AMDGPU: Add baseline test for treating v_pk_mov_b32 like reg_sequence #125656

arsenm · 2025-02-04T09:27:15Z

No description provided.

arsenm · 2025-02-04T09:27:36Z

WIP: AMDGPU: Implement getRegSequenceLikeInputs for v_pk_mov_b32 #125657
AMDGPU: Add baseline test for treating v_pk_mov_b32 like reg_sequence #125656 👈 (View in Graphite)
AMDGPU: Use default shouldRewriteCopySrc #125535 : 1 other dependent PR (#125667 )
PeepholeOpt: Fix looking for def of current copy to coalesce #125533
X86: Remove hack in shouldRewriteCopySrc for subregister handling #125224
main

This stack of pull requests is managed by Graphite. Learn more about stacking.

llvmbot · 2025-02-04T09:28:23Z

@llvm/pr-subscribers-backend-amdgpu

Author: Matt Arsenault (arsenm)

Changes

Patch is 20.10 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/125656.diff

1 Files Affected:

(added) llvm/test/CodeGen/AMDGPU/reg-sequence-like-v-pk-mov-b32.mir (+413)

diff --git a/llvm/test/CodeGen/AMDGPU/reg-sequence-like-v-pk-mov-b32.mir b/llvm/test/CodeGen/AMDGPU/reg-sequence-like-v-pk-mov-b32.mir
new file mode 100644
index 00000000000000..90291221e8e178
--- /dev/null
+++ b/llvm/test/CodeGen/AMDGPU/reg-sequence-like-v-pk-mov-b32.mir
@@ -0,0 +1,413 @@
+# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py UTC_ARGS: --version 5
+# RUN: llc -mtriple=amdgcn -mcpu=gfx90a -run-pass=peephole-opt -o - %s | FileCheck %s
+
+---
+name: v_pk_mov_b32__reg_sequence_shuffle_vgpr_vgpr_1_0
+tracksRegLiveness: true
+body:             |
+  bb.0:
+    liveins: $vgpr0, $vgpr1
+
+    ; CHECK-LABEL: name: v_pk_mov_b32__reg_sequence_shuffle_vgpr_vgpr_1_0
+    ; CHECK: liveins: $vgpr0, $vgpr1
+    ; CHECK-NEXT: {{  $}}
+    ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+    ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+    ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64_align2 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1
+    ; CHECK-NEXT: [[V_PK_MOV_B32_:%[0-9]+]]:vreg_64_align2 = V_PK_MOV_B32 12, [[REG_SEQUENCE]], 8, [[REG_SEQUENCE]], 0, 0, 0, 0, 0, implicit $exec
+    ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub1
+    ; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub0
+    ; CHECK-NEXT: $vgpr4 = COPY [[COPY2]]
+    ; CHECK-NEXT: $vgpr5 = COPY [[COPY3]]
+    ; CHECK-NEXT: S_ENDPGM 0, implicit $vgpr4, implicit $vgpr5
+    %0:vgpr_32 = COPY $vgpr0
+    %1:vgpr_32 = COPY $vgpr1
+    %2:vreg_64_align2 = REG_SEQUENCE %0, %subreg.sub0, %1, %subreg.sub1
+    %3:vreg_64_align2 = V_PK_MOV_B32 12, %2, 8, %2, 0, 0, 0, 0, 0, implicit $exec
+    %4:vgpr_32 = COPY %3.sub1
+    %5:vgpr_32 = COPY %3.sub0
+    $vgpr4 = COPY %4
+    $vgpr5 = COPY %5
+    S_ENDPGM 0, implicit $vgpr4, implicit $vgpr5
+
+...
+
+---
+name: v_pk_mov_b32__reg_sequence_shuffle_vgpr_vgpr_3_0
+tracksRegLiveness: true
+body:             |
+  bb.0:
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+
+    ; CHECK-LABEL: name: v_pk_mov_b32__reg_sequence_shuffle_vgpr_vgpr_3_0
+    ; CHECK: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+    ; CHECK-NEXT: {{  $}}
+    ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+    ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+    ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+    ; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+    ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64_align2 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1
+    ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64_align2 = REG_SEQUENCE [[COPY2]], %subreg.sub0, [[COPY3]], %subreg.sub1
