.s, p0/m, z2.s, z1.s
+; CHECK-NEXT:    mul z0.s, p0/m, z0.s, z1.s
+; CHECK-NEXT:    asr z1.s, z0.s, #31
+; CHECK-NEXT:    movprfx z3, z0
+; CHECK-NEXT:    sxtb z3.s, p0/m, z0.s
+; CHECK-NEXT:    cmpne p1.s, p0/z, z2.s, z1.s
+; CHECK-NEXT:    cmpne p2.s, p0/z, z3.s, z0.s
+; CHECK-NEXT:    orr p0.b, p0/z, p2.b, p1.b
+; CHECK-NEXT:    mov z0.s, p0/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 4 x i8>, <vscale x 4 x i1> } @llvm.smul.with.overflow.nxv4i8(<vscale x 4 x i8> %x, <vscale x 4 x i8> %y)
+  %b = extractvalue { <vscale x 4 x i8>, <vscale x 4 x i1> } %a, 0
+  %c = extractvalue { <vscale x 4 x i8>, <vscale x 4 x i1> } %a, 1
+  %d = select <vscale x 4 x i1> %c, <vscale x 4 x i8> zeroinitializer, <vscale x 4 x i8> %b
+  ret <vscale x 4 x i8> %d
+}
+
+declare { <vscale x 8 x i8>, <vscale x 8 x i1> } @llvm.smul.with.overflow.nxv8i8(<vscale x 8 x i8>, <vscale x 8 x i8>)
+
+define <vscale x 8 x i8> @smulo_nxv8i8(<vscale x 8 x i8> %x, <vscale x 8 x i8> %y) {
+; CHECK-LABEL: smulo_nxv8i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.h
+; CHECK-NEXT:    sxtb z1.h, p0/m, z1.h
+; CHECK-NEXT:    sxtb z0.h, p0/m, z0.h
+; CHECK-NEXT:    movprfx z2, z0
+; CHECK-NEXT:    smulh z2.h, p0/m, z2.h, z1.h
+; CHECK-NEXT:    mul z0.h, p0/m, z0.h, z1.h
+; CHECK-NEXT:    asr z1.h, z0.h, #15
+; CHECK-NEXT:    movprfx z3, z0
+; CHECK-NEXT:    sxtb z3.h, p0/m, z0.h
+; CHECK-NEXT:    cmpne p1.h, p0/z, z2.h, z1.h
+; CHECK-NEXT:    cmpne p2.h, p0/z, z3.h, z0.h
+; CHECK-NEXT:    orr p0.b, p0/z, p2.b, p1.b
+; CHECK-NEXT:    mov z0.h, p0/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 8 x i8>, <vscale x 8 x i1> } @llvm.smul.with.overflow.nxv8i8(<vscale x 8 x i8> %x, <vscale x 8 x i8> %y)
+  %b = extractvalue { <vscale x 8 x i8>, <vscale x 8 x i1> } %a, 0
+  %c = extractvalue { <vscale x 8 x i8>, <vscale x 8 x i1> } %a, 1
+  %d = select <vscale x 8 x i1> %c, <vscale x 8 x i8> zeroinitializer, <vscale x 8 x i8> %b
+  ret <vscale x 8 x i8> %d
+}
+
+declare { <vscale x 16 x i8>, <vscale x 16 x i1> } @llvm.smul.with.overflow.nxv16i8(<vscale x 16 x i8>, <vscale x 16 x i8>)
+
+define <vscale x 16 x i8> @smulo_nxv16i8(<vscale x 16 x i8> %x, <vscale x 16 x i8> %y) {
+; CHECK-LABEL: smulo_nxv16i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.b
+; CHECK-NEXT:    movprfx z2, z0
+; CHECK-NEXT:    smulh z2.b, p0/m, z2.b, z1.b
+; CHECK-NEXT:    mul z0.b, p0/m, z0.b, z1.b
+; CHECK-NEXT:    asr z1.b, z0.b, #7
+; CHECK-NEXT:    cmpne p0.b, p0/z, z2.b, z1.b
+; CHECK-NEXT:    mov z0.b, p0/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 16 x i8>, <vscale x 16 x i1> } @llvm.smul.with.overflow.nxv16i8(<vscale x 16 x i8> %x, <vscale x 16 x i8> %y)
+  %b = extractvalue { <vscale x 16 x i8>, <vscale x 16 x i1> } %a, 0
+  %c = extractvalue { <vscale x 16 x i8>, <vscale x 16 x i1> } %a, 1
+  %d = select <vscale x 16 x i1> %c, <vscale x 16 x i8> zeroinitializer, <vscale x 16 x i8> %b
+  ret <vscale x 16 x i8> %d
+}
+
+declare { <vscale x 32 x i8>, <vscale x 32 x i1> } @llvm.smul.with.overflow.nxv32i8(<vscale x 32 x i8>, <vscale x 32 x i8>)
+
+define <vscale x 32 x i8> @smulo_nxv32i8(<vscale x 32 x i8> %x, <vscale x 32 x i8> %y) {
+; CHECK-LABEL: smulo_nxv32i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.b
+; CHECK-NEXT:    movprfx z4, z1
+; CHECK-NEXT:    smulh z4.b, p0/m, z4.b, z3.b
+; CHECK-NEXT:    mul z1.b, p0/m, z1.b, z3.b
+; CHECK-NEXT:    movprfx z3, z0
+; CHECK-NEXT:    smulh z3.b, p0/m, z3.b, z2.b
+; CHECK-NEXT:    mul z0.b, p0/m, z0.b, z2.b
+; CHECK-NEXT:    asr z2.b, z1.b, #7
+; CHECK-NEXT:    cmpne p1.b, p0/z, z4.b, z2.b
+; CHECK-NEXT:    asr z2.b, z0.b, #7
+; CHECK-NEXT:    cmpne p0.b, p0/z, z3.b, z2.b
+; CHECK-NEXT:    mov z0.b, p0/m, #0 // =0x0
+; CHECK-NEXT:    mov z1.b, p1/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 32 x i8>, <vscale x 32 x i1> } @llvm.smul.with.overflow.nxv32i8(<vscale x 32 x i8> %x, <vscale x 32 x i8> %y)
+  %b = extractvalue { <vscale x 32 x i8>, <vscale x 32 x i1> } %a, 0
+  %c = extractvalue { <vscale x 32 x i8>, <vscale x 32 x i1> } %a, 1
+  %d = select <vscale x 32 x i1> %c, <vscale x 32 x i8> zeroinitializer, <vscale x 32 x i8> %b
+  ret <vscale x 32 x i8> %d
+}
+
+declare { <vscale x 64 x i8>, <vscale x 64 x i1> } @llvm.smul.with.overflow.nxv64i8(<vscale x 64 x i8>, <vscale x 64 x i8>)
+
+define <vscale x 64 x i8> @smulo_nxv64i8(<vscale x 64 x i8> %x, <vscale x 64 x i8> %y) {
+; CHECK-LABEL: smulo_nxv64i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.b
+; CHECK-NEXT:    movprfx z24, z3
+; CHECK-NEXT:    smulh z24.b, p0/m, z24.b, z7.b
+; CHECK-NEXT:    mul z3.b, p0/m, z3.b, z7.b
+; CHECK-NEXT:    movprfx z7, z2
+; CHECK-NEXT:    smulh z7.b, p0/m, z7.b, z6.b
+; CHECK-NEXT:    mul z2.b, p0/m, z2.b, z6.b
+; CHECK-NEXT:    movprfx z6, z1
+; CHECK-NEXT:    smulh z6.b, p0/m, z6.b, z5.b
+; CHECK-NEXT:    mul z1.b, p0/m, z1.b, z5.b
+; CHECK-NEXT:    movprfx z5, z0
+; CHECK-NEXT:    smulh z5.b, p0/m, z5.b, z4.b
+; CHECK-NEXT:    mul z0.b, p0/m, z0.b, z4.b
+; CHECK-NEXT:    asr z4.b, z3.b, #7
+; CHECK-NEXT:    cmpne p1.b, p0/z, z24.b, z4.b
+; CHECK-NEXT:    asr z4.b, z2.b, #7
+; CHECK-NEXT:    asr z24.b, z1.b, #7
+; CHECK-NEXT:    cmpne p2.b, p0/z, z7.b, z4.b
+; CHECK-NEXT:    asr z4.b, z0.b, #7
+; CHECK-NEXT:    cmpne p3.b, p0/z, z6.b, z24.b
+; CHECK-NEXT:    cmpne p0.b, p0/z, z5.b, z4.b
+; CHECK-NEXT:    mov z0.b, p0/m, #0 // =0x0
+; CHECK-NEXT:    mov z1.b, p3/m, #0 // =0x0
+; CHECK-NEXT:    mov z2.b, p2/m, #0 // =0x0
+; CHECK-NEXT:    mov z3.b, p1/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 64 x i8>, <vscale x 64 x i1> } @llvm.smul.with.overflow.nxv64i8(<vscale x 64 x i8> %x, <vscale x 64 x i8> %y)
+  %b = extractvalue { <vscale x 64 x i8>, <vscale x 64 x i1> } %a, 0
+  %c = extractvalue { <vscale x 64 x i8>, <vscale x 64 x i1> } %a, 1
+  %d = select <vscale x 64 x i1> %c, <vscale x 64 x i8> zeroinitializer, <vscale x 64 x i8> %b
+  ret <vscale x 64 x i8> %d
+}
+
+declare { <vscale x 2 x i16>, <vscale x 2 x i1> } @llvm.smul.with.overflow.nxv2i16(<vscale x 2 x i16>, <vscale x 2 x i16>)
+
+define <vscale x 2 x i16> @smulo_nxv2i16(<vscale x 2 x i16> %x, <vscale x 2 x i16> %y) {
+; CHECK-LABEL: smulo_nxv2i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.d
+; CHECK-NEXT:    sxth z1.d, p0/m, z1.d
+; CHECK-NEXT:    sxth z0.d, p0/m, z0.d
+; CHECK-NEXT:    movprfx z2, z0
+; CHECK-NEXT:    smulh z2.d, p0/m, z2.d, z1.d
+; CHECK-NEXT:    mul z0.d, p0/m, z0.d, z1.d
+; CHECK-NEXT:    asr z1.d, z0.d, #63
+; CHECK-NEXT:    movprfx z3, z0
+; CHECK-NEXT:    sxth z3.d, p0/m, z0.d
+; CHECK-NEXT:    cmpne p1.d, p0/z, z2.d, z1.d
+; CHECK-NEXT:    cmpne p2.d, p0/z, z3.d, z0.d
+; CHECK-NEXT:    orr p0.b, p0/z, p2.b, p1.b
+; CHECK-NEXT:    mov z0.d, p0/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 2 x i16>, <vscale x 2 x i1> } @llvm.smul.with.overflow.nxv2i16(<vscale x 2 x i16> %x, <vscale x 2 x i16> %y)
+  %b = extractvalue { <vscale x 2 x i16>, <vscale x 2 x i1> } %a, 0
+  %c = extractvalue { <vscale x 2 x i16>, <vscale x 2 x i1> } %a, 1
+  %d = select <vscale x 2 x i1> %c, <vscale x 2 x i16> zeroinitializer, <vscale x 2 x i16> %b
+  ret <vscale x 2 x i16> %d
+}
+
+declare { <vscale x 4 x i16>, <vscale x 4 x i1> } @llvm.smul.with.overflow.nxv4i16(<vscale x 4 x i16>, <vscale x 4 x i16>)
+
+define <vscale x 4 x i16> @smulo_nxv4i16(<vscale x 4 x i16> %x, <vscale x 4 x i16> %y) {
+; CHECK-LABEL: smulo_nxv4i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.s
+; CHECK-NEXT:    sxth z1.s, p0/m, z1.s
+; CHECK-NEXT:    sxth z0.s, p0/m, z0.s
+; CHECK-NEXT:    movprfx z2, z0
+; CHECK-NEXT:    smulh z2.s, p0/m, z2.s, z1.s
+; CHECK-NEXT:    mul z0.s, p0/m, z0.s, z1.s
+; CHECK-NEXT:    asr z1.s, z0.s, #31
+; CHECK-NEXT:    movprfx z3, z0
+; CHECK-NEXT:    sxth z3.s, p0/m, z0.s
+; CHECK-NEXT:    cmpne p1.s, p0/z, z2.s, z1.s
+; CHECK-NEXT:    cmpne p2.s, p0/z, z3.s, z0.s
+; CHECK-NEXT:    orr p0.b, p0/z, p2.b, p1.b
+; CHECK-NEXT:    mov z0.s, p0/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 4 x i16>, <vscale x 4 x i1> } @llvm.smul.with.overflow.nxv4i16(<vscale x 4 x i16> %x, <vscale x 4 x i16> %y)
+  %b = extractvalue { <vscale x 4 x i16>, <vscale x 4 x i1> } %a, 0
+  %c = extractvalue { <vscale x 4 x i16>, <vscale x 4 x i1> } %a, 1
+  %d = select <vscale x 4 x i1> %c, <vscale x 4 x i16> zeroinitializer, <vscale x 4 x i16> %b
+  ret <vscale x 4 x i16> %d
+}
+
+declare { <vscale x 8 x i16>, <vscale x 8 x i1> } @llvm.smul.with.overflow.nxv8i16(<vscale x 8 x i16>, <vscale x 8 x i16>)
+
+define <vscale x 8 x i16> @smulo_nxv8i16(<vscale x 8 x i16> %x, <vscale x 8 x i16> %y) {
+; CHECK-LABEL: smulo_nxv8i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.h
+; CHECK-NEXT:    movprfx z2, z0
+; CHECK-NEXT:    smulh z2.h, p0/m, z2.h, z1.h
+; CHECK-NEXT:    mul z0.h, p0/m, z0.h, z1.h
+; CHECK-NEXT:    asr z1.h, z0.h, #15
+; CHECK-NEXT:    cmpne p0.h, p0/z, z2.h, z1.h
+; CHECK-NEXT:    mov z0.h, p0/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 8 x i16>, <vscale x 8 x i1> } @llvm.smul.with.overflow.nxv8i16(<vscale x 8 x i16> %x, <vscale x 8 x i16> %y)
+  %b = extractvalue { <vscale x 8 x i16>, <vscale x 8 x i1> } %a, 0
+  %c = extractvalue { <vscale x 8 x i16>, <vscale x 8 x i1> } %a, 1
+  %d = select <vscale x 8 x i1> %c, <vscale x 8 x i16> zeroinitializer, <vscale x 8 x i16> %b
+  ret <vscale x 8 x i16> %d
+}
+
+declare { <vscale x 16 x i16>, <vscale x 16 x i1> } @llvm.smul.with.overflow.nxv16i16(<vscale x 16 x i16>, <vscale x 16 x i16>)
+
+define <vscale x 16 x i16> @smulo_nxv16i16(<vscale x 16 x i16> %x, <vscale x 16 x i16> %y) {
+; CHECK-LABEL: smulo_nxv16i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.h
+; CHECK-NEXT:    movprfx z4, z1
+; CHECK-NEXT:    smulh z4.h, p0/m, z4.h, z3.h
+; CHECK-NEXT:    mul z1.h, p0/m, z1.h, z3.h
+; CHECK-NEXT:    movprfx z3, z0
+; CHECK-NEXT:    smulh z3.h, p0/m, z3.h, z2.h
+; CHECK-NEXT:    mul z0.h, p0/m, z0.h, z2.h
+; CHECK-NEXT:    asr z2.h, z1.h, #15
+; CHECK-NEXT:    cmpne p1.h, p0/z, z4.h, z2.h
+; CHECK-NEXT:    asr z2.h, z0.h, #15
+; CHECK-NEXT:    cmpne p0.h, p0/z, z3.h, z2.h
+; CHECK-NEXT:    mov z0.h, p0/m, #0 // =0x0
+; CHECK-NEXT:    mov z1.h, p1/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 16 x i16>, <vscale x 16 x i1> } @llvm.smul.with.overflow.nxv16i16(<vscale x 16 x i16> %x, <vscale x 16 x i16> %y)
+  %b = extractvalue { <vscale x 16 x i16>, <vscale x 16 x i1> } %a, 0
+  %c = extractvalue { <vscale x 16 x i16>, <vscale x 16 x i1> } %a, 1
+  %d = select <vscale x 16 x i1> %c, <vscale x 16 x i16> zeroinitializer, <vscale x 16 x i16> %b
+  ret <vscale x 16 x i16> %d
+}
+
+declare { <vscale x 32 x i16>, <vscale x 32 x i1> } @llvm.smul.with.overflow.nxv32i16(<vscale x 32 x i16>, <vscale x 32 x i16>)
+
