# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
# RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=tahiti -O0 -run-pass=legalizer -o - %s | FileCheck -check-prefix=GFX6 %s
# RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=fiji -O0 -run-pass=legalizer -o - %s | FileCheck -check-prefix=GFX8 %s
# RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx900 -O0 -run-pass=legalizer -o - %s | FileCheck -check-prefix=GFX9 %s
# RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx1010 -O0 -run-pass=legalizer -o - %s | FileCheck -check-prefix=GFX10 %s
# RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx1100 -O0 -run-pass=legalizer -o - %s | FileCheck -check-prefix=GFX10 %s

---
name: test_mul_s32
body: |
  bb.0:
    liveins: $vgpr0, $vgpr1

    ; GFX6-LABEL: name: test_mul_s32
    ; GFX6: liveins: $vgpr0, $vgpr1
    ; GFX6-NEXT: {{  $}}
    ; GFX6-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
    ; GFX6-NEXT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
    ; GFX6-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[COPY]], [[COPY1]]
    ; GFX6-NEXT: $vgpr0 = COPY [[MUL]](s32)
    ; GFX8-LABEL: name: test_mul_s32
    ; GFX8: liveins: $vgpr0, $vgpr1
    ; GFX8-NEXT: {{  $}}
    ; GFX8-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
    ; GFX8-NEXT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
    ; GFX8-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[COPY]], [[COPY1]]
    ; GFX8-NEXT: $vgpr0 = COPY [[MUL]](s32)
    ; GFX9-LABEL: name: test_mul_s32
    ; GFX9: liveins: $vgpr0, $vgpr1
    ; GFX9-NEXT: {{  $}}
    ; GFX9-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
    ; GFX9-NEXT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
    ; GFX9-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[COPY]], [[COPY1]]
    ; GFX9-NEXT: $vgpr0 = COPY [[MUL]](s32)
    ; GFX10-LABEL: name: test_mul_s32
    ; GFX10: liveins: $vgpr0, $vgpr1
    ; GFX10-NEXT: {{  $}}
    ; GFX10-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
    ; GFX10-NEXT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
    ; GFX10-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[COPY]], [[COPY1]]
    ; GFX10-NEXT: $vgpr0 = COPY [[MUL]](s32)
    %0:_(s32) = COPY $vgpr0
    %1:_(s32) = COPY $vgpr1
    %2:_(s32) = G_MUL %0, %1
    $vgpr0 = COPY %2
...

---
name: test_mul_v2s32
body: |
  bb.0:
    liveins: $vgpr0_vgpr1, $vgpr2_vgpr3

    ; GFX6-LABEL: name: test_mul_v2s32
    ; GFX6: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
    ; GFX6-NEXT: {{  $}}
    ; GFX6-NEXT: [[COPY:%[0-9]+]]:_(<2 x s32>) = COPY $vgpr0_vgpr1
    ; GFX6-NEXT: [[COPY1:%[0-9]+]]:_(<2 x s32>) = COPY $vgpr2_vgpr3
    ; GFX6-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<2 x s32>)
    ; GFX6-NEXT: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<2 x s32>)
    ; GFX6-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[UV]], [[UV2]]
    ; GFX6-NEXT: [[MUL1:%[0-9]+]]:_(s32) = G_MUL [[UV1]], [[UV3]]
    ; GFX6-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[MUL]](s32), [[MUL1]](s32)
    ; GFX6-NEXT: $vgpr0_vgpr1 = COPY [[BUILD_VECTOR]](<2 x s32>)
    ; GFX8-LABEL: name: test_mul_v2s32
    ; GFX8: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
    ; GFX8-NEXT: {{  $}}
    ; GFX8-NEXT: [[COPY:%[0-9]+]]:_(<2 x s32>) = COPY $vgpr0_vgpr1
    ; GFX8-NEXT: [[COPY1:%[0-9]+]]:_(<2 x s32>) = COPY $vgpr2_vgpr3
    ; GFX8-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<2 x s32>)
    ; GFX8-NEXT: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<2 x s32>)
    ; GFX8-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[UV]], [[UV2]]
    ; GFX8-NEXT: [[MUL1:%[0-9]+]]:_(s32) = G_MUL [[UV1]], [[UV3]]
    ; GFX8-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[MUL]](s32), [[MUL1]](s32)
    ; GFX8-NEXT: $vgpr0_vgpr1 = COPY [[BUILD_VECTOR]](<2 x s32>)
    ; GFX9-LABEL: name: test_mul_v2s32
    ; GFX9: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
    ; GFX9-NEXT: {{  $}}
    ; GFX9-NEXT: [[COPY:%[0-9]+]]:_(<2 x s32>) = COPY $vgpr0_vgpr1
    ; GFX9-NEXT: [[COPY1:%[0-9]+]]:_(<2 x s32>) = COPY $vgpr2_vgpr3
    ; GFX9-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<2 x s32>)
    ; GFX9-NEXT: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<2 x s32>)
    ; GFX9-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[UV]], [[UV2]]
    ; GFX9-NEXT: [[MUL1:%[0-9]+]]:_(s32) = G_MUL [[UV1]], [[UV3]]
    ; GFX9-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[MUL]](s32), [[MUL1]](s32)
    ; GFX9-NEXT: $vgpr0_vgpr1 = COPY [[BUILD_VECTOR]](<2 x s32>)
    ; GFX10-LABEL: name: test_mul_v2s32
    ; GFX10: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
    ; GFX10-NEXT: {{  $}}
    ; GFX10-NEXT: [[COPY:%[0-9]+]]:_(<2 x s32>) = COPY $vgpr0_vgpr1
    ; GFX10-NEXT: [[COPY1:%[0-9]+]]:_(<2 x s32>) = COPY $vgpr2_vgpr3
    ; GFX10-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<2 x s32>)
    ; GFX10-NEXT: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<2 x s32>)
    ; GFX10-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[UV]], [[UV2]]
    ; GFX10-NEXT: [[MUL1:%[0-9]+]]:_(s32) = G_MUL [[UV1]], [[UV3]]
    ; GFX10-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[MUL]](s32), [[MUL1]](s32)
    ; GFX10-NEXT: $vgpr0_vgpr1 = COPY [[BUILD_VECTOR]](<2 x s32>)
    %0:_(<2 x s32>) = COPY $vgpr0_vgpr1
    %1:_(<2 x s32>) = COPY $vgpr2_vgpr3
    %2:_(<2 x s32>) = G_MUL %0, %1
    $vgpr0_vgpr1 = COPY %2
...

