# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
# RUN: llc -mtriple=amdgpu9.0a -run-pass='amdgpu-reg-bank-select,amdgpu-reg-bank-legalize' -amdgpu-mfma-vgpr-form=0 %s -o - | FileCheck %s

---
name: mfma_f32_32x32x4bf16_1k_vva
legalized: true
tracksRegLiveness: true
body: |
  bb.0:
    liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $agpr0_agpr1_agpr2_agpr3_agpr4_agpr5_agpr6_agpr7_agpr8_agpr9_agpr10_agpr11_agpr12_agpr13_agpr14_agpr15_agpr16_agpr17_agpr18_agpr19_agpr20_agpr21_agpr22_agpr23_agpr24_agpr25_agpr26_agpr27_agpr28_agpr29_agpr30_agpr31

    ; CHECK-LABEL: name: mfma_f32_32x32x4bf16_1k_vva
    ; CHECK: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $agpr0_agpr1_agpr2_agpr3_agpr4_agpr5_agpr6_agpr7_agpr8_agpr9_agpr10_agpr11_agpr12_agpr13_agpr14_agpr15_agpr16_agpr17_agpr18_agpr19_agpr20_agpr21_agpr22_agpr23_agpr24_agpr25_agpr26_agpr27_agpr28_agpr29_agpr30_agpr31
    ; CHECK-NEXT: {{  $}}
    ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr(s64) = COPY $vgpr0_vgpr1
    ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr(s64) = COPY $vgpr2_vgpr3
    ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr(<32 x s32>) = COPY $agpr0_agpr1_agpr2_agpr3_agpr4_agpr5_agpr6_agpr7_agpr8_agpr9_agpr10_agpr11_agpr12_agpr13_agpr14_agpr15_agpr16_agpr17_agpr18_agpr19_agpr20_agpr21_agpr22_agpr23_agpr24_agpr25_agpr26_agpr27_agpr28_agpr29_agpr30_agpr31
    ; CHECK-NEXT: [[COPY3:%[0-9]+]]:agpr(<32 x s32>) = COPY [[COPY2]](<32 x s32>)
    ; CHECK-NEXT: [[INTRINSIC_CONVERGENT:%[0-9]+]]:agpr(<32 x s32>) = G_INTRINSIC_CONVERGENT intrinsic(@llvm.amdgcn.mfma.f32.32x32x4bf16.1k), [[COPY]](s64), [[COPY1]](s64), [[COPY3]](<32 x s32>), 0, 0, 0
    ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr(<32 x s32>) = COPY [[INTRINSIC_CONVERGENT]](<32 x s32>)
    ; CHECK-NEXT: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15_vgpr16_vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31 = COPY [[COPY4]](<32 x s32>)
    %0:_(s64) = COPY $vgpr0_vgpr1
    %1:_(s64) = COPY $vgpr2_vgpr3
    %2:_(<32 x s32>) = COPY $agpr0_agpr1_agpr2_agpr3_agpr4_agpr5_agpr6_agpr7_agpr8_agpr9_agpr10_agpr11_agpr12_agpr13_agpr14_agpr15_agpr16_agpr17_agpr18_agpr19_agpr20_agpr21_agpr22_agpr23_agpr24_agpr25_agpr26_agpr27_agpr28_agpr29_agpr30_agpr31
    %3:_(<32 x s32>) = G_INTRINSIC_CONVERGENT intrinsic(@llvm.amdgcn.mfma.f32.32x32x4bf16.1k), %0, %1, %2, 0, 0, 0
    $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15_vgpr16_vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31 = COPY %3
...

---
name: mfma_f32_16x16x4bf16_1k_vva
legalized: true
tracksRegLiveness: true
body: |
  bb.0:
    liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $agpr0_agpr1_agpr2_agpr3_agpr4_agpr5_agpr6_agpr7_agpr8_agpr9_agpr10_agpr11_agpr12_agpr13_agpr14_agpr15