+define <vscale x 32 x i16> @smulo_nxv32i16(<vscale x 32 x i16> %x, <vscale x 32 x i16> %y) {
+; CHECK-LABEL: smulo_nxv32i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.h
+; CHECK-NEXT:    movprfx z24, z3
+; CHECK-NEXT:    smulh z24.h, p0/m, z24.h, z7.h
+; CHECK-NEXT:    mul z3.h, p0/m, z3.h, z7.h
+; CHECK-NEXT:    movprfx z7, z2
+; CHECK-NEXT:    smulh z7.h, p0/m, z7.h, z6.h
+; CHECK-NEXT:    mul z2.h, p0/m, z2.h, z6.h
+; CHECK-NEXT:    movprfx z6, z1
+; CHECK-NEXT:    smulh z6.h, p0/m, z6.h, z5.h
+; CHECK-NEXT:    mul z1.h, p0/m, z1.h, z5.h
+; CHECK-NEXT:    movprfx z5, z0
+; CHECK-NEXT:    smulh z5.h, p0/m, z5.h, z4.h
+; CHECK-NEXT:    mul z0.h, p0/m, z0.h, z4.h
+; CHECK-NEXT:    asr z4.h, z3.h, #15
+; CHECK-NEXT:    cmpne p1.h, p0/z, z24.h, z4.h
+; CHECK-NEXT:    asr z4.h, z2.h, #15
+; CHECK-NEXT:    asr z24.h, z1.h, #15
+; CHECK-NEXT:    cmpne p2.h, p0/z, z7.h, z4.h
+; CHECK-NEXT:    asr z4.h, z0.h, #15
+; CHECK-NEXT:    cmpne p3.h, p0/z, z6.h, z24.h
+; CHECK-NEXT:    cmpne p0.h, p0/z, z5.h, z4.h
+; CHECK-NEXT:    mov z0.h, p0/m, #0 // =0x0
+; CHECK-NEXT:    mov z1.h, p3/m, #0 // =0x0
+; CHECK-NEXT:    mov z2.h, p2/m, #0 // =0x0
+; CHECK-NEXT:    mov z3.h, p1/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 32 x i16>, <vscale x 32 x i1> } @llvm.smul.with.overflow.nxv32i16(<vscale x 32 x i16> %x, <vscale x 32 x i16> %y)
+  %b = extractvalue { <vscale x 32 x i16>, <vscale x 32 x i1> } %a, 0
+  %c = extractvalue { <vscale x 32 x i16>, <vscale x 32 x i1> } %a, 1
+  %d = select <vscale x 32 x i1> %c, <vscale x 32 x i16> zeroinitializer, <vscale x 32 x i16> %b
+  ret <vscale x 32 x i16> %d
+}
+
+declare { <vscale x 2 x i32>, <vscale x 2 x i1> } @llvm.smul.with.overflow.nxv2i32(<vscale x 2 x i32>, <vscale x 2 x i32>)
+
+define <vscale x 2 x i32> @smulo_nxv2i32(<vscale x 2 x i32> %x, <vscale x 2 x i32> %y) {
+; CHECK-LABEL: smulo_nxv2i32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.d
+; CHECK-NEXT:    sxtw z1.d, p0/m, z1.d
+; CHECK-NEXT:    sxtw z0.d, p0/m, z0.d
+; CHECK-NEXT:    movprfx z2, z0
+; CHECK-NEXT:    smulh z2.d, p0/m, z2.d, z1.d
+; CHECK-NEXT:    mul z0.d, p0/m, z0.d, z1.d
+; CHECK-NEXT:    asr z1.d, z0.d, #63
+; CHECK-NEXT:    movprfx z3, z0
+; CHECK-NEXT:    sxtw z3.d, p0/m, z0.d
+; CHECK-NEXT:    cmpne p1.d, p0/z, z2.d, z1.d
+; CHECK-NEXT:    cmpne p2.d, p0/z, z3.d, z0.d
+; CHECK-NEXT:    orr p0.b, p0/z, p2.b, p1.b
+; CHECK-NEXT:    mov z0.d, p0/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 2 x i32>, <vscale x 2 x i1> } @llvm.smul.with.overflow.nxv2i32(<vscale x 2 x i32> %x, <vscale x 2 x i32> %y)
+  %b = extractvalue { <vscale x 2 x i32>, <vscale x 2 x i1> } %a, 0
+  %c = extractvalue { <vscale x 2 x i32>, <vscale x 2 x i1> } %a, 1
+  %d = select <vscale x 2 x i1> %c, <vscale x 2 x i32> zeroinitializer, <vscale x 2 x i32> %b
+  ret <vscale x 2 x i32> %d
+}
+
+declare { <vscale x 4 x i32>, <vscale x 4 x i1> } @llvm.smul.with.overflow.nxv4i32(<vscale x 4 x i32>, <vscale x 4 x i32>)
+
+define <vscale x 4 x i32> @smulo_nxv4i32(<vscale x 4 x i32> %x, <vscale x 4 x i32> %y) {
+; CHECK-LABEL: smulo_nxv4i32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.s
+; CHECK-NEXT:    movprfx z2, z0
+; CHECK-NEXT:    smulh z2.s, p0/m, z2.s, z1.s
+; CHECK-NEXT:    mul z0.s, p0/m, z0.s, z1.s
+; CHECK-NEXT:    asr z1.s, z0.s, #31
+; CHECK-NEXT:    cmpne p0.s, p0/z, z2.s, z1.s
+; CHECK-NEXT:    mov z0.s, p0/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 4 x i32>, <vscale x 4 x i1> } @llvm.smul.with.overflow.nxv4i32(<vscale x 4 x i32> %x, <vscale x 4 x i32> %y)
+  %b = extractvalue { <vscale x 4 x i32>, <vscale x 4 x i1> } %a, 0
+  %c = extractvalue { <vscale x 4 x i32>, <vscale x 4 x i1> } %a, 1
+  %d = select <vscale x 4 x i1> %c, <vscale x 4 x i32> zeroinitializer, <vscale x 4 x i32> %b
+  ret <vscale x 4 x i32> %d
+}
+
+declare { <vscale x 8 x i32>, <vscale x 8 x i1> } @llvm.smul.with.overflow.nxv8i32(<vscale x 8 x i32>, <vscale x 8 x i32>)
+
+define <vscale x 8 x i32> @smulo_nxv8i32(<vscale x 8 x i32> %x, <vscale x 8 x i32> %y) {
+; CHECK-LABEL: smulo_nxv8i32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.s
+; CHECK-NEXT:    movprfx z4, z1
+; CHECK-NEXT:    smulh z4.s, p0/m, z4.s, z3.s
+; CHECK-NEXT:    mul z1.s, p0/m, z1.s, z3.s
+; CHECK-NEXT:    movprfx z3, z0
+; CHECK-NEXT:    smulh z3.s, p0/m, z3.s, z2.s
+; CHECK-NEXT:    mul z0.s, p0/m, z0.s, z2.s
+; CHECK-NEXT:    asr z2.s, z1.s, #31
+; CHECK-NEXT:    cmpne p1.s, p0/z, z4.s, z2.s
+; CHECK-NEXT:    asr z2.s, z0.s, #31
+; CHECK-NEXT:    cmpne p0.s, p0/z, z3.s, z2.s
+; CHECK-NEXT:    mov z0.s, p0/m, #0 // =0x0
+; CHECK-NEXT:    mov z1.s, p1/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 8 x i32>, <vscale x 8 x i1> } @llvm.smul.with.overflow.nxv8i32(<vscale x 8 x i32> %x, <vscale x 8 x i32> %y)
+  %b = extractvalue { <vscale x 8 x i32>, <vscale x 8 x i1> } %a, 0
+  %c = extractvalue { <vscale x 8 x i32>, <vscale x 8 x i1> } %a, 1
+  %d = select <vscale x 8 x i1> %c, <vscale x 8 x i32> zeroinitializer, <vscale x 8 x i32> %b
+  ret <vscale x 8 x i32> %d
+}
+
+declare { <vscale x 16 x i32>, <vscale x 16 x i1> } @llvm.smul.with.overflow.nxv16i32(<vscale x 16 x i32>, <vscale x 16 x i32>)
+
+define <vscale x 16 x i32> @smulo_nxv16i32(<vscale x 16 x i32> %x, <vscale x 16 x i32> %y) {
+; CHECK-LABEL: smulo_nxv16i32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.s
+; CHECK-NEXT:    movprfx z24, z3
+; CHECK-NEXT:    smulh z24.s, p0/m, z24.s, z7.s
+; CHECK-NEXT:    mul z3.s, p0/m, z3.s, z7.s
+; CHECK-NEXT:    movprfx z7, z2
+; CHECK-NEXT:    smulh z7.s, p0/m, z7.s, z6.s
+; CHECK-NEXT:    mul z2.s, p0/m, z2.s, z6.s
+; CHECK-NEXT:    movprfx z6, z1
+; CHECK-NEXT:    smulh z6.s, p0/m, z6.s, z5.s
+; CHECK-NEXT:    mul z1.s, p0/m, z1.s, z5.s
+; CHECK-NEXT:    movprfx z5, z0
+; CHECK-NEXT:    smulh z5.s, p0/m, z5.s, z4.s
+; CHECK-NEXT:    mul z0.s, p0/m, z0.s, z4.s
+; CHECK-NEXT:    asr z4.s, z3.s, #31
+; CHECK-NEXT:    cmpne p1.s, p0/z, z24.s, z4.s
+; CHECK-NEXT:    asr z4.s, z2.s, #31
+; CHECK-NEXT:    asr z24.s, z1.s, #31
+; CHECK-NEXT:    cmpne p2.s, p0/z, z7.s, z4.s
+; CHECK-NEXT:    asr z4.s, z0.s, #31
+; CHECK-NEXT:    cmpne p3.s, p0/z, z6.s, z24.s
+; CHECK-NEXT:    cmpne p0.s, p0/z, z5.s, z4.s
+; CHECK-NEXT:    mov z0.s, p0/m, #0 // =0x0
+; CHECK-NEXT:    mov z1.s, p3/m, #0 // =0x0
+; CHECK-NEXT:    mov z2.s, p2/m, #0 // =0x0
+; CHECK-NEXT:    mov z3.s, p1/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 16 x i32>, <vscale x 16 x i1> } @llvm.smul.with.overflow.nxv16i32(<vscale x 16 x i32> %x, <vscale x 16 x i32> %y)
+  %b = extractvalue { <vscale x 16 x i32>, <vscale x 16 x i1> } %a, 0
+  %c = extractvalue { <vscale x 16 x i32>, <vscale x 16 x i1> } %a, 1
+  %d = select <vscale x 16 x i1> %c, <vscale x 16 x i32> zeroinitializer, <vscale x 16 x i32> %b
+  ret <vscale x 16 x i32> %d
+}
+
+declare { <vscale x 2 x i64>, <vscale x 2 x i1> } @llvm.smul.with.overflow.nxv2i64(<vscale x 2 x i64>, <vscale x 2 x i64>)
+
+define <vscale x 2 x i64> @smulo_nxv2i64(<vscale x 2 x i64> %x, <vscale x 2 x i64> %y) {
+; CHECK-LABEL: smulo_nxv2i64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.d
+; CHECK-NEXT:    movprfx z2, z0
+; CHECK-NEXT:    smulh z2.d, p0/m, z2.d, z1.d
+; CHECK-NEXT:    mul z0.d, p0/m, z0.d, z1.d
+; CHECK-NEXT:    asr z1.d, z0.d, #63
+; CHECK-NEXT:    cmpne p0.d, p0/z, z2.d, z1.d
+; CHECK-NEXT:    mov z0.d, p0/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 2 x i64>, <vscale x 2 x i1> } @llvm.smul.with.overflow.nxv2i64(<vscale x 2 x i64> %x, <vscale x 2 x i64> %y)
+  %b = extractvalue { <vscale x 2 x i64>, <vscale x 2 x i1> } %a, 0
+  %c = extractvalue { <vscale x 2 x i64>, <vscale x 2 x i1> } %a, 1
+  %d = select <vscale x 2 x i1> %c, <vscale x 2 x i64> zeroinitializer, <vscale x 2 x i64> %b
+  ret <vscale x 2 x i64> %d
+}
+
+declare { <vscale x 4 x i64>, <vscale x 4 x i1> } @llvm.smul.with.overflow.nxv4i64(<vscale x 4 x i64>, <vscale x 4 x i64>)
+
+define <vscale x 4 x i64> @smulo_nxv4i64(<vscale x 4 x i64> %x, <vscale x 4 x i64> %y) {
+; CHECK-LABEL: smulo_nxv4i64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.d
+; CHECK-NEXT:    movprfx z4, z1
+; CHECK-NEXT:    smulh z4.d, p0/m, z4.d, z3.d
+; CHECK-NEXT:    mul z1.d, p0/m, z1.d, z3.d
+; CHECK-NEXT:    movprfx z3, z0
+; CHECK-NEXT:    smulh z3.d, p0/m, z3.d, z2.d
+; CHECK-NEXT:    mul z0.d, p0/m, z0.d, z2.d
+; CHECK-NEXT:    asr z2.d, z1.d, #63
+; CHECK-NEXT:    cmpne p1.d, p0/z, z4.d, z2.d
+; CHECK-NEXT:    asr z2.d, z0.d, #63
+; CHECK-NEXT:    cmpne p0.d, p0/z, z3.d, z2.d
+; CHECK-NEXT:    mov z0.d, p0/m, #0 // =0x0
+; CHECK-NEXT:    mov z1.d, p1/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 4 x i64>, <vscale x 4 x i1> } @llvm.smul.with.overflow.nxv4i64(<vscale x 4 x i64> %x, <vscale x 4 x i64> %y)
+  %b = extractvalue { <vscale x 4 x i64>, <vscale x 4 x i1> } %a, 0
+  %c = extractvalue { <vscale x 4 x i64>, <vscale x 4 x i1> } %a, 1
+  %d = select <vscale x 4 x i1> %c, <vscale x 4 x i64> zeroinitializer, <vscale x 4 x i64> %b
+  ret <vscale x 4 x i64> %d
+}
+
+declare { <vscale x 8 x i64>, <vscale x 8 x i1> } @llvm.smul.with.overflow.nxv8i64(<vscale x 8 x i64>, <vscale x 8 x i64>)
+
+define <vscale x 8 x i64> @smulo_nxv8i64(<vscale x 8 x i64> %x, <vscale x 8 x i64> %y) {
+; CHECK-LABEL: smulo_nxv8i64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.d
+; CHECK-NEXT:    movprfx z24, z3
+; CHECK-NEXT:    smulh z24.d, p0/m, z24.d, z7.d
+; CHECK-NEXT:    mul z3.d, p0/m, z3.d, z7.d
+; CHECK-NEXT:    movprfx z7, z2
+; CHECK-NEXT:    smulh z7.d, p0/m, z7.d, z6.d
+; CHECK-NEXT:    mul z2.d, p0/m, z2.d, z6.d
+; CHECK-NEXT:    movprfx z6, z1
+; CHECK-NEXT:    smulh z6.d, p0/m, z6.d, z5.d
+; CHECK-NEXT:    mul z1.d, p0/m, z1.d, z5.d
+; CHECK-NEXT:    movprfx z5, z0
+; CHECK-NEXT:    smulh z5.d, p0/m, z5.d, z4.d
+; CHECK-NEXT:    mul z0.d, p0/m, z0.d, z4.d
+; CHECK-NEXT:    asr z4.d, z3.d, #63
+; CHECK-NEXT:    cmpne p1.d, p0/z, z24.d, z4.d
+; CHECK-NEXT:    asr z4.d, z2.d, #63
+; CHECK-NEXT:    asr z24.d, z1.d, #63
+; CHECK-NEXT:    cmpne p2.d, p0/z, z7.d, z4.d
+; CHECK-NEXT:    asr z4.d, z0.d, #63
+; CHECK-NEXT:    cmpne p3.d, p0/z, z6.d, z24.d
+; CHECK-NEXT:    cmpne p0.d, p0/z, z5.d, z4.d
+; CHECK-NEXT:    mov z0.d, p0/m, #0 // =0x0
+; CHECK-NEXT:    mov z1.d, p3/m, #0 // =0x0
+; CHECK-NEXT:    mov z2.d, p2/m, #0 // =0x0
+; CHECK-NEXT:    mov z3.d, p1/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 8 x i64>, <vscale x 8 x i1> } @llvm.smul.with.overflow.nxv8i64(<vscale x 8 x i64> %x, <vscale x 8 x i64> %y)
+  %b = extractvalue { <vscale x 8 x i64>, <vscale x 8 x i1> } %a, 0
+  %c = extractvalue { <vscale x 8 x i64>, <vscale x 8 x i1> } %a, 1
+  %d = select <vscale x 8 x i1> %c, <vscale x 8 x i64> zeroinitializer, <vscale x 8 x i64> %b