---
name: test_mul_s64
body: |
  bb.0:
    liveins: $vgpr0_vgpr1, $vgpr2_vgpr3

    ; GFX6-LABEL: name: test_mul_s64
    ; GFX6: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
    ; GFX6-NEXT: {{  $}}
    ; GFX6-NEXT: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1
    ; GFX6-NEXT: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3
    ; GFX6-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](s64)
    ; GFX6-NEXT: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s64)
    ; GFX6-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[UV]], [[UV2]]
    ; GFX6-NEXT: [[MUL1:%[0-9]+]]:_(s32) = G_MUL [[UV1]], [[UV2]]
    ; GFX6-NEXT: [[MUL2:%[0-9]+]]:_(s32) = G_MUL [[UV]], [[UV3]]
    ; GFX6-NEXT: [[UMULH:%[0-9]+]]:_(s32) = G_UMULH [[UV]], [[UV2]]
    ; GFX6-NEXT: [[ADD:%[0-9]+]]:_(s32) = G_ADD [[MUL1]], [[MUL2]]
    ; GFX6-NEXT: [[ADD1:%[0-9]+]]:_(s32) = G_ADD [[ADD]], [[UMULH]]
    ; GFX6-NEXT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[MUL]](s32), [[ADD1]](s32)
    ; GFX6-NEXT: $vgpr0_vgpr1 = COPY [[MV]](s64)
    ; GFX8-LABEL: name: test_mul_s64
    ; GFX8: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
    ; GFX8-NEXT: {{  $}}
    ; GFX8-NEXT: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1
    ; GFX8-NEXT: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3
    ; GFX8-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](s64)
    ; GFX8-NEXT: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s64)
    ; GFX8-NEXT: [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 0
    ; GFX8-NEXT: [[AMDGPU_MAD_U64_U32_:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_1:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV]](s32), [[UV2]], [[C]]
    ; GFX8-NEXT: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_]](s64)
    ; GFX8-NEXT: [[ANYEXT:%[0-9]+]]:_(s64) = G_ANYEXT [[UV5]](s32)
    ; GFX8-NEXT: [[AMDGPU_MAD_U64_U32_2:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_3:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV]](s32), [[UV3]], [[ANYEXT]]
    ; GFX8-NEXT: [[AMDGPU_MAD_U64_U32_4:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_5:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV1]](s32), [[UV2]], [[AMDGPU_MAD_U64_U32_2]]
    ; GFX8-NEXT: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_4]](s64)
    ; GFX8-NEXT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UV4]](s32), [[UV6]](s32)
    ; GFX8-NEXT: $vgpr0_vgpr1 = COPY [[MV]](s64)
    ; GFX9-LABEL: name: test_mul_s64
    ; GFX9: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
    ; GFX9-NEXT: {{  $}}
    ; GFX9-NEXT: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1
    ; GFX9-NEXT: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3
    ; GFX9-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](s64)
    ; GFX9-NEXT: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s64)
    ; GFX9-NEXT: [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 0
    ; GFX9-NEXT: [[AMDGPU_MAD_U64_U32_:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_1:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV]](s32), [[UV2]], [[C]]
    ; GFX9-NEXT: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_]](s64)
    ; GFX9-NEXT: [[ANYEXT:%[0-9]+]]:_(s64) = G_ANYEXT [[UV5]](s32)
    ; GFX9-NEXT: [[AMDGPU_MAD_U64_U32_2:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_3:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV]](s32), [[UV3]], [[ANYEXT]]
    ; GFX9-NEXT: [[AMDGPU_MAD_U64_U32_4:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_5:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV1]](s32), [[UV2]], [[AMDGPU_MAD_U64_U32_2]]
    ; GFX9-NEXT: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_4]](s64)
    ; GFX9-NEXT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UV4]](s32), [[UV6]](s32)
    ; GFX9-NEXT: $vgpr0_vgpr1 = COPY [[MV]](s64)
    ; GFX10-LABEL: name: test_mul_s64
    ; GFX10: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
    ; GFX10-NEXT: {{  $}}
    ; GFX10-NEXT: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1
    ; GFX10-NEXT: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3
    ; GFX10-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](s64)
    ; GFX10-NEXT: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s64)
    ; GFX10-NEXT: [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 0
    ; GFX10-NEXT: [[AMDGPU_MAD_U64_U32_:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_1:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV]](s32), [[UV2]], [[C]]
    ; GFX10-NEXT: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_]](s64)
    ; GFX10-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[UV]], [[UV3]]
    ; GFX10-NEXT: [[ADD:%[0-9]+]]:_(s32) = G_ADD [[UV5]], [[MUL]]
    ; GFX10-NEXT: [[MUL1:%[0-9]+]]:_(s32) = G_MUL [[UV1]], [[UV2]]
    ; GFX10-NEXT: [[ADD1:%[0-9]+]]:_(s32) = G_ADD [[ADD]], [[MUL1]]
    ; GFX10-NEXT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UV4]](s32), [[ADD1]](s32)
    ; GFX10-NEXT: $vgpr0_vgpr1 = COPY [[MV]](s64)
    %0:_(s64) = COPY $vgpr0_vgpr1
    %1:_(s64) = COPY $vgpr2_vgpr3
    %2:_(s64) = G_MUL %0, %1
    $vgpr0_vgpr1 = COPY %2
...

---
name: test_mul_v2s64
body: |
  bb.0:
    liveins: $vgpr0_vgpr1_vgpr2_vgpr3, $vgpr4_vgpr5_vgpr6_vgpr7