    ; CHECK-LABEL: name: mfma_f32_16x16x4bf16_1k_vva
    ; CHECK: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $agpr0_agpr1_agpr2_agpr3_agpr4_agpr5_agpr6_agpr7_agpr8_agpr9_agpr10_agpr11_agpr12_agpr13_agpr14_agpr15
    ; CHECK-NEXT: {{  $}}
    ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr(s64) = COPY $vgpr0_vgpr1
    ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr(s64) = COPY $vgpr2_vgpr3
    ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr(<16 x s32>) = COPY $agpr0_agpr1_agpr2_agpr3_agpr4_agpr5_agpr6_agpr7_agpr8_agpr9_agpr10_agpr11_agpr12_agpr13_agpr14_agpr15
    ; CHECK-NEXT: [[COPY3:%[0-9]+]]:agpr(<16 x s32>) = COPY [[COPY2]](<16 x s32>)
    ; CHECK-NEXT: [[INTRINSIC_CONVERGENT:%[0-9]+]]:agpr(<16 x s32>) = G_INTRINSIC_CONVERGENT intrinsic(@llvm.amdgcn.mfma.f32.16x16x4bf16.1k), [[COPY]](s64), [[COPY1]](s64), [[COPY3]](<16 x s32>), 0, 0, 0
    ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr(<16 x s32>) = COPY [[INTRINSIC_CONVERGENT]](<16 x s32>)
    ; CHECK-NEXT: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15 = COPY [[COPY4]](<16 x s32>)
    %0:_(s64) = COPY $vgpr0_vgpr1
    %1:_(s64) = COPY $vgpr2_vgpr3
    %2:_(<16 x s32>) = COPY $agpr0_agpr1_agpr2_agpr3_agpr4_agpr5_agpr6_agpr7_agpr8_agpr9_agpr10_agpr11_agpr12_agpr13_agpr14_agpr15
    %3:_(<16 x s32>) = G_INTRINSIC_CONVERGENT intrinsic(@llvm.amdgcn.mfma.f32.16x16x4bf16.1k), %0, %1, %2, 0, 0, 0
    $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15 = COPY %3
...

---
name: mfma_f32_4x4x4bf16_1k_vva
legalized: true
tracksRegLiveness: true
body: |
  bb.0:
    liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $agpr0_agpr1_agpr2_agpr3

    ; CHECK-LABEL: name: mfma_f32_4x4x4bf16_1k_vva
    ; CHECK: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $agpr0_agpr1_agpr2_agpr3
    ; CHECK-NEXT: {{  $}}
    ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr(s64) = COPY $vgpr0_vgpr1
    ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr(s64) = COPY $vgpr2_vgpr3
    ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr(<4 x s32>) = COPY $agpr0_agpr1_agpr2_agpr3
    ; CHECK-NEXT: [[COPY3:%[0-9]+]]:agpr(<4 x s32>) = COPY [[COPY2]](<4 x s32>)
    ; CHECK-NEXT: [[INTRINSIC_CONVERGENT:%[0-9]+]]:agpr(<4 x s32>) = G_INTRINSIC_CONVERGENT intrinsic(@llvm.amdgcn.mfma.f32.4x4x4bf16.1k), [[COPY]](s64), [[COPY1]](s64), [[COPY3]](<4 x s32>), 0, 0, 0
    ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr(<4 x s32>) = COPY [[INTRINSIC_CONVERGENT]](<4 x s32>)
    ; CHECK-NEXT: $vgpr0_vgpr1_vgpr2_vgpr3 = COPY [[COPY4]](<4 x s32>)
    %0:_(s64) = COPY $vgpr0_vgpr1
    %1:_(s64) = COPY $vgpr2_vgpr3
    %2:_(<4 x s32>) = COPY $agpr0_agpr1_agpr2_agpr3
    %3:_(<4 x s32>) = G_INTRINSIC_CONVERGENT intrinsic(@llvm.amdgcn.mfma.f32.4x4x4bf16.1k), %0, %1, %2, 0, 0, 0
    $vgpr0_vgpr1_vgpr2_vgpr3 = COPY %3
...