+  ret <vscale x 8 x i64> %d
+}

diff  --git a/llvm/test/CodeGen/AArch64/sve-umulo-sdnode.ll b/llvm/test/CodeGen/AArch64/sve-umulo-sdnode.ll
new file mode 100644
index 0000000000000..44eac185a97e8
--- /dev/null
+++ b/llvm/test/CodeGen/AArch64/sve-umulo-sdnode.ll
@@ -0,0 +1,482 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sve -verify-machineinstrs < %s | FileCheck %s
+
+declare { <vscale x 2 x i8>, <vscale x 2 x i1> } @llvm.umul.with.overflow.nxv2i8(<vscale x 2 x i8>, <vscale x 2 x i8>)
+
+define <vscale x 2 x i8> @umulo_nxv2i8(<vscale x 2 x i8> %x, <vscale x 2 x i8> %y) {
+; CHECK-LABEL: umulo_nxv2i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.d
+; CHECK-NEXT:    and z1.d, z1.d, #0xff
+; CHECK-NEXT:    and z0.d, z0.d, #0xff
+; CHECK-NEXT:    movprfx z2, z0
+; CHECK-NEXT:    umulh z2.d, p0/m, z2.d, z1.d
+; CHECK-NEXT:    mul z0.d, p0/m, z0.d, z1.d
+; CHECK-NEXT:    lsr z1.d, z0.d, #8
+; CHECK-NEXT:    cmpne p1.d, p0/z, z2.d, #0
+; CHECK-NEXT:    cmpne p2.d, p0/z, z1.d, #0
+; CHECK-NEXT:    orr p0.b, p0/z, p2.b, p1.b
+; CHECK-NEXT:    mov z0.d, p0/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 2 x i8>, <vscale x 2 x i1> } @llvm.umul.with.overflow.nxv2i8(<vscale x 2 x i8> %x, <vscale x 2 x i8> %y)
+  %b = extractvalue { <vscale x 2 x i8>, <vscale x 2 x i1> } %a, 0
+  %c = extractvalue { <vscale x 2 x i8>, <vscale x 2 x i1> } %a, 1
+  %d = select <vscale x 2 x i1> %c, <vscale x 2 x i8> zeroinitializer, <vscale x 2 x i8> %b
+  ret <vscale x 2 x i8> %d
+}
+
+declare { <vscale x 4 x i8>, <vscale x 4 x i1> } @llvm.umul.with.overflow.nxv4i8(<vscale x 4 x i8>, <vscale x 4 x i8>)
+
+define <vscale x 4 x i8> @umulo_nxv4i8(<vscale x 4 x i8> %x, <vscale x 4 x i8> %y) {
+; CHECK-LABEL: umulo_nxv4i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.s
+; CHECK-NEXT:    and z1.s, z1.s, #0xff
+; CHECK-NEXT:    and z0.s, z0.s, #0xff
+; CHECK-NEXT:    movprfx z2, z0
+; CHECK-NEXT:    umulh z2.s, p0/m, z2.s, z1.s
+; CHECK-NEXT:    mul z0.s, p0/m, z0.s, z1.s
+; CHECK-NEXT:    lsr z1.s, z0.s, #8
+; CHECK-NEXT:    cmpne p1.s, p0/z, z2.s, #0
+; CHECK-NEXT:    cmpne p2.s, p0/z, z1.s, #0
+; CHECK-NEXT:    orr p0.b, p0/z, p2.b, p1.b
+; CHECK-NEXT:    mov z0.s, p0/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 4 x i8>, <vscale x 4 x i1> } @llvm.umul.with.overflow.nxv4i8(<vscale x 4 x i8> %x, <vscale x 4 x i8> %y)
+  %b = extractvalue { <vscale x 4 x i8>, <vscale x 4 x i1> } %a, 0
+  %c = extractvalue { <vscale x 4 x i8>, <vscale x 4 x i1> } %a, 1
+  %d = select <vscale x 4 x i1> %c, <vscale x 4 x i8> zeroinitializer, <vscale x 4 x i8> %b
+  ret <vscale x 4 x i8> %d
+}
+
+declare { <vscale x 8 x i8>, <vscale x 8 x i1> } @llvm.umul.with.overflow.nxv8i8(<vscale x 8 x i8>, <vscale x 8 x i8>)
+
+define <vscale x 8 x i8> @umulo_nxv8i8(<vscale x 8 x i8> %x, <vscale x 8 x i8> %y) {
+; CHECK-LABEL: umulo_nxv8i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.h
+; CHECK-NEXT:    and z1.h, z1.h, #0xff
+; CHECK-NEXT:    and z0.h, z0.h, #0xff
+; CHECK-NEXT:    movprfx z2, z0
+; CHECK-NEXT:    umulh z2.h, p0/m, z2.h, z1.h
+; CHECK-NEXT:    mul z0.h, p0/m, z0.h, z1.h
+; CHECK-NEXT:    lsr z1.h, z0.h, #8
+; CHECK-NEXT:    cmpne p1.h, p0/z, z2.h, #0
+; CHECK-NEXT:    cmpne p2.h, p0/z, z1.h, #0
+; CHECK-NEXT:    orr p0.b, p0/z, p2.b, p1.b
+; CHECK-NEXT:    mov z0.h, p0/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 8 x i8>, <vscale x 8 x i1> } @llvm.umul.with.overflow.nxv8i8(<vscale x 8 x i8> %x, <vscale x 8 x i8> %y)
+  %b = extractvalue { <vscale x 8 x i8>, <vscale x 8 x i1> } %a, 0
+  %c = extractvalue { <vscale x 8 x i8>, <vscale x 8 x i1> } %a, 1
+  %d = select <vscale x 8 x i1> %c, <vscale x 8 x i8> zeroinitializer, <vscale x 8 x i8> %b
+  ret <vscale x 8 x i8> %d
+}
+
+declare { <vscale x 16 x i8>, <vscale x 16 x i1> } @llvm.umul.with.overflow.nxv16i8(<vscale x 16 x i8>, <vscale x 16 x i8>)
+
+define <vscale x 16 x i8> @umulo_nxv16i8(<vscale x 16 x i8> %x, <vscale x 16 x i8> %y) {
+; CHECK-LABEL: umulo_nxv16i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.b
+; CHECK-NEXT:    movprfx z2, z0
+; CHECK-NEXT:    mul z2.b, p0/m, z2.b, z1.b
+; CHECK-NEXT:    umulh z0.b, p0/m, z0.b, z1.b
+; CHECK-NEXT:    cmpne p0.b, p0/z, z0.b, #0
+; CHECK-NEXT:    mov z2.b, p0/m, #0 // =0x0
+; CHECK-NEXT:    mov z0.d, z2.d
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 16 x i8>, <vscale x 16 x i1> } @llvm.umul.with.overflow.nxv16i8(<vscale x 16 x i8> %x, <vscale x 16 x i8> %y)
+  %b = extractvalue { <vscale x 16 x i8>, <vscale x 16 x i1> } %a, 0
+  %c = extractvalue { <vscale x 16 x i8>, <vscale x 16 x i1> } %a, 1
+  %d = select <vscale x 16 x i1> %c, <vscale x 16 x i8> zeroinitializer, <vscale x 16 x i8> %b
+  ret <vscale x 16 x i8> %d
+}
+
+declare { <vscale x 32 x i8>, <vscale x 32 x i1> } @llvm.umul.with.overflow.nxv32i8(<vscale x 32 x i8>, <vscale x 32 x i8>)
+
+define <vscale x 32 x i8> @umulo_nxv32i8(<vscale x 32 x i8> %x, <vscale x 32 x i8> %y) {
+; CHECK-LABEL: umulo_nxv32i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.b
+; CHECK-NEXT:    movprfx z4, z1
+; CHECK-NEXT:    mul z4.b, p0/m, z4.b, z3.b
+; CHECK-NEXT:    umulh z3.b, p0/m, z3.b, z1.b
+; CHECK-NEXT:    movprfx z1, z0
+; CHECK-NEXT:    mul z1.b, p0/m, z1.b, z2.b
+; CHECK-NEXT:    umulh z0.b, p0/m, z0.b, z2.b
+; CHECK-NEXT:    cmpne p1.b, p0/z, z3.b, #0
+; CHECK-NEXT:    cmpne p0.b, p0/z, z0.b, #0
+; CHECK-NEXT:    mov z1.b, p0/m, #0 // =0x0
+; CHECK-NEXT:    mov z4.b, p1/m, #0 // =0x0
+; CHECK-NEXT:    mov z0.d, z1.d
+; CHECK-NEXT:    mov z1.d, z4.d
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 32 x i8>, <vscale x 32 x i1> } @llvm.umul.with.overflow.nxv32i8(<vscale x 32 x i8> %x, <vscale x 32 x i8> %y)
+  %b = extractvalue { <vscale x 32 x i8>, <vscale x 32 x i1> } %a, 0
+  %c = extractvalue { <vscale x 32 x i8>, <vscale x 32 x i1> } %a, 1
+  %d = select <vscale x 32 x i1> %c, <vscale x 32 x i8> zeroinitializer, <vscale x 32 x i8> %b
+  ret <vscale x 32 x i8> %d
+}
+
+declare { <vscale x 64 x i8>, <vscale x 64 x i1> } @llvm.umul.with.overflow.nxv64i8(<vscale x 64 x i8>, <vscale x 64 x i8>)
+
+define <vscale x 64 x i8> @umulo_nxv64i8(<vscale x 64 x i8> %x, <vscale x 64 x i8> %y) {
+; CHECK-LABEL: umulo_nxv64i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.b
+; CHECK-NEXT:    movprfx z24, z3
+; CHECK-NEXT:    mul z24.b, p0/m, z24.b, z7.b
+; CHECK-NEXT:    umulh z7.b, p0/m, z7.b, z3.b
+; CHECK-NEXT:    movprfx z3, z2
+; CHECK-NEXT:    mul z3.b, p0/m, z3.b, z6.b
+; CHECK-NEXT:    umulh z6.b, p0/m, z6.b, z2.b
+; CHECK-NEXT:    movprfx z2, z1
+; CHECK-NEXT:    mul z2.b, p0/m, z2.b, z5.b
+; CHECK-NEXT:    umulh z5.b, p0/m, z5.b, z1.b
+; CHECK-NEXT:    movprfx z1, z0
+; CHECK-NEXT:    mul z1.b, p0/m, z1.b, z4.b
+; CHECK-NEXT:    umulh z0.b, p0/m, z0.b, z4.b
+; CHECK-NEXT:    cmpne p1.b, p0/z, z7.b, #0
+; CHECK-NEXT:    cmpne p2.b, p0/z, z6.b, #0
+; CHECK-NEXT:    cmpne p3.b, p0/z, z5.b, #0
+; CHECK-NEXT:    cmpne p0.b, p0/z, z0.b, #0
+; CHECK-NEXT:    mov z1.b, p0/m, #0 // =0x0
+; CHECK-NEXT:    mov z2.b, p3/m, #0 // =0x0
+; CHECK-NEXT:    mov z3.b, p2/m, #0 // =0x0
+; CHECK-NEXT:    mov z24.b, p1/m, #0 // =0x0
+; CHECK-NEXT:    mov z0.d, z1.d
+; CHECK-NEXT:    mov z1.d, z2.d
+; CHECK-NEXT:    mov z2.d, z3.d
+; CHECK-NEXT:    mov z3.d, z24.d
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 64 x i8>, <vscale x 64 x i1> } @llvm.umul.with.overflow.nxv64i8(<vscale x 64 x i8> %x, <vscale x 64 x i8> %y)
+  %b = extractvalue { <vscale x 64 x i8>, <vscale x 64 x i1> } %a, 0
+  %c = extractvalue { <vscale x 64 x i8>, <vscale x 64 x i1> } %a, 1
+  %d = select <vscale x 64 x i1> %c, <vscale x 64 x i8> zeroinitializer, <vscale x 64 x i8> %b
+  ret <vscale x 64 x i8> %d
+}
+
+declare { <vscale x 2 x i16>, <vscale x 2 x i1> } @llvm.umul.with.overflow.nxv2i16(<vscale x 2 x i16>, <vscale x 2 x i16>)
+
+define <vscale x 2 x i16> @umulo_nxv2i16(<vscale x 2 x i16> %x, <vscale x 2 x i16> %y) {
+; CHECK-LABEL: umulo_nxv2i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.d
+; CHECK-NEXT:    and z1.d, z1.d, #0xffff
+; CHECK-NEXT:    and z0.d, z0.d, #0xffff
+; CHECK-NEXT:    movprfx z2, z0
+; CHECK-NEXT:    umulh z2.d, p0/m, z2.d, z1.d
+; CHECK-NEXT:    mul z0.d, p0/m, z0.d, z1.d
+; CHECK-NEXT:    lsr z1.d, z0.d, #16
+; CHECK-NEXT:    cmpne p1.d, p0/z, z2.d, #0
+; CHECK-NEXT:    cmpne p2.d, p0/z, z1.d, #0
+; CHECK-NEXT:    orr p0.b, p0/z, p2.b, p1.b
+; CHECK-NEXT:    mov z0.d, p0/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 2 x i16>, <vscale x 2 x i1> } @llvm.umul.with.overflow.nxv2i16(<vscale x 2 x i16> %x, <vscale x 2 x i16> %y)
+  %b = extractvalue { <vscale x 2 x i16>, <vscale x 2 x i1> } %a, 0
+  %c = extractvalue { <vscale x 2 x i16>, <vscale x 2 x i1> } %a, 1
+  %d = select <vscale x 2 x i1> %c, <vscale x 2 x i16> zeroinitializer, <vscale x 2 x i16> %b
+  ret <vscale x 2 x i16> %d
+}
+
+declare { <vscale x 4 x i16>, <vscale x 4 x i1> } @llvm.umul.with.overflow.nxv4i16(<vscale x 4 x i16>, <vscale x 4 x i16>)
+
+define <vscale x 4 x i16> @umulo_nxv4i16(<vscale x 4 x i16> %x, <vscale x 4 x i16> %y) {
+; CHECK-LABEL: umulo_nxv4i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.s
+; CHECK-NEXT:    and z1.s, z1.s, #0xffff
+; CHECK-NEXT:    and z0.s, z0.s, #0xffff
+; CHECK-NEXT:    movprfx z2, z0
+; CHECK-NEXT:    umulh z2.s, p0/m, z2.s, z1.s
+; CHECK-NEXT:    mul z0.s, p0/m, z0.s, z1.s
+; CHECK-NEXT:    lsr z1.s, z0.s, #16
+; CHECK-NEXT:    cmpne p1.s, p0/z, z2.s, #0
+; CHECK-NEXT:    cmpne p2.s, p0/z, z1.s, #0
+; CHECK-NEXT:    orr p0.b, p0/z, p2.b, p1.b
+; CHECK-NEXT:    mov z0.s, p0/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 4 x i16>, <vscale x 4 x i1> } @llvm.umul.with.overflow.nxv4i16(<vscale x 4 x i16> %x, <vscale x 4 x i16> %y)
+  %b = extractvalue { <vscale x 4 x i16>, <vscale x 4 x i1> } %a, 0
+  %c = extractvalue { <vscale x 4 x i16>, <vscale x 4 x i1> } %a, 1
+  %d = select <vscale x 4 x i1> %c, <vscale x 4 x i16> zeroinitializer, <vscale x 4 x i16> %b
+  ret <vscale x 4 x i16> %d
+}
+
+declare { <vscale x 8 x i16>, <vscale x 8 x i1> } @llvm.umul.with.overflow.nxv8i16(<vscale x 8 x i16>, <vscale x 8 x i16>)
+
+define <vscale x 8 x i16> @umulo_nxv8i16(<vscale x 8 x i16> %x, <vscale x 8 x i16> %y) {
+; CHECK-LABEL: umulo_nxv8i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.h
+; CHECK-NEXT:    movprfx z2, z0
+; CHECK-NEXT:    mul z2.h, p0/m, z2.h, z1.h
+; CHECK-NEXT:    umulh z0.h, p0/m, z0.h, z1.h
+; CHECK-NEXT:    cmpne p0.h, p0/z, z0.h, #0
+; CHECK-NEXT:    mov z2.h, p0/m, #0 // =0x0
+; CHECK-NEXT:    mov z0.d, z2.d
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 8 x i16>, <vscale x 8 x i1> } @llvm.umul.with.overflow.nxv8i16(<vscale x 8 x i16> %x, <vscale x 8 x i16> %y)