    ; GFX6-LABEL: name: test_mul_v2s64
    ; GFX6: liveins: $vgpr0_vgpr1_vgpr2_vgpr3, $vgpr4_vgpr5_vgpr6_vgpr7
    ; GFX6-NEXT: {{  $}}
    ; GFX6-NEXT: [[COPY:%[0-9]+]]:_(<2 x s64>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3
    ; GFX6-NEXT: [[COPY1:%[0-9]+]]:_(<2 x s64>) = COPY $vgpr4_vgpr5_vgpr6_vgpr7
    ; GFX6-NEXT: [[UV:%[0-9]+]]:_(s64), [[UV1:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY]](<2 x s64>)
    ; GFX6-NEXT: [[UV2:%[0-9]+]]:_(s64), [[UV3:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY1]](<2 x s64>)
    ; GFX6-NEXT: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV]](s64)
    ; GFX6-NEXT: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV2]](s64)
    ; GFX6-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[UV4]], [[UV6]]
    ; GFX6-NEXT: [[MUL1:%[0-9]+]]:_(s32) = G_MUL [[UV5]], [[UV6]]
    ; GFX6-NEXT: [[MUL2:%[0-9]+]]:_(s32) = G_MUL [[UV4]], [[UV7]]
    ; GFX6-NEXT: [[UMULH:%[0-9]+]]:_(s32) = G_UMULH [[UV4]], [[UV6]]
    ; GFX6-NEXT: [[ADD:%[0-9]+]]:_(s32) = G_ADD [[MUL1]], [[MUL2]]
    ; GFX6-NEXT: [[ADD1:%[0-9]+]]:_(s32) = G_ADD [[ADD]], [[UMULH]]
    ; GFX6-NEXT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[MUL]](s32), [[ADD1]](s32)
    ; GFX6-NEXT: [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV1]](s64)
    ; GFX6-NEXT: [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV3]](s64)
    ; GFX6-NEXT: [[MUL3:%[0-9]+]]:_(s32) = G_MUL [[UV8]], [[UV10]]
    ; GFX6-NEXT: [[MUL4:%[0-9]+]]:_(s32) = G_MUL [[UV9]], [[UV10]]
    ; GFX6-NEXT: [[MUL5:%[0-9]+]]:_(s32) = G_MUL [[UV8]], [[UV11]]
    ; GFX6-NEXT: [[UMULH1:%[0-9]+]]:_(s32) = G_UMULH [[UV8]], [[UV10]]
    ; GFX6-NEXT: [[ADD2:%[0-9]+]]:_(s32) = G_ADD [[MUL4]], [[MUL5]]
    ; GFX6-NEXT: [[ADD3:%[0-9]+]]:_(s32) = G_ADD [[ADD2]], [[UMULH1]]
    ; GFX6-NEXT: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[MUL3]](s32), [[ADD3]](s32)
    ; GFX6-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[MV]](s64), [[MV1]](s64)
    ; GFX6-NEXT: $vgpr0_vgpr1_vgpr2_vgpr3 = COPY [[BUILD_VECTOR]](<2 x s64>)
    ; GFX8-LABEL: name: test_mul_v2s64
    ; GFX8: liveins: $vgpr0_vgpr1_vgpr2_vgpr3, $vgpr4_vgpr5_vgpr6_vgpr7
    ; GFX8-NEXT: {{  $}}
    ; GFX8-NEXT: [[COPY:%[0-9]+]]:_(<2 x s64>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3
    ; GFX8-NEXT: [[COPY1:%[0-9]+]]:_(<2 x s64>) = COPY $vgpr4_vgpr5_vgpr6_vgpr7
    ; GFX8-NEXT: [[UV:%[0-9]+]]:_(s64), [[UV1:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY]](<2 x s64>)
    ; GFX8-NEXT: [[UV2:%[0-9]+]]:_(s64), [[UV3:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY1]](<2 x s64>)
    ; GFX8-NEXT: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV]](s64)
    ; GFX8-NEXT: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV2]](s64)
    ; GFX8-NEXT: [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 0
    ; GFX8-NEXT: [[AMDGPU_MAD_U64_U32_:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_1:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV4]](s32), [[UV6]], [[C]]
    ; GFX8-NEXT: [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_]](s64)
    ; GFX8-NEXT: [[ANYEXT:%[0-9]+]]:_(s64) = G_ANYEXT [[UV9]](s32)
    ; GFX8-NEXT: [[AMDGPU_MAD_U64_U32_2:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_3:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV4]](s32), [[UV7]], [[ANYEXT]]
    ; GFX8-NEXT: [[AMDGPU_MAD_U64_U32_4:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_5:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV5]](s32), [[UV6]], [[AMDGPU_MAD_U64_U32_2]]
    ; GFX8-NEXT: [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_4]](s64)
    ; GFX8-NEXT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UV8]](s32), [[UV10]](s32)
    ; GFX8-NEXT: [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV1]](s64)
    ; GFX8-NEXT: [[UV14:%[0-9]+]]:_(s32), [[UV15:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV3]](s64)
    ; GFX8-NEXT: [[AMDGPU_MAD_U64_U32_6:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_7:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV12]](s32), [[UV14]], [[C]]
    ; GFX8-NEXT: [[UV16:%[0-9]+]]:_(s32), [[UV17:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_6]](s64)
    ; GFX8-NEXT: [[ANYEXT1:%[0-9]+]]:_(s64) = G_ANYEXT [[UV17]](s32)
    ; GFX8-NEXT: [[AMDGPU_MAD_U64_U32_8:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_9:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV12]](s32), [[UV15]], [[ANYEXT1]]
    ; GFX8-NEXT: [[AMDGPU_MAD_U64_U32_10:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_11:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV13]](s32), [[UV14]], [[AMDGPU_MAD_U64_U32_8]]
    ; GFX8-NEXT: [[UV18:%[0-9]+]]:_(s32), [[UV19:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_10]](s64)
    ; GFX8-NEXT: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UV16]](s32), [[UV18]](s32)
    ; GFX8-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[MV]](s64), [[MV1]](s64)
    ; GFX8-NEXT: $vgpr0_vgpr1_vgpr2_vgpr3 = COPY [[BUILD_VECTOR]](<2 x s64>)
    ; GFX9-LABEL: name: test_mul_v2s64
    ; GFX9: liveins: $vgpr0_vgpr1_vgpr2_vgpr3, $vgpr4_vgpr5_vgpr6_vgpr7
    ; GFX9-NEXT: {{  $}}
    ; GFX9-NEXT: [[COPY:%[0-9]+]]:_(<2 x s64>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3
    ; GFX9-NEXT: [[COPY1:%[0-9]+]]:_(<2 x s64>) = COPY $vgpr4_vgpr5_vgpr6_vgpr7
    ; GFX9-NEXT: [[UV:%[0-9]+]]:_(s64), [[UV1:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY]](<2 x s64>)
    ; GFX9-NEXT: [[UV2:%[0-9]+]]:_(s64), [[UV3:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY1]](<2 x s64>)
    ; GFX9-NEXT: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV]](s64)
    ; GFX9-NEXT: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV2]](s64)
    ; GFX9-NEXT: [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 0
    ; GFX9-NEXT: [[AMDGPU_MAD_U64_U32_:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_1:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV4]](s32), [[UV6]], [[C]]
    ; GFX9-NEXT: [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_]](s64)
    ; GFX9-NEXT: [[ANYEXT:%[0-9]+]]:_(s64) = G_ANYEXT [[UV9]](s32)
    ; GFX9-NEXT: [[AMDGPU_MAD_U64_U32_2:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_3:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV4]](s32), [[UV7]], [[ANYEXT]]
    ; GFX9-NEXT: [[AMDGPU_MAD_U64_U32_4:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_5:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV5]](s32), [[UV6]], [[AMDGPU_MAD_U64_U32_2]]
    ; GFX9-NEXT: [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_4]](s64)
    ; GFX9-NEXT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UV8]](s32), [[UV10]](s32)
    ; GFX9-NEXT: [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV1]](s64)
    ; GFX9-NEXT: [[UV14:%[0-9]+]]:_(s32), [[UV15:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV3]](s64)
    ; GFX9-NEXT: [[AMDGPU_MAD_U64_U32_6:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_7:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV12]](s32), [[UV14]], [[C]]
    ; GFX9-NEXT: [[UV16:%[0-9]+]]:_(s32), [[UV17:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_6]](s64)
    ; GFX9-NEXT: [[ANYEXT1:%[0-9]+]]:_(s64) = G_ANYEXT [[UV17]](s32)
    ; GFX9-NEXT: [[AMDGPU_MAD_U64_U32_8:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_9:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV12]](s32), [[UV15]], [[ANYEXT1]]
    ; GFX9-NEXT: [[AMDGPU_MAD_U64_U32_10:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_11:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV13]](s32), [[UV14]], [[AMDGPU_MAD_U64_U32_8]]
    ; GFX9-NEXT: [[UV18:%[0-9]+]]:_(s32), [[UV19:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_10]](s64)
    ; GFX9-NEXT: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UV16]](s32), [[UV18]](s32)
    ; GFX9-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[MV]](s64), [[MV1]](s64)
    ; GFX9-NEXT: $vgpr0_vgpr1_vgpr2_vgpr3 = COPY [[BUILD_VECTOR]](<2 x s64>)
    ; GFX10-LABEL: name: test_mul_v2s64
    ; GFX10: liveins: $vgpr0_vgpr1_vgpr2_vgpr3, $vgpr4_vgpr5_vgpr6_vgpr7
    ; GFX10-NEXT: {{  $}}
    ; GFX10-NEXT: [[COPY:%[0-9]+]]:_(<2 x s64>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3
    ; GFX10-NEXT: [[COPY1:%[0-9]+]]:_(<2 x s64>) = COPY $vgpr4_vgpr5_vgpr6_vgpr7
    ; GFX10-NEXT: [[UV:%[0-9]+]]:_(s64), [[UV1:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY]](<2 x s64>)
    ; GFX10-NEXT: [[UV2:%[0-9]+]]:_(s64), [[UV3:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY1]](<2 x s64>)
    ; GFX10-NEXT: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV]](s64)
    ; GFX10-NEXT: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV2]](s64)
    ; GFX10-NEXT: [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 0
    ; GFX10-NEXT: [[AMDGPU_MAD_U64_U32_:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_1:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV4]](s32), [[UV6]], [[C]]
    ; GFX10-NEXT: [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_]](s64)
    ; GFX10-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[UV4]], [[UV7]]
    ; GFX10-NEXT: [[ADD:%[0-9]+]]:_(s32) = G_ADD [[UV9]], [[MUL]]
    ; GFX10-NEXT: [[MUL1:%[0-9]+]]:_(s32) = G_MUL [[UV5]], [[UV6]]
    ; GFX10-NEXT: [[ADD1:%[0-9]+]]:_(s32) = G_ADD [[ADD]], [[MUL1]]
    ; GFX10-NEXT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UV8]](s32), [[ADD1]](s32)
    ; GFX10-NEXT: [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV1]](s64)
    ; GFX10-NEXT: [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV3]](s64)
    ; GFX10-NEXT: [[AMDGPU_MAD_U64_U32_2:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_3:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV10]](s32), [[UV12]], [[C]]
    ; GFX10-NEXT: [[UV14:%[0-9]+]]:_(s32), [[UV15:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_2]](s64)
    ; GFX10-NEXT: [[MUL2:%[0-9]+]]:_(s32) = G_MUL [[UV10]], [[UV13]]
    ; GFX10-NEXT: [[ADD2:%[0-9]+]]:_(s32) = G_ADD [[UV15]], [[MUL2]]
    ; GFX10-NEXT: [[MUL3:%[0-9]+]]:_(s32) = G_MUL [[UV11]], [[UV12]]
    ; GFX10-NEXT: [[ADD3:%[0-9]+]]:_(s32) = G_ADD [[ADD2]], [[MUL3]]
    ; GFX10-NEXT: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UV14]](s32), [[ADD3]](s32)
    ; GFX10-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[MV]](s64), [[MV1]](s64)
    ; GFX10-NEXT: $vgpr0_vgpr1_vgpr2_vgpr3 = COPY [[BUILD_VECTOR]](<2 x s64>)
    %0:_(<2 x s64>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3
    %1:_(<2 x s64>) = COPY $vgpr4_vgpr5_vgpr6_vgpr7
    %2:_(<2 x s64>) = G_MUL %0, %1
    $vgpr0_vgpr1_vgpr2_vgpr3 = COPY %2
...