---
name: mfma_f32_32x32x8bf16_1k_vva
legalized: true
tracksRegLiveness: true
body: |
  bb.0:
    liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $agpr0_agpr1_agpr2_agpr3_agpr4_agpr5_agpr6_agpr7_agpr8_agpr9_agpr10_agpr11_agpr12_agpr13_agpr14_agpr15_agpr16_agpr17_agpr18_agpr19_agpr20_agpr21_agpr22_agpr23_agpr24_agpr25_agpr26_agpr27_agpr28_agpr29_agpr30_agpr31

    ; CHECK-LABEL: name: mfma_f32_32x32x8bf16_1k_vva
    ; CHECK: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $agpr0_agpr1_agpr2_agpr3_agpr4_agpr5_agpr6_agpr7_agpr8_agpr9_agpr10_agpr11_agpr12_agpr13_agpr14_agpr15_agpr16_agpr17_agpr18_agpr19_agpr20_agpr21_agpr22_agpr23_agpr24_agpr25_agpr26_agpr27_agpr28_agpr29_agpr30_agpr31
    ; CHECK-NEXT: {{  $}}
    ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr(s64) = COPY $vgpr0_vgpr1
    ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr(s64) = COPY $vgpr2_vgpr3
    ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr(<32 x s32>) = COPY $agpr0_agpr1_agpr2_agpr3_agpr4_agpr5_agpr6_agpr7_agpr8_agpr9_agpr10_agpr11_agpr12_agpr13_agpr14_agpr15_agpr16_agpr17_agpr18_agpr19_agpr20_agpr21_agpr22_agpr23_agpr24_agpr25_agpr26_agpr27_agpr28_agpr29_agpr30_agpr31
    ; CHECK-NEXT: [[COPY3:%[0-9]+]]:agpr(<32 x s32>) = COPY [[COPY2]](<32 x s32>)
    ; CHECK-NEXT: [[INTRINSIC_CONVERGENT:%[0-9]+]]:agpr(<32 x s32>) = G_INTRINSIC_CONVERGENT intrinsic(@llvm.amdgcn.mfma.f32.32x32x8bf16.1k), [[COPY]](s64), [[COPY1]](s64), [[COPY3]](<32 x s32>), 0, 0, 0
    ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr(<32 x s32>) = COPY [[INTRINSIC_CONVERGENT]](<32 x s32>)
    ; CHECK-NEXT: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15_vgpr16_vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31 = COPY [[COPY4]](<32 x s32>)
    %0:_(s64) = COPY $vgpr0_vgpr1
    %1:_(s64) = COPY $vgpr2_vgpr3
    %2:_(<32 x s32>) = COPY $agpr0_agpr1_agpr2_agpr3_agpr4_agpr5_agpr6_agpr7_agpr8_agpr9_agpr10_agpr11_agpr12_agpr13_agpr14_agpr15_agpr16_agpr17_agpr18_agpr19_agpr20_agpr21_agpr22_agpr23_agpr24_agpr25_agpr26_agpr27_agpr28_agpr29_agpr30_agpr31
    %3:_(<32 x s32>) = G_INTRINSIC_CONVERGENT intrinsic(@llvm.amdgcn.mfma.f32.32x32x8bf16.1k), %0, %1, %2, 0, 0, 0
    $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15_vgpr16_vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31 = COPY %3
...

---
name: mfma_f32_16x16x16bf16_1k_vva
legalized: true
tracksRegLiveness: true
body: |
  bb.0:
    liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $agpr0_agpr1_agpr2_agpr3

    ; CHECK-LABEL: name: mfma_f32_16x16x16bf16_1k_vva
    ; CHECK: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $agpr0_agpr1_agpr2_agpr3
    ; CHECK-NEXT: {{  $}}
    ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr(s64) = COPY $vgpr0_vgpr1
    ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr(s64) = COPY $vgpr2_vgpr3
    ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr(<4 x s32>) = COPY $agpr0_agpr1_agpr2_agpr3
    ; CHECK-NEXT: [[COPY3:%[0-9]+]]:agpr(<4 x s32>) = COPY [[COPY2]](<4 x s32>)
    ; CHECK-NEXT: [[INTRINSIC_CONVERGENT:%[0-9]+]]:agpr(<4 x s32>) = G_INTRINSIC_CONVERGENT intrinsic(@llvm.amdgcn.mfma.f32.16x16x16bf16.1k), [[COPY]](s64), [[COPY1]](s64), [[COPY3]](<4 x s32>), 0, 0, 0
    ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr(<4 x s32>) = COPY [[INTRINSIC_CONVERGENT]](<4 x s32>)
    ; CHECK-NEXT: $vgpr0_vgpr1_vgpr2_vgpr3 = COPY [[COPY4]](<4 x s32>)
    %0:_(s64) = COPY $vgpr0_vgpr1
    %1:_(s64) = COPY $vgpr2_vgpr3
    %2:_(<4 x s32>) = COPY $agpr0_agpr1_agpr2_agpr3
    %3:_(<4 x s32>) = G_INTRINSIC_CONVERGENT intrinsic(@llvm.amdgcn.mfma.f32.16x16x16bf16.1k), %0, %1, %2, 0, 0, 0
    $vgpr0_vgpr1_vgpr2_vgpr3 = COPY %3
...