+  %b = extractvalue { <vscale x 8 x i16>, <vscale x 8 x i1> } %a, 0
+  %c = extractvalue { <vscale x 8 x i16>, <vscale x 8 x i1> } %a, 1
+  %d = select <vscale x 8 x i1> %c, <vscale x 8 x i16> zeroinitializer, <vscale x 8 x i16> %b
+  ret <vscale x 8 x i16> %d
+}
+
+declare { <vscale x 16 x i16>, <vscale x 16 x i1> } @llvm.umul.with.overflow.nxv16i16(<vscale x 16 x i16>, <vscale x 16 x i16>)
+
+define <vscale x 16 x i16> @umulo_nxv16i16(<vscale x 16 x i16> %x, <vscale x 16 x i16> %y) {
+; CHECK-LABEL: umulo_nxv16i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.h
+; CHECK-NEXT:    movprfx z4, z1
+; CHECK-NEXT:    mul z4.h, p0/m, z4.h, z3.h
+; CHECK-NEXT:    umulh z3.h, p0/m, z3.h, z1.h
+; CHECK-NEXT:    movprfx z1, z0
+; CHECK-NEXT:    mul z1.h, p0/m, z1.h, z2.h
+; CHECK-NEXT:    umulh z0.h, p0/m, z0.h, z2.h
+; CHECK-NEXT:    cmpne p1.h, p0/z, z3.h, #0
+; CHECK-NEXT:    cmpne p0.h, p0/z, z0.h, #0
+; CHECK-NEXT:    mov z1.h, p0/m, #0 // =0x0
+; CHECK-NEXT:    mov z4.h, p1/m, #0 // =0x0
+; CHECK-NEXT:    mov z0.d, z1.d
+; CHECK-NEXT:    mov z1.d, z4.d
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 16 x i16>, <vscale x 16 x i1> } @llvm.umul.with.overflow.nxv16i16(<vscale x 16 x i16> %x, <vscale x 16 x i16> %y)
+  %b = extractvalue { <vscale x 16 x i16>, <vscale x 16 x i1> } %a, 0
+  %c = extractvalue { <vscale x 16 x i16>, <vscale x 16 x i1> } %a, 1
+  %d = select <vscale x 16 x i1> %c, <vscale x 16 x i16> zeroinitializer, <vscale x 16 x i16> %b
+  ret <vscale x 16 x i16> %d
+}
+
+declare { <vscale x 32 x i16>, <vscale x 32 x i1> } @llvm.umul.with.overflow.nxv32i16(<vscale x 32 x i16>, <vscale x 32 x i16>)
+
+define <vscale x 32 x i16> @umulo_nxv32i16(<vscale x 32 x i16> %x, <vscale x 32 x i16> %y) {
+; CHECK-LABEL: umulo_nxv32i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.h
+; CHECK-NEXT:    movprfx z24, z3
+; CHECK-NEXT:    mul z24.h, p0/m, z24.h, z7.h
+; CHECK-NEXT:    umulh z7.h, p0/m, z7.h, z3.h
+; CHECK-NEXT:    movprfx z3, z2
+; CHECK-NEXT:    mul z3.h, p0/m, z3.h, z6.h
+; CHECK-NEXT:    umulh z6.h, p0/m, z6.h, z2.h
+; CHECK-NEXT:    movprfx z2, z1
+; CHECK-NEXT:    mul z2.h, p0/m, z2.h, z5.h
+; CHECK-NEXT:    umulh z5.h, p0/m, z5.h, z1.h
+; CHECK-NEXT:    movprfx z1, z0
+; CHECK-NEXT:    mul z1.h, p0/m, z1.h, z4.h
+; CHECK-NEXT:    umulh z0.h, p0/m, z0.h, z4.h
+; CHECK-NEXT:    cmpne p1.h, p0/z, z7.h, #0
+; CHECK-NEXT:    cmpne p2.h, p0/z, z6.h, #0
+; CHECK-NEXT:    cmpne p3.h, p0/z, z5.h, #0
+; CHECK-NEXT:    cmpne p0.h, p0/z, z0.h, #0
+; CHECK-NEXT:    mov z1.h, p0/m, #0 // =0x0
+; CHECK-NEXT:    mov z2.h, p3/m, #0 // =0x0
+; CHECK-NEXT:    mov z3.h, p2/m, #0 // =0x0
+; CHECK-NEXT:    mov z24.h, p1/m, #0 // =0x0
+; CHECK-NEXT:    mov z0.d, z1.d
+; CHECK-NEXT:    mov z1.d, z2.d
+; CHECK-NEXT:    mov z2.d, z3.d
+; CHECK-NEXT:    mov z3.d, z24.d
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 32 x i16>, <vscale x 32 x i1> } @llvm.umul.with.overflow.nxv32i16(<vscale x 32 x i16> %x, <vscale x 32 x i16> %y)
+  %b = extractvalue { <vscale x 32 x i16>, <vscale x 32 x i1> } %a, 0
+  %c = extractvalue { <vscale x 32 x i16>, <vscale x 32 x i1> } %a, 1
+  %d = select <vscale x 32 x i1> %c, <vscale x 32 x i16> zeroinitializer, <vscale x 32 x i16> %b
+  ret <vscale x 32 x i16> %d
+}
+
+declare { <vscale x 2 x i32>, <vscale x 2 x i1> } @llvm.umul.with.overflow.nxv2i32(<vscale x 2 x i32>, <vscale x 2 x i32>)
+
+define <vscale x 2 x i32> @umulo_nxv2i32(<vscale x 2 x i32> %x, <vscale x 2 x i32> %y) {
+; CHECK-LABEL: umulo_nxv2i32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.d
+; CHECK-NEXT:    and z1.d, z1.d, #0xffffffff
+; CHECK-NEXT:    and z0.d, z0.d, #0xffffffff
+; CHECK-NEXT:    movprfx z2, z0
+; CHECK-NEXT:    umulh z2.d, p0/m, z2.d, z1.d
+; CHECK-NEXT:    mul z0.d, p0/m, z0.d, z1.d
+; CHECK-NEXT:    lsr z1.d, z0.d, #32
+; CHECK-NEXT:    cmpne p1.d, p0/z, z2.d, #0
+; CHECK-NEXT:    cmpne p2.d, p0/z, z1.d, #0
+; CHECK-NEXT:    orr p0.b, p0/z, p2.b, p1.b
+; CHECK-NEXT:    mov z0.d, p0/m, #0 // =0x0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 2 x i32>, <vscale x 2 x i1> } @llvm.umul.with.overflow.nxv2i32(<vscale x 2 x i32> %x, <vscale x 2 x i32> %y)
+  %b = extractvalue { <vscale x 2 x i32>, <vscale x 2 x i1> } %a, 0
+  %c = extractvalue { <vscale x 2 x i32>, <vscale x 2 x i1> } %a, 1
+  %d = select <vscale x 2 x i1> %c, <vscale x 2 x i32> zeroinitializer, <vscale x 2 x i32> %b
+  ret <vscale x 2 x i32> %d
+}
+
+declare { <vscale x 4 x i32>, <vscale x 4 x i1> } @llvm.umul.with.overflow.nxv4i32(<vscale x 4 x i32>, <vscale x 4 x i32>)
+
+define <vscale x 4 x i32> @umulo_nxv4i32(<vscale x 4 x i32> %x, <vscale x 4 x i32> %y) {
+; CHECK-LABEL: umulo_nxv4i32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.s
+; CHECK-NEXT:    movprfx z2, z0
+; CHECK-NEXT:    mul z2.s, p0/m, z2.s, z1.s
+; CHECK-NEXT:    umulh z0.s, p0/m, z0.s, z1.s
+; CHECK-NEXT:    cmpne p0.s, p0/z, z0.s, #0
+; CHECK-NEXT:    mov z2.s, p0/m, #0 // =0x0
+; CHECK-NEXT:    mov z0.d, z2.d
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 4 x i32>, <vscale x 4 x i1> } @llvm.umul.with.overflow.nxv4i32(<vscale x 4 x i32> %x, <vscale x 4 x i32> %y)
+  %b = extractvalue { <vscale x 4 x i32>, <vscale x 4 x i1> } %a, 0
+  %c = extractvalue { <vscale x 4 x i32>, <vscale x 4 x i1> } %a, 1
+  %d = select <vscale x 4 x i1> %c, <vscale x 4 x i32> zeroinitializer, <vscale x 4 x i32> %b
+  ret <vscale x 4 x i32> %d
+}
+
+declare { <vscale x 8 x i32>, <vscale x 8 x i1> } @llvm.umul.with.overflow.nxv8i32(<vscale x 8 x i32>, <vscale x 8 x i32>)
+
+define <vscale x 8 x i32> @umulo_nxv8i32(<vscale x 8 x i32> %x, <vscale x 8 x i32> %y) {
+; CHECK-LABEL: umulo_nxv8i32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.s
+; CHECK-NEXT:    movprfx z4, z1
+; CHECK-NEXT:    mul z4.s, p0/m, z4.s, z3.s
+; CHECK-NEXT:    umulh z3.s, p0/m, z3.s, z1.s
+; CHECK-NEXT:    movprfx z1, z0
+; CHECK-NEXT:    mul z1.s, p0/m, z1.s, z2.s
+; CHECK-NEXT:    umulh z0.s, p0/m, z0.s, z2.s
+; CHECK-NEXT:    cmpne p1.s, p0/z, z3.s, #0
+; CHECK-NEXT:    cmpne p0.s, p0/z, z0.s, #0
+; CHECK-NEXT:    mov z1.s, p0/m, #0 // =0x0
+; CHECK-NEXT:    mov z4.s, p1/m, #0 // =0x0
+; CHECK-NEXT:    mov z0.d, z1.d
+; CHECK-NEXT:    mov z1.d, z4.d
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 8 x i32>, <vscale x 8 x i1> } @llvm.umul.with.overflow.nxv8i32(<vscale x 8 x i32> %x, <vscale x 8 x i32> %y)
+  %b = extractvalue { <vscale x 8 x i32>, <vscale x 8 x i1> } %a, 0
+  %c = extractvalue { <vscale x 8 x i32>, <vscale x 8 x i1> } %a, 1
+  %d = select <vscale x 8 x i1> %c, <vscale x 8 x i32> zeroinitializer, <vscale x 8 x i32> %b
+  ret <vscale x 8 x i32> %d
+}
+
+declare { <vscale x 16 x i32>, <vscale x 16 x i1> } @llvm.umul.with.overflow.nxv16i32(<vscale x 16 x i32>, <vscale x 16 x i32>)
+
+define <vscale x 16 x i32> @umulo_nxv16i32(<vscale x 16 x i32> %x, <vscale x 16 x i32> %y) {
+; CHECK-LABEL: umulo_nxv16i32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.s
+; CHECK-NEXT:    movprfx z24, z3
+; CHECK-NEXT:    mul z24.s, p0/m, z24.s, z7.s
+; CHECK-NEXT:    umulh z7.s, p0/m, z7.s, z3.s
+; CHECK-NEXT:    movprfx z3, z2
+; CHECK-NEXT:    mul z3.s, p0/m, z3.s, z6.s
+; CHECK-NEXT:    umulh z6.s, p0/m, z6.s, z2.s
+; CHECK-NEXT:    movprfx z2, z1
+; CHECK-NEXT:    mul z2.s, p0/m, z2.s, z5.s
+; CHECK-NEXT:    umulh z5.s, p0/m, z5.s, z1.s
+; CHECK-NEXT:    movprfx z1, z0
+; CHECK-NEXT:    mul z1.s, p0/m, z1.s, z4.s
+; CHECK-NEXT:    umulh z0.s, p0/m, z0.s, z4.s
+; CHECK-NEXT:    cmpne p1.s, p0/z, z7.s, #0
+; CHECK-NEXT:    cmpne p2.s, p0/z, z6.s, #0
+; CHECK-NEXT:    cmpne p3.s, p0/z, z5.s, #0
+; CHECK-NEXT:    cmpne p0.s, p0/z, z0.s, #0
+; CHECK-NEXT:    mov z1.s, p0/m, #0 // =0x0
+; CHECK-NEXT:    mov z2.s, p3/m, #0 // =0x0
+; CHECK-NEXT:    mov z3.s, p2/m, #0 // =0x0
+; CHECK-NEXT:    mov z24.s, p1/m, #0 // =0x0
+; CHECK-NEXT:    mov z0.d, z1.d
+; CHECK-NEXT:    mov z1.d, z2.d
+; CHECK-NEXT:    mov z2.d, z3.d
+; CHECK-NEXT:    mov z3.d, z24.d
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 16 x i32>, <vscale x 16 x i1> } @llvm.umul.with.overflow.nxv16i32(<vscale x 16 x i32> %x, <vscale x 16 x i32> %y)
+  %b = extractvalue { <vscale x 16 x i32>, <vscale x 16 x i1> } %a, 0
+  %c = extractvalue { <vscale x 16 x i32>, <vscale x 16 x i1> } %a, 1
+  %d = select <vscale x 16 x i1> %c, <vscale x 16 x i32> zeroinitializer, <vscale x 16 x i32> %b
+  ret <vscale x 16 x i32> %d
+}
+
+declare { <vscale x 2 x i64>, <vscale x 2 x i1> } @llvm.umul.with.overflow.nxv2i64(<vscale x 2 x i64>, <vscale x 2 x i64>)
+
+define <vscale x 2 x i64> @umulo_nxv2i64(<vscale x 2 x i64> %x, <vscale x 2 x i64> %y) {
+; CHECK-LABEL: umulo_nxv2i64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.d
+; CHECK-NEXT:    movprfx z2, z0
+; CHECK-NEXT:    mul z2.d, p0/m, z2.d, z1.d
+; CHECK-NEXT:    umulh z0.d, p0/m, z0.d, z1.d
+; CHECK-NEXT:    cmpne p0.d, p0/z, z0.d, #0
+; CHECK-NEXT:    mov z2.d, p0/m, #0 // =0x0
+; CHECK-NEXT:    mov z0.d, z2.d
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 2 x i64>, <vscale x 2 x i1> } @llvm.umul.with.overflow.nxv2i64(<vscale x 2 x i64> %x, <vscale x 2 x i64> %y)
+  %b = extractvalue { <vscale x 2 x i64>, <vscale x 2 x i1> } %a, 0
+  %c = extractvalue { <vscale x 2 x i64>, <vscale x 2 x i1> } %a, 1
+  %d = select <vscale x 2 x i1> %c, <vscale x 2 x i64> zeroinitializer, <vscale x 2 x i64> %b
+  ret <vscale x 2 x i64> %d
+}
+
+declare { <vscale x 4 x i64>, <vscale x 4 x i1> } @llvm.umul.with.overflow.nxv4i64(<vscale x 4 x i64>, <vscale x 4 x i64>)
+
+define <vscale x 4 x i64> @umulo_nxv4i64(<vscale x 4 x i64> %x, <vscale x 4 x i64> %y) {
+; CHECK-LABEL: umulo_nxv4i64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.d
+; CHECK-NEXT:    movprfx z4, z1
+; CHECK-NEXT:    mul z4.d, p0/m, z4.d, z3.d
+; CHECK-NEXT:    umulh z3.d, p0/m, z3.d, z1.d
+; CHECK-NEXT:    movprfx z1, z0
+; CHECK-NEXT:    mul z1.d, p0/m, z1.d, z2.d
+; CHECK-NEXT:    umulh z0.d, p0/m, z0.d, z2.d
+; CHECK-NEXT:    cmpne p1.d, p0/z, z3.d, #0
+; CHECK-NEXT:    cmpne p0.d, p0/z, z0.d, #0
+; CHECK-NEXT:    mov z1.d, p0/m, #0 // =0x0
+; CHECK-NEXT:    mov z4.d, p1/m, #0 // =0x0
+; CHECK-NEXT:    mov z0.d, z1.d
+; CHECK-NEXT:    mov z1.d, z4.d
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 4 x i64>, <vscale x 4 x i1> } @llvm.umul.with.overflow.nxv4i64(<vscale x 4 x i64> %x, <vscale x 4 x i64> %y)
+  %b = extractvalue { <vscale x 4 x i64>, <vscale x 4 x i1> } %a, 0
+  %c = extractvalue { <vscale x 4 x i64>, <vscale x 4 x i1> } %a, 1
+  %d = select <vscale x 4 x i1> %c, <vscale x 4 x i64> zeroinitializer, <vscale x 4 x i64> %b
+  ret <vscale x 4 x i64> %d
+}
+
+declare { <vscale x 8 x i64>, <vscale x 8 x i1> } @llvm.umul.with.overflow.nxv8i64(<vscale x 8 x i64>, <vscale x 8 x i64>)