---
name: test_mul_s16
body: |
  bb.0:
    liveins: $vgpr0, $vgpr1

    ; GFX6-LABEL: name: test_mul_s16
    ; GFX6: liveins: $vgpr0, $vgpr1
    ; GFX6-NEXT: {{  $}}
    ; GFX6-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
    ; GFX6-NEXT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
    ; GFX6-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[COPY]], [[COPY1]]
    ; GFX6-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 65535
    ; GFX6-NEXT: [[AND:%[0-9]+]]:_(s32) = G_AND [[MUL]], [[C]]
    ; GFX6-NEXT: $vgpr0 = COPY [[AND]](s32)
    ; GFX8-LABEL: name: test_mul_s16
    ; GFX8: liveins: $vgpr0, $vgpr1
    ; GFX8-NEXT: {{  $}}
    ; GFX8-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
    ; GFX8-NEXT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
    ; GFX8-NEXT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
    ; GFX8-NEXT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
    ; GFX8-NEXT: [[MUL:%[0-9]+]]:_(s16) = G_MUL [[TRUNC]], [[TRUNC1]]
    ; GFX8-NEXT: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[MUL]](s16)
    ; GFX8-NEXT: $vgpr0 = COPY [[ZEXT]](s32)
    ; GFX9-LABEL: name: test_mul_s16
    ; GFX9: liveins: $vgpr0, $vgpr1
    ; GFX9-NEXT: {{  $}}
    ; GFX9-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
    ; GFX9-NEXT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
    ; GFX9-NEXT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
    ; GFX9-NEXT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
    ; GFX9-NEXT: [[MUL:%[0-9]+]]:_(s16) = G_MUL [[TRUNC]], [[TRUNC1]]
    ; GFX9-NEXT: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[MUL]](s16)
    ; GFX9-NEXT: $vgpr0 = COPY [[ZEXT]](s32)
    ; GFX10-LABEL: name: test_mul_s16
    ; GFX10: liveins: $vgpr0, $vgpr1
    ; GFX10-NEXT: {{  $}}
    ; GFX10-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
    ; GFX10-NEXT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
    ; GFX10-NEXT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
    ; GFX10-NEXT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
    ; GFX10-NEXT: [[MUL:%[0-9]+]]:_(s16) = G_MUL [[TRUNC]], [[TRUNC1]]
    ; GFX10-NEXT: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[MUL]](s16)
    ; GFX10-NEXT: $vgpr0 = COPY [[ZEXT]](s32)
    %0:_(s32) = COPY $vgpr0
    %1:_(s32) = COPY $vgpr1
    %2:_(s16) = G_TRUNC %0
    %3:_(s16) = G_TRUNC %1
    %4:_(s16) = G_MUL %2, %3
    %5:_(s32) = G_ZEXT %4
    $vgpr0 = COPY %5
...

---
name: test_mul_v2s16
body: |
  bb.0:
    liveins: $vgpr0, $vgpr1

    ; GFX6-LABEL: name: test_mul_v2s16
    ; GFX6: liveins: $vgpr0, $vgpr1
    ; GFX6-NEXT: {{  $}}
    ; GFX6-NEXT: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
    ; GFX6-NEXT: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
    ; GFX6-NEXT: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
    ; GFX6-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
    ; GFX6-NEXT: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
    ; GFX6-NEXT: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>)
    ; GFX6-NEXT: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
    ; GFX6-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[BITCAST]], [[BITCAST1]]
    ; GFX6-NEXT: [[MUL1:%[0-9]+]]:_(s32) = G_MUL [[LSHR]], [[LSHR1]]
    ; GFX6-NEXT: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 65535
    ; GFX6-NEXT: [[AND:%[0-9]+]]:_(s32) = G_AND [[MUL]], [[C1]]
    ; GFX6-NEXT: [[AND1:%[0-9]+]]:_(s32) = G_AND [[MUL1]], [[C1]]
    ; GFX6-NEXT: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[AND1]], [[C]](s32)
    ; GFX6-NEXT: [[OR:%[0-9]+]]:_(s32) = G_OR [[AND]], [[SHL]]
    ; GFX6-NEXT: [[BITCAST2:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
    ; GFX6-NEXT: $vgpr0 = COPY [[BITCAST2]](<2 x s16>)
    ; GFX8-LABEL: name: test_mul_v2s16
    ; GFX8: liveins: $vgpr0, $vgpr1
    ; GFX8-NEXT: {{  $}}
    ; GFX8-NEXT: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
    ; GFX8-NEXT: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
    ; GFX8-NEXT: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
    ; GFX8-NEXT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
    ; GFX8-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
    ; GFX8-NEXT: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
    ; GFX8-NEXT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
    ; GFX8-NEXT: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>)
    ; GFX8-NEXT: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
    ; GFX8-NEXT: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
    ; GFX8-NEXT: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
    ; GFX8-NEXT: [[MUL:%[0-9]+]]:_(s16) = G_MUL [[TRUNC]], [[TRUNC2]]
    ; GFX8-NEXT: [[MUL1:%[0-9]+]]:_(s16) = G_MUL [[TRUNC1]], [[TRUNC3]]
    ; GFX8-NEXT: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[MUL]](s16)
    ; GFX8-NEXT: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[MUL1]](s16)
    ; GFX8-NEXT: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32)
    ; GFX8-NEXT: [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]]
    ; GFX8-NEXT: [[BITCAST2:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
    ; GFX8-NEXT: $vgpr0 = COPY [[BITCAST2]](<2 x s16>)
    ; GFX9-LABEL: name: test_mul_v2s16
    ; GFX9: liveins: $vgpr0, $vgpr1
    ; GFX9-NEXT: {{  $}}
    ; GFX9-NEXT: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
    ; GFX9-NEXT: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
    ; GFX9-NEXT: [[MUL:%[0-9]+]]:_(<2 x s16>) = G_MUL [[COPY]], [[COPY1]]
    ; GFX9-NEXT: $vgpr0 = COPY [[MUL]](<2 x s16>)
    ; GFX10-LABEL: name: test_mul_v2s16
    ; GFX10: liveins: $vgpr0, $vgpr1
    ; GFX10-NEXT: {{  $}}
    ; GFX10-NEXT: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
    ; GFX10-NEXT: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
    ; GFX10-NEXT: [[MUL:%[0-9]+]]:_(<2 x s16>) = G_MUL [[COPY]], [[COPY1]]
    ; GFX10-NEXT: $vgpr0 = COPY [[MUL]](<2 x s16>)
    %0:_(<2 x s16>) = COPY $vgpr0
    %1:_(<2 x s16>) = COPY $vgpr1
    %2:_(<2 x s16>) = G_MUL %0, %1
    $vgpr0 = COPY %2
...