---
name: mfma_f64_16x16x4f64_vva
legalized: true
tracksRegLiveness: true
body: |
  bb.0:
    liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $agpr0_agpr1_agpr2_agpr3_agpr4_agpr5_agpr6_agpr7

    ; CHECK-LABEL: name: mfma_f64_16x16x4f64_vva
    ; CHECK: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $agpr0_agpr1_agpr2_agpr3_agpr4_agpr5_agpr6_agpr7
    ; CHECK-NEXT: {{  $}}
    ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr(s64) = COPY $vgpr0_vgpr1
    ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr(s64) = COPY $vgpr2_vgpr3
    ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr(<8 x s32>) = COPY $agpr0_agpr1_agpr2_agpr3_agpr4_agpr5_agpr6_agpr7
    ; CHECK-NEXT: [[COPY3:%[0-9]+]]:agpr(<8 x s32>) = COPY [[COPY2]](<8 x s32>)
    ; CHECK-NEXT: [[INTRINSIC_CONVERGENT:%[0-9]+]]:agpr(<8 x s32>) = G_INTRINSIC_CONVERGENT intrinsic(@llvm.amdgcn.mfma.f64.16x16x4f64), [[COPY]](s64), [[COPY1]](s64), [[COPY3]](<8 x s32>), 0, 0, 0
    ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr(<8 x s32>) = COPY [[INTRINSIC_CONVERGENT]](<8 x s32>)
    ; CHECK-NEXT: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 = COPY [[COPY4]](<8 x s32>)
    %0:_(s64) = COPY $vgpr0_vgpr1
    %1:_(s64) = COPY $vgpr2_vgpr3
    %2:_(<8 x s32>) = COPY $agpr0_agpr1_agpr2_agpr3_agpr4_agpr5_agpr6_agpr7
    %3:_(<8 x s32>) = G_INTRINSIC_CONVERGENT intrinsic(@llvm.amdgcn.mfma.f64.16x16x4f64), %0, %1, %2, 0, 0, 0
    $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 = COPY %3
...

---
name: mfma_f64_4x4x4f64_vva
legalized: true
tracksRegLiveness: true
body: |
  bb.0:
    liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $agpr0_agpr1

    ; CHECK-LABEL: name: mfma_f64_4x4x4f64_vva
    ; CHECK: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $agpr0_agpr1
    ; CHECK-NEXT: {{  $}}
    ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr(s64) = COPY $vgpr0_vgpr1
    ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr(s64) = COPY $vgpr2_vgpr3
    ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr(<2 x s32>) = COPY $agpr0_agpr1
    ; CHECK-NEXT: [[COPY3:%[0-9]+]]:agpr(<2 x s32>) = COPY [[COPY2]](<2 x s32>)
    ; CHECK-NEXT: [[INTRINSIC_CONVERGENT:%[0-9]+]]:agpr(<2 x s32>) = G_INTRINSIC_CONVERGENT intrinsic(@llvm.amdgcn.mfma.f64.4x4x4f64), [[COPY]](s64), [[COPY1]](s64), [[COPY3]](<2 x s32>), 0, 0, 0
    ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr(<2 x s32>) = COPY [[INTRINSIC_CONVERGENT]](<2 x s32>)
    ; CHECK-NEXT: $vgpr0_vgpr1 = COPY [[COPY4]](<2 x s32>)
    %0:_(s64) = COPY $vgpr0_vgpr1
    %1:_(s64) = COPY $vgpr2_vgpr3
    %2:_(<2 x s32>) = COPY $agpr0_agpr1
    %3:_(<2 x s32>) = G_INTRINSIC_CONVERGENT intrinsic(@llvm.amdgcn.mfma.f64.4x4x4f64), %0, %1, %2, 0, 0, 0
    $vgpr0_vgpr1 = COPY %3
...