+
+define <vscale x 8 x i64> @umulo_nxv8i64(<vscale x 8 x i64> %x, <vscale x 8 x i64> %y) {
+; CHECK-LABEL: umulo_nxv8i64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.d
+; CHECK-NEXT:    movprfx z24, z3
+; CHECK-NEXT:    mul z24.d, p0/m, z24.d, z7.d
+; CHECK-NEXT:    umulh z7.d, p0/m, z7.d, z3.d
+; CHECK-NEXT:    movprfx z3, z2
+; CHECK-NEXT:    mul z3.d, p0/m, z3.d, z6.d
+; CHECK-NEXT:    umulh z6.d, p0/m, z6.d, z2.d
+; CHECK-NEXT:    movprfx z2, z1
+; CHECK-NEXT:    mul z2.d, p0/m, z2.d, z5.d
+; CHECK-NEXT:    umulh z5.d, p0/m, z5.d, z1.d
+; CHECK-NEXT:    movprfx z1, z0
+; CHECK-NEXT:    mul z1.d, p0/m, z1.d, z4.d
+; CHECK-NEXT:    umulh z0.d, p0/m, z0.d, z4.d
+; CHECK-NEXT:    cmpne p1.d, p0/z, z7.d, #0
+; CHECK-NEXT:    cmpne p2.d, p0/z, z6.d, #0
+; CHECK-NEXT:    cmpne p3.d, p0/z, z5.d, #0
+; CHECK-NEXT:    cmpne p0.d, p0/z, z0.d, #0
+; CHECK-NEXT:    mov z1.d, p0/m, #0 // =0x0
+; CHECK-NEXT:    mov z2.d, p3/m, #0 // =0x0
+; CHECK-NEXT:    mov z3.d, p2/m, #0 // =0x0
+; CHECK-NEXT:    mov z24.d, p1/m, #0 // =0x0
+; CHECK-NEXT:    mov z0.d, z1.d
+; CHECK-NEXT:    mov z1.d, z2.d
+; CHECK-NEXT:    mov z2.d, z3.d
+; CHECK-NEXT:    mov z3.d, z24.d
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 8 x i64>, <vscale x 8 x i1> } @llvm.umul.with.overflow.nxv8i64(<vscale x 8 x i64> %x, <vscale x 8 x i64> %y)
+  %b = extractvalue { <vscale x 8 x i64>, <vscale x 8 x i1> } %a, 0
+  %c = extractvalue { <vscale x 8 x i64>, <vscale x 8 x i1> } %a, 1
+  %d = select <vscale x 8 x i1> %c, <vscale x 8 x i64> zeroinitializer, <vscale x 8 x i64> %b
+  ret <vscale x 8 x i64> %d
+}

diff  --git a/llvm/test/CodeGen/RISCV/rvv/smulo-sdnode.ll b/llvm/test/CodeGen/RISCV/rvv/smulo-sdnode.ll
new file mode 100644
index 0000000000000..cbdd09ff3739e
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/smulo-sdnode.ll
@@ -0,0 +1,424 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=riscv64 -mattr=+experimental-v -verify-machineinstrs < %s | FileCheck %s
+
+declare { <vscale x 1 x i8>, <vscale x 1 x i1> } @llvm.smul.with.overflow.nxv1i8(<vscale x 1 x i8>, <vscale x 1 x i8>)
+
+define <vscale x 1 x i8> @smulo_nxv1i8(<vscale x 1 x i8> %x, <vscale x 1 x i8> %y) {
+; CHECK-LABEL: smulo_nxv1i8:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e8, mf8, ta, mu
+; CHECK-NEXT:    vmulh.vv v25, v8, v9
+; CHECK-NEXT:    vmul.vv v26, v8, v9
+; CHECK-NEXT:    vsra.vi v27, v26, 7
+; CHECK-NEXT:    vmsne.vv v0, v25, v27
+; CHECK-NEXT:    vmerge.vim v8, v26, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 1 x i8>, <vscale x 1 x i1> } @llvm.smul.with.overflow.nxv1i8(<vscale x 1 x i8> %x, <vscale x 1 x i8> %y)
+  %b = extractvalue { <vscale x 1 x i8>, <vscale x 1 x i1> } %a, 0
+  %c = extractvalue { <vscale x 1 x i8>, <vscale x 1 x i1> } %a, 1
+  %d = select <vscale x 1 x i1> %c, <vscale x 1 x i8> zeroinitializer, <vscale x 1 x i8> %b
+  ret <vscale x 1 x i8> %d
+}
+
+declare { <vscale x 2 x i8>, <vscale x 2 x i1> } @llvm.smul.with.overflow.nxv2i8(<vscale x 2 x i8>, <vscale x 2 x i8>)
+
+define <vscale x 2 x i8> @smulo_nxv2i8(<vscale x 2 x i8> %x, <vscale x 2 x i8> %y) {
+; CHECK-LABEL: smulo_nxv2i8:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e8, mf4, ta, mu
+; CHECK-NEXT:    vmulh.vv v25, v8, v9
+; CHECK-NEXT:    vmul.vv v26, v8, v9
+; CHECK-NEXT:    vsra.vi v27, v26, 7
+; CHECK-NEXT:    vmsne.vv v0, v25, v27
+; CHECK-NEXT:    vmerge.vim v8, v26, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 2 x i8>, <vscale x 2 x i1> } @llvm.smul.with.overflow.nxv2i8(<vscale x 2 x i8> %x, <vscale x 2 x i8> %y)
+  %b = extractvalue { <vscale x 2 x i8>, <vscale x 2 x i1> } %a, 0
+  %c = extractvalue { <vscale x 2 x i8>, <vscale x 2 x i1> } %a, 1
+  %d = select <vscale x 2 x i1> %c, <vscale x 2 x i8> zeroinitializer, <vscale x 2 x i8> %b
+  ret <vscale x 2 x i8> %d
+}
+
+declare { <vscale x 4 x i8>, <vscale x 4 x i1> } @llvm.smul.with.overflow.nxv4i8(<vscale x 4 x i8>, <vscale x 4 x i8>)
+
+define <vscale x 4 x i8> @smulo_nxv4i8(<vscale x 4 x i8> %x, <vscale x 4 x i8> %y) {
+; CHECK-LABEL: smulo_nxv4i8:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e8, mf2, ta, mu
+; CHECK-NEXT:    vmulh.vv v25, v8, v9
+; CHECK-NEXT:    vmul.vv v26, v8, v9
+; CHECK-NEXT:    vsra.vi v27, v26, 7
+; CHECK-NEXT:    vmsne.vv v0, v25, v27
+; CHECK-NEXT:    vmerge.vim v8, v26, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 4 x i8>, <vscale x 4 x i1> } @llvm.smul.with.overflow.nxv4i8(<vscale x 4 x i8> %x, <vscale x 4 x i8> %y)
+  %b = extractvalue { <vscale x 4 x i8>, <vscale x 4 x i1> } %a, 0
+  %c = extractvalue { <vscale x 4 x i8>, <vscale x 4 x i1> } %a, 1
+  %d = select <vscale x 4 x i1> %c, <vscale x 4 x i8> zeroinitializer, <vscale x 4 x i8> %b
+  ret <vscale x 4 x i8> %d
+}
+
+declare { <vscale x 8 x i8>, <vscale x 8 x i1> } @llvm.smul.with.overflow.nxv8i8(<vscale x 8 x i8>, <vscale x 8 x i8>)
+
+define <vscale x 8 x i8> @smulo_nxv8i8(<vscale x 8 x i8> %x, <vscale x 8 x i8> %y) {
+; CHECK-LABEL: smulo_nxv8i8:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e8, m1, ta, mu
+; CHECK-NEXT:    vmulh.vv v25, v8, v9
+; CHECK-NEXT:    vmul.vv v26, v8, v9
+; CHECK-NEXT:    vsra.vi v27, v26, 7
+; CHECK-NEXT:    vmsne.vv v0, v25, v27
+; CHECK-NEXT:    vmerge.vim v8, v26, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 8 x i8>, <vscale x 8 x i1> } @llvm.smul.with.overflow.nxv8i8(<vscale x 8 x i8> %x, <vscale x 8 x i8> %y)
+  %b = extractvalue { <vscale x 8 x i8>, <vscale x 8 x i1> } %a, 0
+  %c = extractvalue { <vscale x 8 x i8>, <vscale x 8 x i1> } %a, 1
+  %d = select <vscale x 8 x i1> %c, <vscale x 8 x i8> zeroinitializer, <vscale x 8 x i8> %b
+  ret <vscale x 8 x i8> %d
+}
+
+declare { <vscale x 16 x i8>, <vscale x 16 x i1> } @llvm.smul.with.overflow.nxv16i8(<vscale x 16 x i8>, <vscale x 16 x i8>)
+
+define <vscale x 16 x i8> @smulo_nxv16i8(<vscale x 16 x i8> %x, <vscale x 16 x i8> %y) {
+; CHECK-LABEL: smulo_nxv16i8:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e8, m2, ta, mu
+; CHECK-NEXT:    vmulh.vv v26, v8, v10
+; CHECK-NEXT:    vmul.vv v28, v8, v10
+; CHECK-NEXT:    vsra.vi v30, v28, 7
+; CHECK-NEXT:    vmsne.vv v0, v26, v30
+; CHECK-NEXT:    vmerge.vim v8, v28, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 16 x i8>, <vscale x 16 x i1> } @llvm.smul.with.overflow.nxv16i8(<vscale x 16 x i8> %x, <vscale x 16 x i8> %y)
+  %b = extractvalue { <vscale x 16 x i8>, <vscale x 16 x i1> } %a, 0
+  %c = extractvalue { <vscale x 16 x i8>, <vscale x 16 x i1> } %a, 1
+  %d = select <vscale x 16 x i1> %c, <vscale x 16 x i8> zeroinitializer, <vscale x 16 x i8> %b
+  ret <vscale x 16 x i8> %d
+}
+
+declare { <vscale x 32 x i8>, <vscale x 32 x i1> } @llvm.smul.with.overflow.nxv32i8(<vscale x 32 x i8>, <vscale x 32 x i8>)
+
+define <vscale x 32 x i8> @smulo_nxv32i8(<vscale x 32 x i8> %x, <vscale x 32 x i8> %y) {
+; CHECK-LABEL: smulo_nxv32i8:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e8, m4, ta, mu
+; CHECK-NEXT:    vmulh.vv v28, v8, v12
+; CHECK-NEXT:    vmul.vv v8, v8, v12
+; CHECK-NEXT:    vsra.vi v12, v8, 7
+; CHECK-NEXT:    vmsne.vv v0, v28, v12
+; CHECK-NEXT:    vmerge.vim v8, v8, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 32 x i8>, <vscale x 32 x i1> } @llvm.smul.with.overflow.nxv32i8(<vscale x 32 x i8> %x, <vscale x 32 x i8> %y)
+  %b = extractvalue { <vscale x 32 x i8>, <vscale x 32 x i1> } %a, 0
+  %c = extractvalue { <vscale x 32 x i8>, <vscale x 32 x i1> } %a, 1
+  %d = select <vscale x 32 x i1> %c, <vscale x 32 x i8> zeroinitializer, <vscale x 32 x i8> %b
+  ret <vscale x 32 x i8> %d
+}
+
+declare { <vscale x 64 x i8>, <vscale x 64 x i1> } @llvm.smul.with.overflow.nxv64i8(<vscale x 64 x i8>, <vscale x 64 x i8>)
+
+define <vscale x 64 x i8> @smulo_nxv64i8(<vscale x 64 x i8> %x, <vscale x 64 x i8> %y) {
+; CHECK-LABEL: smulo_nxv64i8:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e8, m8, ta, mu
+; CHECK-NEXT:    vmulh.vv v24, v8, v16
+; CHECK-NEXT:    vmul.vv v8, v8, v16
+; CHECK-NEXT:    vsra.vi v16, v8, 7
+; CHECK-NEXT:    vmsne.vv v0, v24, v16
+; CHECK-NEXT:    vmerge.vim v8, v8, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 64 x i8>, <vscale x 64 x i1> } @llvm.smul.with.overflow.nxv64i8(<vscale x 64 x i8> %x, <vscale x 64 x i8> %y)
+  %b = extractvalue { <vscale x 64 x i8>, <vscale x 64 x i1> } %a, 0
+  %c = extractvalue { <vscale x 64 x i8>, <vscale x 64 x i1> } %a, 1
+  %d = select <vscale x 64 x i1> %c, <vscale x 64 x i8> zeroinitializer, <vscale x 64 x i8> %b
+  ret <vscale x 64 x i8> %d
+}
+
+declare { <vscale x 1 x i16>, <vscale x 1 x i1> } @llvm.smul.with.overflow.nxv1i16(<vscale x 1 x i16>, <vscale x 1 x i16>)
+
+define <vscale x 1 x i16> @smulo_nxv1i16(<vscale x 1 x i16> %x, <vscale x 1 x i16> %y) {
+; CHECK-LABEL: smulo_nxv1i16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, mu
+; CHECK-NEXT:    vmulh.vv v25, v8, v9
+; CHECK-NEXT:    vmul.vv v26, v8, v9
+; CHECK-NEXT:    vsra.vi v27, v26, 15
+; CHECK-NEXT:    vmsne.vv v0, v25, v27
+; CHECK-NEXT:    vmerge.vim v8, v26, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 1 x i16>, <vscale x 1 x i1> } @llvm.smul.with.overflow.nxv1i16(<vscale x 1 x i16> %x, <vscale x 1 x i16> %y)
+  %b = extractvalue { <vscale x 1 x i16>, <vscale x 1 x i1> } %a, 0
+  %c = extractvalue { <vscale x 1 x i16>, <vscale x 1 x i1> } %a, 1
+  %d = select <vscale x 1 x i1> %c, <vscale x 1 x i16> zeroinitializer, <vscale x 1 x i16> %b
+  ret <vscale x 1 x i16> %d
+}
+
+declare { <vscale x 2 x i16>, <vscale x 2 x i1> } @llvm.smul.with.overflow.nxv2i16(<vscale x 2 x i16>, <vscale x 2 x i16>)
+
+define <vscale x 2 x i16> @smulo_nxv2i16(<vscale x 2 x i16> %x, <vscale x 2 x i16> %y) {
+; CHECK-LABEL: smulo_nxv2i16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, mu
+; CHECK-NEXT:    vmulh.vv v25, v8, v9
+; CHECK-NEXT:    vmul.vv v26, v8, v9
+; CHECK-NEXT:    vsra.vi v27, v26, 15
+; CHECK-NEXT:    vmsne.vv v0, v25, v27
+; CHECK-NEXT:    vmerge.vim v8, v26, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 2 x i16>, <vscale x 2 x i1> } @llvm.smul.with.overflow.nxv2i16(<vscale x 2 x i16> %x, <vscale x 2 x i16> %y)
+  %b = extractvalue { <vscale x 2 x i16>, <vscale x 2 x i1> } %a, 0
+  %c = extractvalue { <vscale x 2 x i16>, <vscale x 2 x i1> } %a, 1
+  %d = select <vscale x 2 x i1> %c, <vscale x 2 x i16> zeroinitializer, <vscale x 2 x i16> %b
+  ret <vscale x 2 x i16> %d
+}
+
+declare { <vscale x 4 x i16>, <vscale x 4 x i1> } @llvm.smul.with.overflow.nxv4i16(<vscale x 4 x i16>, <vscale x 4 x i16>)
+
+define <vscale x 4 x i16> @smulo_nxv4i16(<vscale x 4 x i16> %x, <vscale x 4 x i16> %y) {
+; CHECK-LABEL: smulo_nxv4i16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, mu
+; CHECK-NEXT:    vmulh.vv v25, v8, v9
+; CHECK-NEXT:    vmul.vv v26, v8, v9