---
name: test_mul_v3s16
body: |
  bb.0:
    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
    ; GFX6-LABEL: name: test_mul_v3s16
    ; GFX6: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
    ; GFX6-NEXT: {{  $}}
    ; GFX6-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
    ; GFX6-NEXT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
    ; GFX6-NEXT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
    ; GFX6-NEXT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
    ; GFX6-NEXT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
    ; GFX6-NEXT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
    ; GFX6-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[COPY]], [[COPY3]]
    ; GFX6-NEXT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[MUL]](s32)
    ; GFX6-NEXT: [[MUL1:%[0-9]+]]:_(s32) = G_MUL [[COPY1]], [[COPY4]]
    ; GFX6-NEXT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[MUL1]](s32)
    ; GFX6-NEXT: [[MUL2:%[0-9]+]]:_(s32) = G_MUL [[COPY2]], [[COPY5]]
    ; GFX6-NEXT: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[MUL2]](s32)
    ; GFX6-NEXT: S_ENDPGM 0, implicit [[TRUNC]](s16), implicit [[TRUNC1]](s16), implicit [[TRUNC2]](s16)
    ; GFX8-LABEL: name: test_mul_v3s16
    ; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
    ; GFX8-NEXT: {{  $}}
    ; GFX8-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
    ; GFX8-NEXT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
    ; GFX8-NEXT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
    ; GFX8-NEXT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
    ; GFX8-NEXT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
    ; GFX8-NEXT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
    ; GFX8-NEXT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
    ; GFX8-NEXT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
    ; GFX8-NEXT: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
    ; GFX8-NEXT: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[COPY3]](s32)
    ; GFX8-NEXT: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[COPY4]](s32)
    ; GFX8-NEXT: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[COPY5]](s32)
    ; GFX8-NEXT: [[MUL:%[0-9]+]]:_(s16) = G_MUL [[TRUNC]], [[TRUNC3]]
    ; GFX8-NEXT: [[MUL1:%[0-9]+]]:_(s16) = G_MUL [[TRUNC1]], [[TRUNC4]]
    ; GFX8-NEXT: [[MUL2:%[0-9]+]]:_(s16) = G_MUL [[TRUNC2]], [[TRUNC5]]
    ; GFX8-NEXT: S_ENDPGM 0, implicit [[MUL]](s16), implicit [[MUL1]](s16), implicit [[MUL2]](s16)
    ; GFX9-LABEL: name: test_mul_v3s16
    ; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
    ; GFX9-NEXT: {{  $}}
    ; GFX9-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
    ; GFX9-NEXT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
    ; GFX9-NEXT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
    ; GFX9-NEXT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
    ; GFX9-NEXT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
    ; GFX9-NEXT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
    ; GFX9-NEXT: [[BUILD_VECTOR_TRUNC:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR_TRUNC [[COPY]](s32), [[COPY1]](s32)
    ; GFX9-NEXT: [[DEF:%[0-9]+]]:_(s32) = G_IMPLICIT_DEF
    ; GFX9-NEXT: [[COPY6:%[0-9]+]]:_(s32) = COPY [[DEF]](s32)
    ; GFX9-NEXT: [[BUILD_VECTOR_TRUNC1:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR_TRUNC [[COPY2]](s32), [[COPY6]](s32)
    ; GFX9-NEXT: [[BUILD_VECTOR_TRUNC2:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR_TRUNC [[COPY3]](s32), [[COPY4]](s32)
    ; GFX9-NEXT: [[BUILD_VECTOR_TRUNC3:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR_TRUNC [[COPY5]](s32), [[DEF]](s32)
    ; GFX9-NEXT: [[MUL:%[0-9]+]]:_(<2 x s16>) = G_MUL [[BUILD_VECTOR_TRUNC]], [[BUILD_VECTOR_TRUNC2]]
    ; GFX9-NEXT: [[MUL1:%[0-9]+]]:_(<2 x s16>) = G_MUL [[BUILD_VECTOR_TRUNC1]], [[BUILD_VECTOR_TRUNC3]]
    ; GFX9-NEXT: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[MUL]](<2 x s16>)
    ; GFX9-NEXT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
    ; GFX9-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
    ; GFX9-NEXT: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
    ; GFX9-NEXT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
    ; GFX9-NEXT: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[MUL1]](<2 x s16>)
    ; GFX9-NEXT: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
    ; GFX9-NEXT: S_ENDPGM 0, implicit [[TRUNC]](s16), implicit [[TRUNC1]](s16), implicit [[TRUNC2]](s16)
    ; GFX10-LABEL: name: test_mul_v3s16
    ; GFX10: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
    ; GFX10-NEXT: {{  $}}
    ; GFX10-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
    ; GFX10-NEXT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
    ; GFX10-NEXT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
    ; GFX10-NEXT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
    ; GFX10-NEXT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
    ; GFX10-NEXT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
    ; GFX10-NEXT: [[BUILD_VECTOR_TRUNC:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR_TRUNC [[COPY]](s32), [[COPY1]](s32)
    ; GFX10-NEXT: [[DEF:%[0-9]+]]:_(s32) = G_IMPLICIT_DEF
    ; GFX10-NEXT: [[COPY6:%[0-9]+]]:_(s32) = COPY [[DEF]](s32)
    ; GFX10-NEXT: [[BUILD_VECTOR_TRUNC1:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR_TRUNC [[COPY2]](s32), [[COPY6]](s32)
    ; GFX10-NEXT: [[BUILD_VECTOR_TRUNC2:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR_TRUNC [[COPY3]](s32), [[COPY4]](s32)
    ; GFX10-NEXT: [[BUILD_VECTOR_TRUNC3:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR_TRUNC [[COPY5]](s32), [[DEF]](s32)
    ; GFX10-NEXT: [[MUL:%[0-9]+]]:_(<2 x s16>) = G_MUL [[BUILD_VECTOR_TRUNC]], [[BUILD_VECTOR_TRUNC2]]
    ; GFX10-NEXT: [[MUL1:%[0-9]+]]:_(<2 x s16>) = G_MUL [[BUILD_VECTOR_TRUNC1]], [[BUILD_VECTOR_TRUNC3]]
    ; GFX10-NEXT: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[MUL]](<2 x s16>)
    ; GFX10-NEXT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
    ; GFX10-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
    ; GFX10-NEXT: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
    ; GFX10-NEXT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
    ; GFX10-NEXT: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[MUL1]](<2 x s16>)
    ; GFX10-NEXT: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
    ; GFX10-NEXT: S_ENDPGM 0, implicit [[TRUNC]](s16), implicit [[TRUNC1]](s16), implicit [[TRUNC2]](s16)
    %0:_(s32) = COPY $vgpr0
    %1:_(s32) = COPY $vgpr1
    %2:_(s32) = COPY $vgpr2
    %3:_(s32) = COPY $vgpr3
    %4:_(s32) = COPY $vgpr4
    %5:_(s32) = COPY $vgpr5
    %6:_(s16) = G_TRUNC %0
    %7:_(s16) = G_TRUNC %1
    %8:_(s16) = G_TRUNC %2
    %9:_(s16) = G_TRUNC %3
    %10:_(s16) = G_TRUNC %4
    %11:_(s16) = G_TRUNC %5
    %12:_(<3 x s16>) = G_BUILD_VECTOR %6, %7, %8
    %13:_(<3 x s16>) = G_BUILD_VECTOR %9, %10, %11
    %14:_(<3 x s16>) = G_MUL %12, %13
    %15:_(s16), %16:_(s16), %17:_(s16) = G_UNMERGE_VALUES %14
    S_ENDPGM 0, implicit %15, implicit %16, implicit %17
...

