llvm/llvm/test/CodeGen/AMDGPU/vgpr-remat.mir

# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
# RUN: llc -mtriple=amdgcn -mcpu=gfx900 -run-pass=register-coalescer -o - %s | FileCheck %s

# Check that we get two move-immediates into %1 and %2, instead of a copy from
# %1 to %2, because that would introduce a dependency and maybe a stall.
---
name: f
tracksRegLiveness: true
body: |
  ; CHECK-LABEL: name: f
  ; CHECK: bb.0:
  ; CHECK-NEXT:   successors: %bb.2(0x40000000), %bb.1(0x40000000)
  ; CHECK-NEXT:   liveins: $sgpr0
  ; CHECK-NEXT: {{  $}}
  ; CHECK-NEXT:   undef [[V_MOV_B32_e32_:%[0-9]+]].sub0:vreg_96 = V_MOV_B32_e32 0, implicit $exec
  ; CHECK-NEXT:   [[V_MOV_B32_e32_:%[0-9]+]].sub1:vreg_96 = V_MOV_B32_e32 0, implicit $exec
  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0
  ; CHECK-NEXT:   $exec = S_MOV_B64_term [[COPY]]
  ; CHECK-NEXT:   S_CBRANCH_EXECZ %bb.2, implicit $exec
  ; CHECK-NEXT:   S_BRANCH %bb.1
  ; CHECK-NEXT: {{  $}}
  ; CHECK-NEXT: bb.1:
  ; CHECK-NEXT:   successors: %bb.2(0x80000000)
  ; CHECK-NEXT: {{  $}}
  ; CHECK-NEXT:   [[V_MOV_B32_e32_:%[0-9]+]].sub0:vreg_96 = V_MUL_F32_e32 [[V_MOV_B32_e32_]].sub0, [[V_MOV_B32_e32_]].sub0, implicit $mode, implicit $exec
  ; CHECK-NEXT:   [[V_MOV_B32_e32_:%[0-9]+]].sub1:vreg_96 = V_MUL_F32_e32 [[V_MOV_B32_e32_]].sub1, [[V_MOV_B32_e32_]].sub1, implicit $mode, implicit $exec
  ; CHECK-NEXT: {{  $}}
  ; CHECK-NEXT: bb.2:
  ; CHECK-NEXT:   S_ENDPGM 0, implicit [[V_MOV_B32_e32_]]
  bb.0:
    liveins: $sgpr0
    %0:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
    %1:vgpr_32 = COPY %0:vgpr_32
    %2:vgpr_32 = COPY %0:vgpr_32
    %3:sreg_64 = COPY $sgpr0
    $exec = S_MOV_B64_term %3:sreg_64
    S_CBRANCH_EXECZ %bb.2, implicit $exec
    S_BRANCH %bb.1

  bb.1:
    %1:vgpr_32 = V_MUL_F32_e32 %1:vgpr_32, %1:vgpr_32, implicit $mode, implicit $exec
    %2:vgpr_32 = V_MUL_F32_e32 %2:vgpr_32, %2:vgpr_32, implicit $mode, implicit $exec

  bb.2:
    undef %4.sub0:vreg_96 = COPY %1:vgpr_32
    %4.sub1:vreg_96 = COPY %2:vgpr_32
    S_ENDPGM 0, implicit %4
...