+; CHECK-NEXT:    vsra.vi v27, v26, 15
+; CHECK-NEXT:    vmsne.vv v0, v25, v27
+; CHECK-NEXT:    vmerge.vim v8, v26, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 4 x i16>, <vscale x 4 x i1> } @llvm.smul.with.overflow.nxv4i16(<vscale x 4 x i16> %x, <vscale x 4 x i16> %y)
+  %b = extractvalue { <vscale x 4 x i16>, <vscale x 4 x i1> } %a, 0
+  %c = extractvalue { <vscale x 4 x i16>, <vscale x 4 x i1> } %a, 1
+  %d = select <vscale x 4 x i1> %c, <vscale x 4 x i16> zeroinitializer, <vscale x 4 x i16> %b
+  ret <vscale x 4 x i16> %d
+}
+
+declare { <vscale x 8 x i16>, <vscale x 8 x i1> } @llvm.smul.with.overflow.nxv8i16(<vscale x 8 x i16>, <vscale x 8 x i16>)
+
+define <vscale x 8 x i16> @smulo_nxv8i16(<vscale x 8 x i16> %x, <vscale x 8 x i16> %y) {
+; CHECK-LABEL: smulo_nxv8i16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, mu
+; CHECK-NEXT:    vmulh.vv v26, v8, v10
+; CHECK-NEXT:    vmul.vv v28, v8, v10
+; CHECK-NEXT:    vsra.vi v30, v28, 15
+; CHECK-NEXT:    vmsne.vv v0, v26, v30
+; CHECK-NEXT:    vmerge.vim v8, v28, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 8 x i16>, <vscale x 8 x i1> } @llvm.smul.with.overflow.nxv8i16(<vscale x 8 x i16> %x, <vscale x 8 x i16> %y)
+  %b = extractvalue { <vscale x 8 x i16>, <vscale x 8 x i1> } %a, 0
+  %c = extractvalue { <vscale x 8 x i16>, <vscale x 8 x i1> } %a, 1
+  %d = select <vscale x 8 x i1> %c, <vscale x 8 x i16> zeroinitializer, <vscale x 8 x i16> %b
+  ret <vscale x 8 x i16> %d
+}
+
+declare { <vscale x 16 x i16>, <vscale x 16 x i1> } @llvm.smul.with.overflow.nxv16i16(<vscale x 16 x i16>, <vscale x 16 x i16>)
+
+define <vscale x 16 x i16> @smulo_nxv16i16(<vscale x 16 x i16> %x, <vscale x 16 x i16> %y) {
+; CHECK-LABEL: smulo_nxv16i16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, mu
+; CHECK-NEXT:    vmulh.vv v28, v8, v12
+; CHECK-NEXT:    vmul.vv v8, v8, v12
+; CHECK-NEXT:    vsra.vi v12, v8, 15
+; CHECK-NEXT:    vmsne.vv v0, v28, v12
+; CHECK-NEXT:    vmerge.vim v8, v8, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 16 x i16>, <vscale x 16 x i1> } @llvm.smul.with.overflow.nxv16i16(<vscale x 16 x i16> %x, <vscale x 16 x i16> %y)
+  %b = extractvalue { <vscale x 16 x i16>, <vscale x 16 x i1> } %a, 0
+  %c = extractvalue { <vscale x 16 x i16>, <vscale x 16 x i1> } %a, 1
+  %d = select <vscale x 16 x i1> %c, <vscale x 16 x i16> zeroinitializer, <vscale x 16 x i16> %b
+  ret <vscale x 16 x i16> %d
+}
+
+declare { <vscale x 32 x i16>, <vscale x 32 x i1> } @llvm.smul.with.overflow.nxv32i16(<vscale x 32 x i16>, <vscale x 32 x i16>)
+
+define <vscale x 32 x i16> @smulo_nxv32i16(<vscale x 32 x i16> %x, <vscale x 32 x i16> %y) {
+; CHECK-LABEL: smulo_nxv32i16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e16, m8, ta, mu
+; CHECK-NEXT:    vmulh.vv v24, v8, v16
+; CHECK-NEXT:    vmul.vv v8, v8, v16
+; CHECK-NEXT:    vsra.vi v16, v8, 15
+; CHECK-NEXT:    vmsne.vv v0, v24, v16
+; CHECK-NEXT:    vmerge.vim v8, v8, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 32 x i16>, <vscale x 32 x i1> } @llvm.smul.with.overflow.nxv32i16(<vscale x 32 x i16> %x, <vscale x 32 x i16> %y)
+  %b = extractvalue { <vscale x 32 x i16>, <vscale x 32 x i1> } %a, 0
+  %c = extractvalue { <vscale x 32 x i16>, <vscale x 32 x i1> } %a, 1
+  %d = select <vscale x 32 x i1> %c, <vscale x 32 x i16> zeroinitializer, <vscale x 32 x i16> %b
+  ret <vscale x 32 x i16> %d
+}
+
+declare { <vscale x 1 x i32>, <vscale x 1 x i1> } @llvm.smul.with.overflow.nxv1i32(<vscale x 1 x i32>, <vscale x 1 x i32>)
+
+define <vscale x 1 x i32> @smulo_nxv1i32(<vscale x 1 x i32> %x, <vscale x 1 x i32> %y) {
+; CHECK-LABEL: smulo_nxv1i32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e32, mf2, ta, mu
+; CHECK-NEXT:    vmulh.vv v25, v8, v9
+; CHECK-NEXT:    vmul.vv v26, v8, v9
+; CHECK-NEXT:    vsra.vi v27, v26, 31
+; CHECK-NEXT:    vmsne.vv v0, v25, v27
+; CHECK-NEXT:    vmerge.vim v8, v26, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 1 x i32>, <vscale x 1 x i1> } @llvm.smul.with.overflow.nxv1i32(<vscale x 1 x i32> %x, <vscale x 1 x i32> %y)
+  %b = extractvalue { <vscale x 1 x i32>, <vscale x 1 x i1> } %a, 0
+  %c = extractvalue { <vscale x 1 x i32>, <vscale x 1 x i1> } %a, 1
+  %d = select <vscale x 1 x i1> %c, <vscale x 1 x i32> zeroinitializer, <vscale x 1 x i32> %b
+  ret <vscale x 1 x i32> %d
+}
+
+declare { <vscale x 2 x i32>, <vscale x 2 x i1> } @llvm.smul.with.overflow.nxv2i32(<vscale x 2 x i32>, <vscale x 2 x i32>)
+
+define <vscale x 2 x i32> @smulo_nxv2i32(<vscale x 2 x i32> %x, <vscale x 2 x i32> %y) {
+; CHECK-LABEL: smulo_nxv2i32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e32, m1, ta, mu
+; CHECK-NEXT:    vmulh.vv v25, v8, v9
+; CHECK-NEXT:    vmul.vv v26, v8, v9
+; CHECK-NEXT:    vsra.vi v27, v26, 31
+; CHECK-NEXT:    vmsne.vv v0, v25, v27
+; CHECK-NEXT:    vmerge.vim v8, v26, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 2 x i32>, <vscale x 2 x i1> } @llvm.smul.with.overflow.nxv2i32(<vscale x 2 x i32> %x, <vscale x 2 x i32> %y)
+  %b = extractvalue { <vscale x 2 x i32>, <vscale x 2 x i1> } %a, 0
+  %c = extractvalue { <vscale x 2 x i32>, <vscale x 2 x i1> } %a, 1
+  %d = select <vscale x 2 x i1> %c, <vscale x 2 x i32> zeroinitializer, <vscale x 2 x i32> %b
+  ret <vscale x 2 x i32> %d
+}
+
+declare { <vscale x 4 x i32>, <vscale x 4 x i1> } @llvm.smul.with.overflow.nxv4i32(<vscale x 4 x i32>, <vscale x 4 x i32>)
+
+define <vscale x 4 x i32> @smulo_nxv4i32(<vscale x 4 x i32> %x, <vscale x 4 x i32> %y) {
+; CHECK-LABEL: smulo_nxv4i32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e32, m2, ta, mu
+; CHECK-NEXT:    vmulh.vv v26, v8, v10
+; CHECK-NEXT:    vmul.vv v28, v8, v10
+; CHECK-NEXT:    vsra.vi v30, v28, 31
+; CHECK-NEXT:    vmsne.vv v0, v26, v30
+; CHECK-NEXT:    vmerge.vim v8, v28, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 4 x i32>, <vscale x 4 x i1> } @llvm.smul.with.overflow.nxv4i32(<vscale x 4 x i32> %x, <vscale x 4 x i32> %y)
+  %b = extractvalue { <vscale x 4 x i32>, <vscale x 4 x i1> } %a, 0
+  %c = extractvalue { <vscale x 4 x i32>, <vscale x 4 x i1> } %a, 1
+  %d = select <vscale x 4 x i1> %c, <vscale x 4 x i32> zeroinitializer, <vscale x 4 x i32> %b
+  ret <vscale x 4 x i32> %d
+}
+
+declare { <vscale x 8 x i32>, <vscale x 8 x i1> } @llvm.smul.with.overflow.nxv8i32(<vscale x 8 x i32>, <vscale x 8 x i32>)
+
+define <vscale x 8 x i32> @smulo_nxv8i32(<vscale x 8 x i32> %x, <vscale x 8 x i32> %y) {
+; CHECK-LABEL: smulo_nxv8i32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e32, m4, ta, mu
+; CHECK-NEXT:    vmulh.vv v28, v8, v12
+; CHECK-NEXT:    vmul.vv v8, v8, v12
+; CHECK-NEXT:    vsra.vi v12, v8, 31
+; CHECK-NEXT:    vmsne.vv v0, v28, v12
+; CHECK-NEXT:    vmerge.vim v8, v8, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 8 x i32>, <vscale x 8 x i1> } @llvm.smul.with.overflow.nxv8i32(<vscale x 8 x i32> %x, <vscale x 8 x i32> %y)
+  %b = extractvalue { <vscale x 8 x i32>, <vscale x 8 x i1> } %a, 0
+  %c = extractvalue { <vscale x 8 x i32>, <vscale x 8 x i1> } %a, 1
+  %d = select <vscale x 8 x i1> %c, <vscale x 8 x i32> zeroinitializer, <vscale x 8 x i32> %b
+  ret <vscale x 8 x i32> %d
+}
+
+declare { <vscale x 16 x i32>, <vscale x 16 x i1> } @llvm.smul.with.overflow.nxv16i32(<vscale x 16 x i32>, <vscale x 16 x i32>)
+
+define <vscale x 16 x i32> @smulo_nxv16i32(<vscale x 16 x i32> %x, <vscale x 16 x i32> %y) {
+; CHECK-LABEL: smulo_nxv16i32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e32, m8, ta, mu
+; CHECK-NEXT:    vmulh.vv v24, v8, v16
+; CHECK-NEXT:    vmul.vv v8, v8, v16
+; CHECK-NEXT:    vsra.vi v16, v8, 31
+; CHECK-NEXT:    vmsne.vv v0, v24, v16
+; CHECK-NEXT:    vmerge.vim v8, v8, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 16 x i32>, <vscale x 16 x i1> } @llvm.smul.with.overflow.nxv16i32(<vscale x 16 x i32> %x, <vscale x 16 x i32> %y)
+  %b = extractvalue { <vscale x 16 x i32>, <vscale x 16 x i1> } %a, 0
+  %c = extractvalue { <vscale x 16 x i32>, <vscale x 16 x i1> } %a, 1
+  %d = select <vscale x 16 x i1> %c, <vscale x 16 x i32> zeroinitializer, <vscale x 16 x i32> %b
+  ret <vscale x 16 x i32> %d
+}
+
+declare { <vscale x 1 x i64>, <vscale x 1 x i1> } @llvm.smul.with.overflow.nxv1i64(<vscale x 1 x i64>, <vscale x 1 x i64>)
+
+define <vscale x 1 x i64> @smulo_nxv1i64(<vscale x 1 x i64> %x, <vscale x 1 x i64> %y) {
+; CHECK-LABEL: smulo_nxv1i64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e64, m1, ta, mu
+; CHECK-NEXT:    vmulh.vv v25, v8, v9
+; CHECK-NEXT:    vmul.vv v26, v8, v9
+; CHECK-NEXT:    addi a0, zero, 63
+; CHECK-NEXT:    vsra.vx v27, v26, a0
+; CHECK-NEXT:    vmsne.vv v0, v25, v27
+; CHECK-NEXT:    vmerge.vim v8, v26, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 1 x i64>, <vscale x 1 x i1> } @llvm.smul.with.overflow.nxv1i64(<vscale x 1 x i64> %x, <vscale x 1 x i64> %y)
+  %b = extractvalue { <vscale x 1 x i64>, <vscale x 1 x i1> } %a, 0
+  %c = extractvalue { <vscale x 1 x i64>, <vscale x 1 x i1> } %a, 1
+  %d = select <vscale x 1 x i1> %c, <vscale x 1 x i64> zeroinitializer, <vscale x 1 x i64> %b
+  ret <vscale x 1 x i64> %d
+}
+
+declare { <vscale x 2 x i64>, <vscale x 2 x i1> } @llvm.smul.with.overflow.nxv2i64(<vscale x 2 x i64>, <vscale x 2 x i64>)
+
+define <vscale x 2 x i64> @smulo_nxv2i64(<vscale x 2 x i64> %x, <vscale x 2 x i64> %y) {
+; CHECK-LABEL: smulo_nxv2i64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e64, m2, ta, mu
+; CHECK-NEXT:    vmulh.vv v26, v8, v10
+; CHECK-NEXT:    vmul.vv v28, v8, v10
+; CHECK-NEXT:    addi a0, zero, 63
+; CHECK-NEXT:    vsra.vx v30, v28, a0
+; CHECK-NEXT:    vmsne.vv v0, v26, v30
+; CHECK-NEXT:    vmerge.vim v8, v28, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 2 x i64>, <vscale x 2 x i1> } @llvm.smul.with.overflow.nxv2i64(<vscale x 2 x i64> %x, <vscale x 2 x i64> %y)
+  %b = extractvalue { <vscale x 2 x i64>, <vscale x 2 x i1> } %a, 0
+  %c = extractvalue { <vscale x 2 x i64>, <vscale x 2 x i1> } %a, 1
+  %d = select <vscale x 2 x i1> %c, <vscale x 2 x i64> zeroinitializer, <vscale x 2 x i64> %b
+  ret <vscale x 2 x i64> %d
+}
+
+declare { <vscale x 4 x i64>, <vscale x 4 x i1> } @llvm.smul.with.overflow.nxv4i64(<vscale x 4 x i64>, <vscale x 4 x i64>)
+
+define <vscale x 4 x i64> @smulo_nxv4i64(<vscale x 4 x i64> %x, <vscale x 4 x i64> %y) {
+; CHECK-LABEL: smulo_nxv4i64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e64, m4, ta, mu
+; CHECK-NEXT:    vmulh.vv v28, v8, v12
+; CHECK-NEXT:    vmul.vv v8, v8, v12
+; CHECK-NEXT:    addi a0, zero, 63
+; CHECK-NEXT:    vsra.vx v12, v8, a0
+; CHECK-NEXT:    vmsne.vv v0, v28, v12
+; CHECK-NEXT:    vmerge.vim v8, v8, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 4 x i64>, <vscale x 4 x i1> } @llvm.smul.with.overflow.nxv4i64(<vscale x 4 x i64> %x, <vscale x 4 x i64> %y)