---
name: test_mul_v4s16
body: |
  bb.0:
    liveins: $vgpr0_vgpr1, $vgpr2_vgpr3

    ; GFX6-LABEL: name: test_mul_v4s16
    ; GFX6: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
    ; GFX6-NEXT: {{  $}}
    ; GFX6-NEXT: [[COPY:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr0_vgpr1
    ; GFX6-NEXT: [[COPY1:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr2_vgpr3
    ; GFX6-NEXT: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY]](<4 x s16>)
    ; GFX6-NEXT: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[UV]](<2 x s16>)
    ; GFX6-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
    ; GFX6-NEXT: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
    ; GFX6-NEXT: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[UV1]](<2 x s16>)
    ; GFX6-NEXT: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
    ; GFX6-NEXT: [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY1]](<4 x s16>)
    ; GFX6-NEXT: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[UV2]](<2 x s16>)
    ; GFX6-NEXT: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32)
    ; GFX6-NEXT: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[UV3]](<2 x s16>)
    ; GFX6-NEXT: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32)
    ; GFX6-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[BITCAST]], [[BITCAST2]]
    ; GFX6-NEXT: [[MUL1:%[0-9]+]]:_(s32) = G_MUL [[LSHR]], [[LSHR2]]
    ; GFX6-NEXT: [[MUL2:%[0-9]+]]:_(s32) = G_MUL [[BITCAST1]], [[BITCAST3]]
    ; GFX6-NEXT: [[MUL3:%[0-9]+]]:_(s32) = G_MUL [[LSHR1]], [[LSHR3]]
    ; GFX6-NEXT: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 65535
    ; GFX6-NEXT: [[AND:%[0-9]+]]:_(s32) = G_AND [[MUL]], [[C1]]
    ; GFX6-NEXT: [[AND1:%[0-9]+]]:_(s32) = G_AND [[MUL1]], [[C1]]
    ; GFX6-NEXT: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[AND1]], [[C]](s32)
    ; GFX6-NEXT: [[OR:%[0-9]+]]:_(s32) = G_OR [[AND]], [[SHL]]
    ; GFX6-NEXT: [[BITCAST4:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
    ; GFX6-NEXT: [[AND2:%[0-9]+]]:_(s32) = G_AND [[MUL2]], [[C1]]
    ; GFX6-NEXT: [[AND3:%[0-9]+]]:_(s32) = G_AND [[MUL3]], [[C1]]
    ; GFX6-NEXT: [[SHL1:%[0-9]+]]:_(s32) = G_SHL [[AND3]], [[C]](s32)
    ; GFX6-NEXT: [[OR1:%[0-9]+]]:_(s32) = G_OR [[AND2]], [[SHL1]]
    ; GFX6-NEXT: [[BITCAST5:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR1]](s32)
    ; GFX6-NEXT: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[BITCAST4]](<2 x s16>), [[BITCAST5]](<2 x s16>)
    ; GFX6-NEXT: $vgpr0_vgpr1 = COPY [[CONCAT_VECTORS]](<4 x s16>)
    ; GFX8-LABEL: name: test_mul_v4s16
    ; GFX8: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
    ; GFX8-NEXT: {{  $}}
    ; GFX8-NEXT: [[COPY:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr0_vgpr1
    ; GFX8-NEXT: [[COPY1:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr2_vgpr3
    ; GFX8-NEXT: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY]](<4 x s16>)
    ; GFX8-NEXT: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[UV]](<2 x s16>)
    ; GFX8-NEXT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
    ; GFX8-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
    ; GFX8-NEXT: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
    ; GFX8-NEXT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
    ; GFX8-NEXT: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[UV1]](<2 x s16>)
    ; GFX8-NEXT: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
    ; GFX8-NEXT: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
    ; GFX8-NEXT: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
    ; GFX8-NEXT: [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY1]](<4 x s16>)
    ; GFX8-NEXT: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[UV2]](<2 x s16>)
    ; GFX8-NEXT: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32)
    ; GFX8-NEXT: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32)
    ; GFX8-NEXT: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32)
    ; GFX8-NEXT: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[UV3]](<2 x s16>)
    ; GFX8-NEXT: [[TRUNC6:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST3]](s32)
    ; GFX8-NEXT: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32)
    ; GFX8-NEXT: [[TRUNC7:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR3]](s32)
    ; GFX8-NEXT: [[MUL:%[0-9]+]]:_(s16) = G_MUL [[TRUNC]], [[TRUNC4]]
    ; GFX8-NEXT: [[MUL1:%[0-9]+]]:_(s16) = G_MUL [[TRUNC1]], [[TRUNC5]]
    ; GFX8-NEXT: [[MUL2:%[0-9]+]]:_(s16) = G_MUL [[TRUNC2]], [[TRUNC6]]
    ; GFX8-NEXT: [[MUL3:%[0-9]+]]:_(s16) = G_MUL [[TRUNC3]], [[TRUNC7]]
    ; GFX8-NEXT: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[MUL]](s16)
    ; GFX8-NEXT: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[MUL1]](s16)
    ; GFX8-NEXT: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32)
    ; GFX8-NEXT: [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]]
    ; GFX8-NEXT: [[BITCAST4:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
    ; GFX8-NEXT: [[ZEXT2:%[0-9]+]]:_(s32) = G_ZEXT [[MUL2]](s16)
    ; GFX8-NEXT: [[ZEXT3:%[0-9]+]]:_(s32) = G_ZEXT [[MUL3]](s16)
    ; GFX8-NEXT: [[SHL1:%[0-9]+]]:_(s32) = G_SHL [[ZEXT3]], [[C]](s32)
    ; GFX8-NEXT: [[OR1:%[0-9]+]]:_(s32) = G_OR [[ZEXT2]], [[SHL1]]
    ; GFX8-NEXT: [[BITCAST5:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR1]](s32)
    ; GFX8-NEXT: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[BITCAST4]](<2 x s16>), [[BITCAST5]](<2 x s16>)
    ; GFX8-NEXT: $vgpr0_vgpr1 = COPY [[CONCAT_VECTORS]](<4 x s16>)
    ; GFX9-LABEL: name: test_mul_v4s16
    ; GFX9: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
    ; GFX9-NEXT: {{  $}}
    ; GFX9-NEXT: [[COPY:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr0_vgpr1
    ; GFX9-NEXT: [[COPY1:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr2_vgpr3
    ; GFX9-NEXT: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY]](<4 x s16>)
    ; GFX9-NEXT: [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY1]](<4 x s16>)
    ; GFX9-NEXT: [[MUL:%[0-9]+]]:_(<2 x s16>) = G_MUL [[UV]], [[UV2]]
    ; GFX9-NEXT: [[MUL1:%[0-9]+]]:_(<2 x s16>) = G_MUL [[UV1]], [[UV3]]
    ; GFX9-NEXT: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[MUL]](<2 x s16>), [[MUL1]](<2 x s16>)
    ; GFX9-NEXT: $vgpr0_vgpr1 = COPY [[CONCAT_VECTORS]](<4 x s16>)
    ; GFX10-LABEL: name: test_mul_v4s16
    ; GFX10: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
    ; GFX10-NEXT: {{  $}}
    ; GFX10-NEXT: [[COPY:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr0_vgpr1
    ; GFX10-NEXT: [[COPY1:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr2_vgpr3
    ; GFX10-NEXT: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY]](<4 x s16>)
    ; GFX10-NEXT: [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY1]](<4 x s16>)
    ; GFX10-NEXT: [[MUL:%[0-9]+]]:_(<2 x s16>) = G_MUL [[UV]], [[UV2]]
    ; GFX10-NEXT: [[MUL1:%[0-9]+]]:_(<2 x s16>) = G_MUL [[UV1]], [[UV3]]
    ; GFX10-NEXT: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[MUL]](<2 x s16>), [[MUL1]](<2 x s16>)
    ; GFX10-NEXT: $vgpr0_vgpr1 = COPY [[CONCAT_VECTORS]](<4 x s16>)
    %0:_(<4 x s16>) = COPY $vgpr0_vgpr1
    %1:_(<4 x s16>) = COPY $vgpr2_vgpr3
    %2:_(<4 x s16>) = G_MUL %0, %1
    $vgpr0_vgpr1 = COPY %2
...

---
name: test_mul_s24
body: |
  bb.0:
    liveins: $vgpr0, $vgpr1

    ; GFX6-LABEL: name: test_mul_s24
    ; GFX6: liveins: $vgpr0, $vgpr1
    ; GFX6-NEXT: {{  $}}
    ; GFX6-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
    ; GFX6-NEXT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
    ; GFX6-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[COPY]], [[COPY1]]
    ; GFX6-NEXT: $vgpr0 = COPY [[MUL]](s32)
    ; GFX8-LABEL: name: test_mul_s24
    ; GFX8: liveins: $vgpr0, $vgpr1
    ; GFX8-NEXT: {{  $}}
    ; GFX8-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
    ; GFX8-NEXT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
    ; GFX8-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[COPY]], [[COPY1]]
    ; GFX8-NEXT: $vgpr0 = COPY [[MUL]](s32)
    ; GFX9-LABEL: name: test_mul_s24
    ; GFX9: liveins: $vgpr0, $vgpr1
    ; GFX9-NEXT: {{  $}}
    ; GFX9-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
    ; GFX9-NEXT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
    ; GFX9-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[COPY]], [[COPY1]]
    ; GFX9-NEXT: $vgpr0 = COPY [[MUL]](s32)
    ; GFX10-LABEL: name: test_mul_s24
    ; GFX10: liveins: $vgpr0, $vgpr1
    ; GFX10-NEXT: {{  $}}
    ; GFX10-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
    ; GFX10-NEXT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
    ; GFX10-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[COPY]], [[COPY1]]
    ; GFX10-NEXT: $vgpr0 = COPY [[MUL]](s32)
    %0:_(s32) = COPY $vgpr0
    %1:_(s32) = COPY $vgpr1
    %2:_(s24) = G_TRUNC %0
    %3:_(s24) = G_TRUNC %1
    %4:_(s24) = G_MUL %2, %3
    %5:_(s32) = G_ANYEXT %4
    $vgpr0 = COPY %5
...

---
name: test_mul_s33
body: |
  bb.0:
    liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
    ; GFX6-LABEL: name: test_mul_s33
    ; GFX6: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
    ; GFX6-NEXT: {{  $}}
    ; GFX6-NEXT: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1
    ; GFX6-NEXT: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3
    ; GFX6-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](s64)
    ; GFX6-NEXT: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s64)
    ; GFX6-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[UV]], [[UV2]]
    ; GFX6-NEXT: [[MUL1:%[0-9]+]]:_(s32) = G_MUL [[UV1]], [[UV2]]
    ; GFX6-NEXT: [[MUL2:%[0-9]+]]:_(s32) = G_MUL [[UV]], [[UV3]]
    ; GFX6-NEXT: [[UMULH:%[0-9]+]]:_(s32) = G_UMULH [[UV]], [[UV2]]
    ; GFX6-NEXT: [[ADD:%[0-9]+]]:_(s32) = G_ADD [[MUL1]], [[MUL2]]
    ; GFX6-NEXT: [[ADD1:%[0-9]+]]:_(s32) = G_ADD [[ADD]], [[UMULH]]
    ; GFX6-NEXT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[MUL]](s32), [[ADD1]](s32)
    ; GFX6-NEXT: $vgpr0_vgpr1 = COPY [[MV]](s64)
    ; GFX8-LABEL: name: test_mul_s33
    ; GFX8: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
    ; GFX8-NEXT: {{  $}}
    ; GFX8-NEXT: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1
    ; GFX8-NEXT: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3
    ; GFX8-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](s64)
    ; GFX8-NEXT: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s64)
    ; GFX8-NEXT: [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 0
    ; GFX8-NEXT: [[AMDGPU_MAD_U64_U32_:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_1:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV]](s32), [[UV2]], [[C]]
    ; GFX8-NEXT: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_]](s64)
    ; GFX8-NEXT: [[ANYEXT:%[0-9]+]]:_(s64) = G_ANYEXT [[UV5]](s32)
    ; GFX8-NEXT: [[AMDGPU_MAD_U64_U32_2:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_3:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV]](s32), [[UV3]], [[ANYEXT]]
    ; GFX8-NEXT: [[AMDGPU_MAD_U64_U32_4:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_5:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV1]](s32), [[UV2]], [[AMDGPU_MAD_U64_U32_2]]
    ; GFX8-NEXT: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_4]](s64)
    ; GFX8-NEXT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UV4]](s32), [[UV6]](s32)
    ; GFX8-NEXT: $vgpr0_vgpr1 = COPY [[MV]](s64)
    ; GFX9-LABEL: name: test_mul_s33
    ; GFX9: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
    ; GFX9-NEXT: {{  $}}
    ; GFX9-NEXT: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1
    ; GFX9-NEXT: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3
    ; GFX9-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](s64)
    ; GFX9-NEXT: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s64)
    ; GFX9-NEXT: [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 0
    ; GFX9-NEXT: [[AMDGPU_MAD_U64_U32_:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_1:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV]](s32), [[UV2]], [[C]]
    ; GFX9-NEXT: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_]](s64)
    ; GFX9-NEXT: [[ANYEXT:%[0-9]+]]:_(s64) = G_ANYEXT [[UV5]](s32)
    ; GFX9-NEXT: [[AMDGPU_MAD_U64_U32_2:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_3:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV]](s32), [[UV3]], [[ANYEXT]]
    ; GFX9-NEXT: [[AMDGPU_MAD_U64_U32_4:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_5:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV1]](s32), [[UV2]], [[AMDGPU_MAD_U64_U32_2]]
    ; GFX9-NEXT: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_4]](s64)
    ; GFX9-NEXT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UV4]](s32), [[UV6]](s32)
    ; GFX9-NEXT: $vgpr0_vgpr1 = COPY [[MV]](s64)
    ; GFX10-LABEL: name: test_mul_s33
    ; GFX10: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
    ; GFX10-NEXT: {{  $}}
    ; GFX10-NEXT: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1
    ; GFX10-NEXT: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3
    ; GFX10-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](s64)
    ; GFX10-NEXT: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s64)
    ; GFX10-NEXT: [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 0
    ; GFX10-NEXT: [[AMDGPU_MAD_U64_U32_:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_1:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV]](s32), [[UV2]], [[C]]
    ; GFX10-NEXT: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_]](s64)
    ; GFX10-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[UV]], [[UV3]]
    ; GFX10-NEXT: [[ADD:%[0-9]+]]:_(s32) = G_ADD [[UV5]], [[MUL]]
    ; GFX10-NEXT: [[MUL1:%[0-9]+]]:_(s32) = G_MUL [[UV1]], [[UV2]]
    ; GFX10-NEXT: [[ADD1:%[0-9]+]]:_(s32) = G_ADD [[ADD]], [[MUL1]]
    ; GFX10-NEXT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UV4]](s32), [[ADD1]](s32)
    ; GFX10-NEXT: $vgpr0_vgpr1 = COPY [[MV]](s64)
    %0:_(s64) = COPY $vgpr0_vgpr1
    %1:_(s64) = COPY $vgpr2_vgpr3
    %2:_(s33) = G_TRUNC %0
    %3:_(s33) = G_TRUNC %1
    %4:_(s33) = G_MUL %2, %3
    %5:_(s64) = G_ANYEXT %4
    $vgpr0_vgpr1 = COPY %5
...

---
name: test_mul_s96
body: |
  bb.0:
    liveins: $vgpr0_vgpr1_vgpr2, $vgpr3_vgpr4_vgpr5