+  %b = extractvalue { <vscale x 4 x i64>, <vscale x 4 x i1> } %a, 0
+  %c = extractvalue { <vscale x 4 x i64>, <vscale x 4 x i1> } %a, 1
+  %d = select <vscale x 4 x i1> %c, <vscale x 4 x i64> zeroinitializer, <vscale x 4 x i64> %b
+  ret <vscale x 4 x i64> %d
+}
+
+declare { <vscale x 8 x i64>, <vscale x 8 x i1> } @llvm.smul.with.overflow.nxv8i64(<vscale x 8 x i64>, <vscale x 8 x i64>)
+
+define <vscale x 8 x i64> @smulo_nxv8i64(<vscale x 8 x i64> %x, <vscale x 8 x i64> %y) {
+; CHECK-LABEL: smulo_nxv8i64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e64, m8, ta, mu
+; CHECK-NEXT:    vmulh.vv v24, v8, v16
+; CHECK-NEXT:    vmul.vv v8, v8, v16
+; CHECK-NEXT:    addi a0, zero, 63
+; CHECK-NEXT:    vsra.vx v16, v8, a0
+; CHECK-NEXT:    vmsne.vv v0, v24, v16
+; CHECK-NEXT:    vmerge.vim v8, v8, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 8 x i64>, <vscale x 8 x i1> } @llvm.smul.with.overflow.nxv8i64(<vscale x 8 x i64> %x, <vscale x 8 x i64> %y)
+  %b = extractvalue { <vscale x 8 x i64>, <vscale x 8 x i1> } %a, 0
+  %c = extractvalue { <vscale x 8 x i64>, <vscale x 8 x i1> } %a, 1
+  %d = select <vscale x 8 x i1> %c, <vscale x 8 x i64> zeroinitializer, <vscale x 8 x i64> %b
+  ret <vscale x 8 x i64> %d
+}

diff  --git a/llvm/test/CodeGen/RISCV/rvv/umulo-sdnode.ll b/llvm/test/CodeGen/RISCV/rvv/umulo-sdnode.ll
new file mode 100644
index 0000000000000..1d100ed7d2fca
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/umulo-sdnode.ll
@@ -0,0 +1,398 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=riscv64 -mattr=+experimental-v -verify-machineinstrs < %s | FileCheck %s
+
+declare { <vscale x 1 x i8>, <vscale x 1 x i1> } @llvm.umul.with.overflow.nxv1i8(<vscale x 1 x i8>, <vscale x 1 x i8>)
+
+define <vscale x 1 x i8> @umulo_nxv1i8(<vscale x 1 x i8> %x, <vscale x 1 x i8> %y) {
+; CHECK-LABEL: umulo_nxv1i8:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e8, mf8, ta, mu
+; CHECK-NEXT:    vmulhu.vv v25, v8, v9
+; CHECK-NEXT:    vmsne.vi v0, v25, 0
+; CHECK-NEXT:    vmul.vv v25, v8, v9
+; CHECK-NEXT:    vmerge.vim v8, v25, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 1 x i8>, <vscale x 1 x i1> } @llvm.umul.with.overflow.nxv1i8(<vscale x 1 x i8> %x, <vscale x 1 x i8> %y)
+  %b = extractvalue { <vscale x 1 x i8>, <vscale x 1 x i1> } %a, 0
+  %c = extractvalue { <vscale x 1 x i8>, <vscale x 1 x i1> } %a, 1
+  %d = select <vscale x 1 x i1> %c, <vscale x 1 x i8> zeroinitializer, <vscale x 1 x i8> %b
+  ret <vscale x 1 x i8> %d
+}
+
+declare { <vscale x 2 x i8>, <vscale x 2 x i1> } @llvm.umul.with.overflow.nxv2i8(<vscale x 2 x i8>, <vscale x 2 x i8>)
+
+define <vscale x 2 x i8> @umulo_nxv2i8(<vscale x 2 x i8> %x, <vscale x 2 x i8> %y) {
+; CHECK-LABEL: umulo_nxv2i8:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e8, mf4, ta, mu
+; CHECK-NEXT:    vmulhu.vv v25, v8, v9
+; CHECK-NEXT:    vmsne.vi v0, v25, 0
+; CHECK-NEXT:    vmul.vv v25, v8, v9
+; CHECK-NEXT:    vmerge.vim v8, v25, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 2 x i8>, <vscale x 2 x i1> } @llvm.umul.with.overflow.nxv2i8(<vscale x 2 x i8> %x, <vscale x 2 x i8> %y)
+  %b = extractvalue { <vscale x 2 x i8>, <vscale x 2 x i1> } %a, 0
+  %c = extractvalue { <vscale x 2 x i8>, <vscale x 2 x i1> } %a, 1
+  %d = select <vscale x 2 x i1> %c, <vscale x 2 x i8> zeroinitializer, <vscale x 2 x i8> %b
+  ret <vscale x 2 x i8> %d
+}
+
+declare { <vscale x 4 x i8>, <vscale x 4 x i1> } @llvm.umul.with.overflow.nxv4i8(<vscale x 4 x i8>, <vscale x 4 x i8>)
+
+define <vscale x 4 x i8> @umulo_nxv4i8(<vscale x 4 x i8> %x, <vscale x 4 x i8> %y) {
+; CHECK-LABEL: umulo_nxv4i8:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e8, mf2, ta, mu
+; CHECK-NEXT:    vmulhu.vv v25, v8, v9
+; CHECK-NEXT:    vmsne.vi v0, v25, 0
+; CHECK-NEXT:    vmul.vv v25, v8, v9
+; CHECK-NEXT:    vmerge.vim v8, v25, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 4 x i8>, <vscale x 4 x i1> } @llvm.umul.with.overflow.nxv4i8(<vscale x 4 x i8> %x, <vscale x 4 x i8> %y)
+  %b = extractvalue { <vscale x 4 x i8>, <vscale x 4 x i1> } %a, 0
+  %c = extractvalue { <vscale x 4 x i8>, <vscale x 4 x i1> } %a, 1
+  %d = select <vscale x 4 x i1> %c, <vscale x 4 x i8> zeroinitializer, <vscale x 4 x i8> %b
+  ret <vscale x 4 x i8> %d
+}
+
+declare { <vscale x 8 x i8>, <vscale x 8 x i1> } @llvm.umul.with.overflow.nxv8i8(<vscale x 8 x i8>, <vscale x 8 x i8>)
+
+define <vscale x 8 x i8> @umulo_nxv8i8(<vscale x 8 x i8> %x, <vscale x 8 x i8> %y) {
+; CHECK-LABEL: umulo_nxv8i8:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e8, m1, ta, mu
+; CHECK-NEXT:    vmulhu.vv v25, v8, v9
+; CHECK-NEXT:    vmsne.vi v0, v25, 0
+; CHECK-NEXT:    vmul.vv v25, v8, v9
+; CHECK-NEXT:    vmerge.vim v8, v25, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 8 x i8>, <vscale x 8 x i1> } @llvm.umul.with.overflow.nxv8i8(<vscale x 8 x i8> %x, <vscale x 8 x i8> %y)
+  %b = extractvalue { <vscale x 8 x i8>, <vscale x 8 x i1> } %a, 0
+  %c = extractvalue { <vscale x 8 x i8>, <vscale x 8 x i1> } %a, 1
+  %d = select <vscale x 8 x i1> %c, <vscale x 8 x i8> zeroinitializer, <vscale x 8 x i8> %b
+  ret <vscale x 8 x i8> %d
+}
+
+declare { <vscale x 16 x i8>, <vscale x 16 x i1> } @llvm.umul.with.overflow.nxv16i8(<vscale x 16 x i8>, <vscale x 16 x i8>)
+
+define <vscale x 16 x i8> @umulo_nxv16i8(<vscale x 16 x i8> %x, <vscale x 16 x i8> %y) {
+; CHECK-LABEL: umulo_nxv16i8:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e8, m2, ta, mu
+; CHECK-NEXT:    vmulhu.vv v26, v8, v10
+; CHECK-NEXT:    vmsne.vi v0, v26, 0
+; CHECK-NEXT:    vmul.vv v26, v8, v10
+; CHECK-NEXT:    vmerge.vim v8, v26, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 16 x i8>, <vscale x 16 x i1> } @llvm.umul.with.overflow.nxv16i8(<vscale x 16 x i8> %x, <vscale x 16 x i8> %y)
+  %b = extractvalue { <vscale x 16 x i8>, <vscale x 16 x i1> } %a, 0
+  %c = extractvalue { <vscale x 16 x i8>, <vscale x 16 x i1> } %a, 1
+  %d = select <vscale x 16 x i1> %c, <vscale x 16 x i8> zeroinitializer, <vscale x 16 x i8> %b
+  ret <vscale x 16 x i8> %d
+}
+
+declare { <vscale x 32 x i8>, <vscale x 32 x i1> } @llvm.umul.with.overflow.nxv32i8(<vscale x 32 x i8>, <vscale x 32 x i8>)
+
+define <vscale x 32 x i8> @umulo_nxv32i8(<vscale x 32 x i8> %x, <vscale x 32 x i8> %y) {
+; CHECK-LABEL: umulo_nxv32i8:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e8, m4, ta, mu
+; CHECK-NEXT:    vmulhu.vv v28, v8, v12
+; CHECK-NEXT:    vmsne.vi v0, v28, 0
+; CHECK-NEXT:    vmul.vv v28, v8, v12
+; CHECK-NEXT:    vmerge.vim v8, v28, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 32 x i8>, <vscale x 32 x i1> } @llvm.umul.with.overflow.nxv32i8(<vscale x 32 x i8> %x, <vscale x 32 x i8> %y)
+  %b = extractvalue { <vscale x 32 x i8>, <vscale x 32 x i1> } %a, 0
+  %c = extractvalue { <vscale x 32 x i8>, <vscale x 32 x i1> } %a, 1
+  %d = select <vscale x 32 x i1> %c, <vscale x 32 x i8> zeroinitializer, <vscale x 32 x i8> %b
+  ret <vscale x 32 x i8> %d
+}
+
+declare { <vscale x 64 x i8>, <vscale x 64 x i1> } @llvm.umul.with.overflow.nxv64i8(<vscale x 64 x i8>, <vscale x 64 x i8>)
+
+define <vscale x 64 x i8> @umulo_nxv64i8(<vscale x 64 x i8> %x, <vscale x 64 x i8> %y) {
+; CHECK-LABEL: umulo_nxv64i8:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e8, m8, ta, mu
+; CHECK-NEXT:    vmulhu.vv v24, v8, v16
+; CHECK-NEXT:    vmsne.vi v0, v24, 0
+; CHECK-NEXT:    vmul.vv v8, v8, v16
+; CHECK-NEXT:    vmerge.vim v8, v8, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 64 x i8>, <vscale x 64 x i1> } @llvm.umul.with.overflow.nxv64i8(<vscale x 64 x i8> %x, <vscale x 64 x i8> %y)
+  %b = extractvalue { <vscale x 64 x i8>, <vscale x 64 x i1> } %a, 0
+  %c = extractvalue { <vscale x 64 x i8>, <vscale x 64 x i1> } %a, 1
+  %d = select <vscale x 64 x i1> %c, <vscale x 64 x i8> zeroinitializer, <vscale x 64 x i8> %b
+  ret <vscale x 64 x i8> %d
+}
+
+declare { <vscale x 1 x i16>, <vscale x 1 x i1> } @llvm.umul.with.overflow.nxv1i16(<vscale x 1 x i16>, <vscale x 1 x i16>)
+
+define <vscale x 1 x i16> @umulo_nxv1i16(<vscale x 1 x i16> %x, <vscale x 1 x i16> %y) {
+; CHECK-LABEL: umulo_nxv1i16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, mu
+; CHECK-NEXT:    vmulhu.vv v25, v8, v9
+; CHECK-NEXT:    vmsne.vi v0, v25, 0
+; CHECK-NEXT:    vmul.vv v25, v8, v9
+; CHECK-NEXT:    vmerge.vim v8, v25, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 1 x i16>, <vscale x 1 x i1> } @llvm.umul.with.overflow.nxv1i16(<vscale x 1 x i16> %x, <vscale x 1 x i16> %y)
+  %b = extractvalue { <vscale x 1 x i16>, <vscale x 1 x i1> } %a, 0
+  %c = extractvalue { <vscale x 1 x i16>, <vscale x 1 x i1> } %a, 1
+  %d = select <vscale x 1 x i1> %c, <vscale x 1 x i16> zeroinitializer, <vscale x 1 x i16> %b
+  ret <vscale x 1 x i16> %d
+}
+
+declare { <vscale x 2 x i16>, <vscale x 2 x i1> } @llvm.umul.with.overflow.nxv2i16(<vscale x 2 x i16>, <vscale x 2 x i16>)
+
+define <vscale x 2 x i16> @umulo_nxv2i16(<vscale x 2 x i16> %x, <vscale x 2 x i16> %y) {
+; CHECK-LABEL: umulo_nxv2i16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, mu
+; CHECK-NEXT:    vmulhu.vv v25, v8, v9
+; CHECK-NEXT:    vmsne.vi v0, v25, 0
+; CHECK-NEXT:    vmul.vv v25, v8, v9
+; CHECK-NEXT:    vmerge.vim v8, v25, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 2 x i16>, <vscale x 2 x i1> } @llvm.umul.with.overflow.nxv2i16(<vscale x 2 x i16> %x, <vscale x 2 x i16> %y)
+  %b = extractvalue { <vscale x 2 x i16>, <vscale x 2 x i1> } %a, 0
+  %c = extractvalue { <vscale x 2 x i16>, <vscale x 2 x i1> } %a, 1
+  %d = select <vscale x 2 x i1> %c, <vscale x 2 x i16> zeroinitializer, <vscale x 2 x i16> %b
+  ret <vscale x 2 x i16> %d
+}
+
+declare { <vscale x 4 x i16>, <vscale x 4 x i1> } @llvm.umul.with.overflow.nxv4i16(<vscale x 4 x i16>, <vscale x 4 x i16>)
+
+define <vscale x 4 x i16> @umulo_nxv4i16(<vscale x 4 x i16> %x, <vscale x 4 x i16> %y) {
+; CHECK-LABEL: umulo_nxv4i16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, mu
+; CHECK-NEXT:    vmulhu.vv v25, v8, v9
+; CHECK-NEXT:    vmsne.vi v0, v25, 0
+; CHECK-NEXT:    vmul.vv v25, v8, v9
+; CHECK-NEXT:    vmerge.vim v8, v25, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 4 x i16>, <vscale x 4 x i1> } @llvm.umul.with.overflow.nxv4i16(<vscale x 4 x i16> %x, <vscale x 4 x i16> %y)
+  %b = extractvalue { <vscale x 4 x i16>, <vscale x 4 x i1> } %a, 0
+  %c = extractvalue { <vscale x 4 x i16>, <vscale x 4 x i1> } %a, 1
+  %d = select <vscale x 4 x i1> %c, <vscale x 4 x i16> zeroinitializer, <vscale x 4 x i16> %b
+  ret <vscale x 4 x i16> %d
+}
+
+declare { <vscale x 8 x i16>, <vscale x 8 x i1> } @llvm.umul.with.overflow.nxv8i16(<vscale x 8 x i16>, <vscale x 8 x i16>)
+