    ; GFX6-LABEL: name: test_mul_s96
    ; GFX6: liveins: $vgpr0_vgpr1_vgpr2, $vgpr3_vgpr4_vgpr5
    ; GFX6-NEXT: {{  $}}
    ; GFX6-NEXT: [[COPY:%[0-9]+]]:_(s96) = COPY $vgpr0_vgpr1_vgpr2
    ; GFX6-NEXT: [[COPY1:%[0-9]+]]:_(s96) = COPY $vgpr3_vgpr4_vgpr5
    ; GFX6-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](s96)
    ; GFX6-NEXT: [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s96)
    ; GFX6-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[UV]], [[UV3]]
    ; GFX6-NEXT: [[MUL1:%[0-9]+]]:_(s32) = G_MUL [[UV1]], [[UV3]]
    ; GFX6-NEXT: [[MUL2:%[0-9]+]]:_(s32) = G_MUL [[UV]], [[UV4]]
    ; GFX6-NEXT: [[UMULH:%[0-9]+]]:_(s32) = G_UMULH [[UV]], [[UV3]]
    ; GFX6-NEXT: [[UADDO:%[0-9]+]]:_(s32), [[UADDO1:%[0-9]+]]:_(s1) = G_UADDO [[MUL1]], [[MUL2]]
    ; GFX6-NEXT: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO1]](s1)
    ; GFX6-NEXT: [[UADDO2:%[0-9]+]]:_(s32), [[UADDO3:%[0-9]+]]:_(s1) = G_UADDO [[UADDO]], [[UMULH]]
    ; GFX6-NEXT: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO3]](s1)
    ; GFX6-NEXT: [[ADD:%[0-9]+]]:_(s32) = G_ADD [[ZEXT]], [[ZEXT1]]
    ; GFX6-NEXT: [[MUL3:%[0-9]+]]:_(s32) = G_MUL [[UV2]], [[UV3]]
    ; GFX6-NEXT: [[MUL4:%[0-9]+]]:_(s32) = G_MUL [[UV1]], [[UV4]]
    ; GFX6-NEXT: [[MUL5:%[0-9]+]]:_(s32) = G_MUL [[UV]], [[UV5]]
    ; GFX6-NEXT: [[UMULH1:%[0-9]+]]:_(s32) = G_UMULH [[UV1]], [[UV3]]
    ; GFX6-NEXT: [[UMULH2:%[0-9]+]]:_(s32) = G_UMULH [[UV]], [[UV4]]
    ; GFX6-NEXT: [[ADD1:%[0-9]+]]:_(s32) = G_ADD [[MUL3]], [[MUL4]]
    ; GFX6-NEXT: [[ADD2:%[0-9]+]]:_(s32) = G_ADD [[ADD1]], [[MUL5]]
    ; GFX6-NEXT: [[ADD3:%[0-9]+]]:_(s32) = G_ADD [[ADD2]], [[UMULH1]]
    ; GFX6-NEXT: [[ADD4:%[0-9]+]]:_(s32) = G_ADD [[ADD3]], [[UMULH2]]
    ; GFX6-NEXT: [[ADD5:%[0-9]+]]:_(s32) = G_ADD [[ADD4]], [[ADD]]
    ; GFX6-NEXT: [[MV:%[0-9]+]]:_(s96) = G_MERGE_VALUES [[MUL]](s32), [[UADDO2]](s32), [[ADD5]](s32)
    ; GFX6-NEXT: $vgpr0_vgpr1_vgpr2 = COPY [[MV]](s96)
    ; GFX8-LABEL: name: test_mul_s96
    ; GFX8: liveins: $vgpr0_vgpr1_vgpr2, $vgpr3_vgpr4_vgpr5
    ; GFX8-NEXT: {{  $}}
    ; GFX8-NEXT: [[COPY:%[0-9]+]]:_(s96) = COPY $vgpr0_vgpr1_vgpr2
    ; GFX8-NEXT: [[COPY1:%[0-9]+]]:_(s96) = COPY $vgpr3_vgpr4_vgpr5
    ; GFX8-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](s96)
    ; GFX8-NEXT: [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s96)
    ; GFX8-NEXT: [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 0
    ; GFX8-NEXT: [[AMDGPU_MAD_U64_U32_:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_1:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV]](s32), [[UV3]], [[C]]
    ; GFX8-NEXT: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_]](s64)
    ; GFX8-NEXT: [[AMDGPU_MAD_U64_U32_2:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_3:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV]](s32), [[UV5]], [[C]]
    ; GFX8-NEXT: [[AMDGPU_MAD_U64_U32_4:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_5:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV1]](s32), [[UV4]], [[AMDGPU_MAD_U64_U32_2]]
    ; GFX8-NEXT: [[AMDGPU_MAD_U64_U32_6:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_7:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV2]](s32), [[UV3]], [[AMDGPU_MAD_U64_U32_4]]
    ; GFX8-NEXT: [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_6]](s64)
    ; GFX8-NEXT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UV7]](s32), [[UV8]](s32)
    ; GFX8-NEXT: [[AMDGPU_MAD_U64_U32_8:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_9:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV]](s32), [[UV4]], [[MV]]
    ; GFX8-NEXT: [[AMDGPU_MAD_U64_U32_10:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_11:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV1]](s32), [[UV3]], [[AMDGPU_MAD_U64_U32_8]]
    ; GFX8-NEXT: [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_10]](s64)
    ; GFX8-NEXT: [[MV1:%[0-9]+]]:_(s96) = G_MERGE_VALUES [[UV6]](s32), [[UV10]](s32), [[UV11]](s32)
    ; GFX8-NEXT: $vgpr0_vgpr1_vgpr2 = COPY [[MV1]](s96)
    ; GFX9-LABEL: name: test_mul_s96
    ; GFX9: liveins: $vgpr0_vgpr1_vgpr2, $vgpr3_vgpr4_vgpr5
    ; GFX9-NEXT: {{  $}}
    ; GFX9-NEXT: [[COPY:%[0-9]+]]:_(s96) = COPY $vgpr0_vgpr1_vgpr2
    ; GFX9-NEXT: [[COPY1:%[0-9]+]]:_(s96) = COPY $vgpr3_vgpr4_vgpr5
    ; GFX9-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](s96)
    ; GFX9-NEXT: [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s96)
    ; GFX9-NEXT: [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 0
    ; GFX9-NEXT: [[AMDGPU_MAD_U64_U32_:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_1:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV]](s32), [[UV3]], [[C]]
    ; GFX9-NEXT: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_]](s64)
    ; GFX9-NEXT: [[AMDGPU_MAD_U64_U32_2:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_3:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV]](s32), [[UV5]], [[C]]
    ; GFX9-NEXT: [[AMDGPU_MAD_U64_U32_4:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_5:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV1]](s32), [[UV4]], [[AMDGPU_MAD_U64_U32_2]]
    ; GFX9-NEXT: [[AMDGPU_MAD_U64_U32_6:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_7:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV2]](s32), [[UV3]], [[AMDGPU_MAD_U64_U32_4]]
    ; GFX9-NEXT: [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_6]](s64)
    ; GFX9-NEXT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UV7]](s32), [[UV8]](s32)
    ; GFX9-NEXT: [[AMDGPU_MAD_U64_U32_8:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_9:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV]](s32), [[UV4]], [[MV]]
    ; GFX9-NEXT: [[AMDGPU_MAD_U64_U32_10:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_11:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV1]](s32), [[UV3]], [[AMDGPU_MAD_U64_U32_8]]
    ; GFX9-NEXT: [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_10]](s64)
    ; GFX9-NEXT: [[MV1:%[0-9]+]]:_(s96) = G_MERGE_VALUES [[UV6]](s32), [[UV10]](s32), [[UV11]](s32)
    ; GFX9-NEXT: $vgpr0_vgpr1_vgpr2 = COPY [[MV1]](s96)
    ; GFX10-LABEL: name: test_mul_s96
    ; GFX10: liveins: $vgpr0_vgpr1_vgpr2, $vgpr3_vgpr4_vgpr5
    ; GFX10-NEXT: {{  $}}
    ; GFX10-NEXT: [[COPY:%[0-9]+]]:_(s96) = COPY $vgpr0_vgpr1_vgpr2
    ; GFX10-NEXT: [[COPY1:%[0-9]+]]:_(s96) = COPY $vgpr3_vgpr4_vgpr5
    ; GFX10-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](s96)
    ; GFX10-NEXT: [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s96)
    ; GFX10-NEXT: [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 0
    ; GFX10-NEXT: [[AMDGPU_MAD_U64_U32_:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_1:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV]](s32), [[UV3]], [[C]]
    ; GFX10-NEXT: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_]](s64)
    ; GFX10-NEXT: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[UV]], [[UV5]]
    ; GFX10-NEXT: [[MUL1:%[0-9]+]]:_(s32) = G_MUL [[UV1]], [[UV4]]
    ; GFX10-NEXT: [[ADD:%[0-9]+]]:_(s32) = G_ADD [[MUL]], [[MUL1]]
    ; GFX10-NEXT: [[MUL2:%[0-9]+]]:_(s32) = G_MUL [[UV2]], [[UV3]]
    ; GFX10-NEXT: [[ADD1:%[0-9]+]]:_(s32) = G_ADD [[ADD]], [[MUL2]]
    ; GFX10-NEXT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UV7]](s32), [[ADD1]](s32)
    ; GFX10-NEXT: [[AMDGPU_MAD_U64_U32_2:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_3:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV]](s32), [[UV4]], [[MV]]
    ; GFX10-NEXT: [[AMDGPU_MAD_U64_U32_4:%[0-9]+]]:_(s64), [[AMDGPU_MAD_U64_U32_5:%[0-9]+]]:_(s1) = G_AMDGPU_MAD_U64_U32 [[UV1]](s32), [[UV3]], [[AMDGPU_MAD_U64_U32_2]]
    ; GFX10-NEXT: [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_MAD_U64_U32_4]](s64)
    ; GFX10-NEXT: [[MV1:%[0-9]+]]:_(s96) = G_MERGE_VALUES [[UV6]](s32), [[UV8]](s32), [[UV9]](s32)
    ; GFX10-NEXT: $vgpr0_vgpr1_vgpr2 = COPY [[MV1]](s96)
    %0:_(s96) = COPY $vgpr0_vgpr1_vgpr2
    %1:_(s96) = COPY $vgpr3_vgpr4_vgpr5
    %2:_(s96) = G_MUL %0, %1
    $vgpr0_vgpr1_vgpr2 = COPY %2
...