+define <vscale x 8 x i16> @umulo_nxv8i16(<vscale x 8 x i16> %x, <vscale x 8 x i16> %y) {
+; CHECK-LABEL: umulo_nxv8i16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, mu
+; CHECK-NEXT:    vmulhu.vv v26, v8, v10
+; CHECK-NEXT:    vmsne.vi v0, v26, 0
+; CHECK-NEXT:    vmul.vv v26, v8, v10
+; CHECK-NEXT:    vmerge.vim v8, v26, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 8 x i16>, <vscale x 8 x i1> } @llvm.umul.with.overflow.nxv8i16(<vscale x 8 x i16> %x, <vscale x 8 x i16> %y)
+  %b = extractvalue { <vscale x 8 x i16>, <vscale x 8 x i1> } %a, 0
+  %c = extractvalue { <vscale x 8 x i16>, <vscale x 8 x i1> } %a, 1
+  %d = select <vscale x 8 x i1> %c, <vscale x 8 x i16> zeroinitializer, <vscale x 8 x i16> %b
+  ret <vscale x 8 x i16> %d
+}
+
+declare { <vscale x 16 x i16>, <vscale x 16 x i1> } @llvm.umul.with.overflow.nxv16i16(<vscale x 16 x i16>, <vscale x 16 x i16>)
+
+define <vscale x 16 x i16> @umulo_nxv16i16(<vscale x 16 x i16> %x, <vscale x 16 x i16> %y) {
+; CHECK-LABEL: umulo_nxv16i16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, mu
+; CHECK-NEXT:    vmulhu.vv v28, v8, v12
+; CHECK-NEXT:    vmsne.vi v0, v28, 0
+; CHECK-NEXT:    vmul.vv v28, v8, v12
+; CHECK-NEXT:    vmerge.vim v8, v28, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 16 x i16>, <vscale x 16 x i1> } @llvm.umul.with.overflow.nxv16i16(<vscale x 16 x i16> %x, <vscale x 16 x i16> %y)
+  %b = extractvalue { <vscale x 16 x i16>, <vscale x 16 x i1> } %a, 0
+  %c = extractvalue { <vscale x 16 x i16>, <vscale x 16 x i1> } %a, 1
+  %d = select <vscale x 16 x i1> %c, <vscale x 16 x i16> zeroinitializer, <vscale x 16 x i16> %b
+  ret <vscale x 16 x i16> %d
+}
+
+declare { <vscale x 32 x i16>, <vscale x 32 x i1> } @llvm.umul.with.overflow.nxv32i16(<vscale x 32 x i16>, <vscale x 32 x i16>)
+
+define <vscale x 32 x i16> @umulo_nxv32i16(<vscale x 32 x i16> %x, <vscale x 32 x i16> %y) {
+; CHECK-LABEL: umulo_nxv32i16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e16, m8, ta, mu
+; CHECK-NEXT:    vmulhu.vv v24, v8, v16
+; CHECK-NEXT:    vmsne.vi v0, v24, 0
+; CHECK-NEXT:    vmul.vv v8, v8, v16
+; CHECK-NEXT:    vmerge.vim v8, v8, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 32 x i16>, <vscale x 32 x i1> } @llvm.umul.with.overflow.nxv32i16(<vscale x 32 x i16> %x, <vscale x 32 x i16> %y)
+  %b = extractvalue { <vscale x 32 x i16>, <vscale x 32 x i1> } %a, 0
+  %c = extractvalue { <vscale x 32 x i16>, <vscale x 32 x i1> } %a, 1
+  %d = select <vscale x 32 x i1> %c, <vscale x 32 x i16> zeroinitializer, <vscale x 32 x i16> %b
+  ret <vscale x 32 x i16> %d
+}
+
+declare { <vscale x 1 x i32>, <vscale x 1 x i1> } @llvm.umul.with.overflow.nxv1i32(<vscale x 1 x i32>, <vscale x 1 x i32>)
+
+define <vscale x 1 x i32> @umulo_nxv1i32(<vscale x 1 x i32> %x, <vscale x 1 x i32> %y) {
+; CHECK-LABEL: umulo_nxv1i32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e32, mf2, ta, mu
+; CHECK-NEXT:    vmulhu.vv v25, v8, v9
+; CHECK-NEXT:    vmsne.vi v0, v25, 0
+; CHECK-NEXT:    vmul.vv v25, v8, v9
+; CHECK-NEXT:    vmerge.vim v8, v25, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 1 x i32>, <vscale x 1 x i1> } @llvm.umul.with.overflow.nxv1i32(<vscale x 1 x i32> %x, <vscale x 1 x i32> %y)
+  %b = extractvalue { <vscale x 1 x i32>, <vscale x 1 x i1> } %a, 0
+  %c = extractvalue { <vscale x 1 x i32>, <vscale x 1 x i1> } %a, 1
+  %d = select <vscale x 1 x i1> %c, <vscale x 1 x i32> zeroinitializer, <vscale x 1 x i32> %b
+  ret <vscale x 1 x i32> %d
+}
+
+declare { <vscale x 2 x i32>, <vscale x 2 x i1> } @llvm.umul.with.overflow.nxv2i32(<vscale x 2 x i32>, <vscale x 2 x i32>)
+
+define <vscale x 2 x i32> @umulo_nxv2i32(<vscale x 2 x i32> %x, <vscale x 2 x i32> %y) {
+; CHECK-LABEL: umulo_nxv2i32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e32, m1, ta, mu
+; CHECK-NEXT:    vmulhu.vv v25, v8, v9
+; CHECK-NEXT:    vmsne.vi v0, v25, 0
+; CHECK-NEXT:    vmul.vv v25, v8, v9
+; CHECK-NEXT:    vmerge.vim v8, v25, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 2 x i32>, <vscale x 2 x i1> } @llvm.umul.with.overflow.nxv2i32(<vscale x 2 x i32> %x, <vscale x 2 x i32> %y)
+  %b = extractvalue { <vscale x 2 x i32>, <vscale x 2 x i1> } %a, 0
+  %c = extractvalue { <vscale x 2 x i32>, <vscale x 2 x i1> } %a, 1
+  %d = select <vscale x 2 x i1> %c, <vscale x 2 x i32> zeroinitializer, <vscale x 2 x i32> %b
+  ret <vscale x 2 x i32> %d
+}
+
+declare { <vscale x 4 x i32>, <vscale x 4 x i1> } @llvm.umul.with.overflow.nxv4i32(<vscale x 4 x i32>, <vscale x 4 x i32>)
+
+define <vscale x 4 x i32> @umulo_nxv4i32(<vscale x 4 x i32> %x, <vscale x 4 x i32> %y) {
+; CHECK-LABEL: umulo_nxv4i32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e32, m2, ta, mu
+; CHECK-NEXT:    vmulhu.vv v26, v8, v10
+; CHECK-NEXT:    vmsne.vi v0, v26, 0
+; CHECK-NEXT:    vmul.vv v26, v8, v10
+; CHECK-NEXT:    vmerge.vim v8, v26, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 4 x i32>, <vscale x 4 x i1> } @llvm.umul.with.overflow.nxv4i32(<vscale x 4 x i32> %x, <vscale x 4 x i32> %y)
+  %b = extractvalue { <vscale x 4 x i32>, <vscale x 4 x i1> } %a, 0
+  %c = extractvalue { <vscale x 4 x i32>, <vscale x 4 x i1> } %a, 1
+  %d = select <vscale x 4 x i1> %c, <vscale x 4 x i32> zeroinitializer, <vscale x 4 x i32> %b
+  ret <vscale x 4 x i32> %d
+}
+
+declare { <vscale x 8 x i32>, <vscale x 8 x i1> } @llvm.umul.with.overflow.nxv8i32(<vscale x 8 x i32>, <vscale x 8 x i32>)
+
+define <vscale x 8 x i32> @umulo_nxv8i32(<vscale x 8 x i32> %x, <vscale x 8 x i32> %y) {
+; CHECK-LABEL: umulo_nxv8i32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e32, m4, ta, mu
+; CHECK-NEXT:    vmulhu.vv v28, v8, v12
+; CHECK-NEXT:    vmsne.vi v0, v28, 0
+; CHECK-NEXT:    vmul.vv v28, v8, v12
+; CHECK-NEXT:    vmerge.vim v8, v28, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 8 x i32>, <vscale x 8 x i1> } @llvm.umul.with.overflow.nxv8i32(<vscale x 8 x i32> %x, <vscale x 8 x i32> %y)
+  %b = extractvalue { <vscale x 8 x i32>, <vscale x 8 x i1> } %a, 0
+  %c = extractvalue { <vscale x 8 x i32>, <vscale x 8 x i1> } %a, 1
+  %d = select <vscale x 8 x i1> %c, <vscale x 8 x i32> zeroinitializer, <vscale x 8 x i32> %b
+  ret <vscale x 8 x i32> %d
+}
+
+declare { <vscale x 16 x i32>, <vscale x 16 x i1> } @llvm.umul.with.overflow.nxv16i32(<vscale x 16 x i32>, <vscale x 16 x i32>)
+
+define <vscale x 16 x i32> @umulo_nxv16i32(<vscale x 16 x i32> %x, <vscale x 16 x i32> %y) {
+; CHECK-LABEL: umulo_nxv16i32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e32, m8, ta, mu
+; CHECK-NEXT:    vmulhu.vv v24, v8, v16
+; CHECK-NEXT:    vmsne.vi v0, v24, 0
+; CHECK-NEXT:    vmul.vv v8, v8, v16
+; CHECK-NEXT:    vmerge.vim v8, v8, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 16 x i32>, <vscale x 16 x i1> } @llvm.umul.with.overflow.nxv16i32(<vscale x 16 x i32> %x, <vscale x 16 x i32> %y)
+  %b = extractvalue { <vscale x 16 x i32>, <vscale x 16 x i1> } %a, 0
+  %c = extractvalue { <vscale x 16 x i32>, <vscale x 16 x i1> } %a, 1
+  %d = select <vscale x 16 x i1> %c, <vscale x 16 x i32> zeroinitializer, <vscale x 16 x i32> %b
+  ret <vscale x 16 x i32> %d
+}
+
+declare { <vscale x 1 x i64>, <vscale x 1 x i1> } @llvm.umul.with.overflow.nxv1i64(<vscale x 1 x i64>, <vscale x 1 x i64>)
+
+define <vscale x 1 x i64> @umulo_nxv1i64(<vscale x 1 x i64> %x, <vscale x 1 x i64> %y) {
+; CHECK-LABEL: umulo_nxv1i64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e64, m1, ta, mu
+; CHECK-NEXT:    vmulhu.vv v25, v8, v9
+; CHECK-NEXT:    vmsne.vi v0, v25, 0
+; CHECK-NEXT:    vmul.vv v25, v8, v9
+; CHECK-NEXT:    vmerge.vim v8, v25, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 1 x i64>, <vscale x 1 x i1> } @llvm.umul.with.overflow.nxv1i64(<vscale x 1 x i64> %x, <vscale x 1 x i64> %y)
+  %b = extractvalue { <vscale x 1 x i64>, <vscale x 1 x i1> } %a, 0
+  %c = extractvalue { <vscale x 1 x i64>, <vscale x 1 x i1> } %a, 1
+  %d = select <vscale x 1 x i1> %c, <vscale x 1 x i64> zeroinitializer, <vscale x 1 x i64> %b
+  ret <vscale x 1 x i64> %d
+}
+
+declare { <vscale x 2 x i64>, <vscale x 2 x i1> } @llvm.umul.with.overflow.nxv2i64(<vscale x 2 x i64>, <vscale x 2 x i64>)
+
+define <vscale x 2 x i64> @umulo_nxv2i64(<vscale x 2 x i64> %x, <vscale x 2 x i64> %y) {
+; CHECK-LABEL: umulo_nxv2i64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e64, m2, ta, mu
+; CHECK-NEXT:    vmulhu.vv v26, v8, v10
+; CHECK-NEXT:    vmsne.vi v0, v26, 0
+; CHECK-NEXT:    vmul.vv v26, v8, v10
+; CHECK-NEXT:    vmerge.vim v8, v26, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 2 x i64>, <vscale x 2 x i1> } @llvm.umul.with.overflow.nxv2i64(<vscale x 2 x i64> %x, <vscale x 2 x i64> %y)
+  %b = extractvalue { <vscale x 2 x i64>, <vscale x 2 x i1> } %a, 0
+  %c = extractvalue { <vscale x 2 x i64>, <vscale x 2 x i1> } %a, 1
+  %d = select <vscale x 2 x i1> %c, <vscale x 2 x i64> zeroinitializer, <vscale x 2 x i64> %b
+  ret <vscale x 2 x i64> %d
+}
+
+declare { <vscale x 4 x i64>, <vscale x 4 x i1> } @llvm.umul.with.overflow.nxv4i64(<vscale x 4 x i64>, <vscale x 4 x i64>)
+
+define <vscale x 4 x i64> @umulo_nxv4i64(<vscale x 4 x i64> %x, <vscale x 4 x i64> %y) {
+; CHECK-LABEL: umulo_nxv4i64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e64, m4, ta, mu
+; CHECK-NEXT:    vmulhu.vv v28, v8, v12
+; CHECK-NEXT:    vmsne.vi v0, v28, 0
+; CHECK-NEXT:    vmul.vv v28, v8, v12
+; CHECK-NEXT:    vmerge.vim v8, v28, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 4 x i64>, <vscale x 4 x i1> } @llvm.umul.with.overflow.nxv4i64(<vscale x 4 x i64> %x, <vscale x 4 x i64> %y)
+  %b = extractvalue { <vscale x 4 x i64>, <vscale x 4 x i1> } %a, 0
+  %c = extractvalue { <vscale x 4 x i64>, <vscale x 4 x i1> } %a, 1
+  %d = select <vscale x 4 x i1> %c, <vscale x 4 x i64> zeroinitializer, <vscale x 4 x i64> %b
+  ret <vscale x 4 x i64> %d
+}
+
+declare { <vscale x 8 x i64>, <vscale x 8 x i1> } @llvm.umul.with.overflow.nxv8i64(<vscale x 8 x i64>, <vscale x 8 x i64>)
+
+define <vscale x 8 x i64> @umulo_nxv8i64(<vscale x 8 x i64> %x, <vscale x 8 x i64> %y) {
+; CHECK-LABEL: umulo_nxv8i64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a0, zero, e64, m8, ta, mu
+; CHECK-NEXT:    vmulhu.vv v24, v8, v16
+; CHECK-NEXT:    vmsne.vi v0, v24, 0
+; CHECK-NEXT:    vmul.vv v8, v8, v16
+; CHECK-NEXT:    vmerge.vim v8, v8, 0, v0
+; CHECK-NEXT:    ret
+  %a = call { <vscale x 8 x i64>, <vscale x 8 x i1> } @llvm.umul.with.overflow.nxv8i64(<vscale x 8 x i64> %x, <vscale x 8 x i64> %y)
+  %b = extractvalue { <vscale x 8 x i64>, <vscale x 8 x i1> } %a, 0
+  %c = extractvalue { <vscale x 8 x i64>, <vscale x 8 x i1> } %a, 1
+  %d = select <vscale x 8 x i1> %c, <vscale x 8 x i64> zeroinitializer, <vscale x 8 x i64> %b
+  ret <vscale x 8 x i64> %d
+}


        