+    ; CHECK-NEXT: [[V_PK_MOV_B32_:%[0-9]+]]:vreg_64_align2 = V_PK_MOV_B32 12, [[REG_SEQUENCE]], 8, [[REG_SEQUENCE1]], 0, 0, 0, 0, 0, implicit $exec
+    ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub1
+    ; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub0
+    ; CHECK-NEXT: $vgpr4 = COPY [[COPY4]]
+    ; CHECK-NEXT: $vgpr5 = COPY [[COPY5]]
+    ; CHECK-NEXT: S_ENDPGM 0, implicit $vgpr4, implicit $vgpr5
+    %0:vgpr_32 = COPY $vgpr0
+    %1:vgpr_32 = COPY $vgpr1
+    %2:vgpr_32 = COPY $vgpr2
+    %3:vgpr_32 = COPY $vgpr3
+    %4:vreg_64_align2 = REG_SEQUENCE %0, %subreg.sub0, %1, %subreg.sub1
+    %5:vreg_64_align2 = REG_SEQUENCE %2, %subreg.sub0, %3, %subreg.sub1
+    %6:vreg_64_align2 = V_PK_MOV_B32 12, %4, 8, %5, 0, 0, 0, 0, 0, implicit $exec
+    %7:vgpr_32 = COPY %6.sub1
+    %8:vgpr_32 = COPY %6.sub0
+    $vgpr4 = COPY %7
+    $vgpr5 = COPY %8
+    S_ENDPGM 0, implicit $vgpr4, implicit $vgpr5
+
+...
+
+---
+name: v_pk_mov_b32__reg_sequence_shuffle_sgpr_sgpr_1_0
+tracksRegLiveness: true
+body:             |
+  bb.0:
+    liveins: $sgpr8, $sgpr9
+
+    ; CHECK-LABEL: name: v_pk_mov_b32__reg_sequence_shuffle_sgpr_sgpr_1_0
+    ; CHECK: liveins: $sgpr8, $sgpr9
+    ; CHECK-NEXT: {{  $}}
+    ; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr8
+    ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr9
+    ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sreg_64 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1
+    ; CHECK-NEXT: [[V_PK_MOV_B32_:%[0-9]+]]:vreg_64_align2 = V_PK_MOV_B32 12, [[REG_SEQUENCE]], 8, [[REG_SEQUENCE]], 0, 0, 0, 0, 0, implicit $exec
+    ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub1
+    ; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub0
+    ; CHECK-NEXT: $vgpr4 = COPY [[COPY2]]
+    ; CHECK-NEXT: $vgpr5 = COPY [[COPY3]]
+    ; CHECK-NEXT: S_ENDPGM 0, implicit $vgpr4, implicit $vgpr5
+    %0:sreg_32 = COPY $sgpr8
+    %1:sreg_32 = COPY $sgpr9
+    %2:sreg_64 = REG_SEQUENCE %0, %subreg.sub0, %1, %subreg.sub1
+    %3:vreg_64_align2 = V_PK_MOV_B32 12, %2, 8, %2, 0, 0, 0, 0, 0, implicit $exec
+    %4:vgpr_32 = COPY %3.sub1
+    %5:vgpr_32 = COPY %3.sub0
+    $vgpr4 = COPY %4
+    $vgpr5 = COPY %5
+    S_ENDPGM 0, implicit $vgpr4, implicit $vgpr5
+
+...
+
+---
+name: v_pk_mov_b32__reg_sequence_shuffle_vgpr_sgpr_3_0
+tracksRegLiveness: true
+body:             |
+  bb.0:
+    liveins: $vgpr8, $vgpr9, $sgpr10, $sgpr11
+
+    ; CHECK-LABEL: name: v_pk_mov_b32__reg_sequence_shuffle_vgpr_sgpr_3_0
+    ; CHECK: liveins: $vgpr8, $vgpr9, $sgpr10, $sgpr11
+    ; CHECK-NEXT: {{  $}}
+    ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr8
+    ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr9
+    ; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_64 = COPY $sgpr10
+    ; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_64 = COPY $sgpr11
+    ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64_align2 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1
+    ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sreg_64 = REG_SEQUENCE [[COPY2]], %subreg.sub0, [[COPY3]], %subreg.sub1
+    ; CHECK-NEXT: [[V_PK_MOV_B32_:%[0-9]+]]:vreg_64_align2 = V_PK_MOV_B32 12, [[REG_SEQUENCE]], 8, [[REG_SEQUENCE1]], 0, 0, 0, 0, 0, implicit $exec
+    ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub1
+    ; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub0
+    ; CHECK-NEXT: $vgpr4 = COPY [[COPY4]]
+    ; CHECK-NEXT: $vgpr5 = COPY [[COPY5]]
+    ; CHECK-NEXT: S_ENDPGM 0, implicit $vgpr4, implicit $vgpr5
+    %0:vgpr_32 = COPY $vgpr8
+    %1:vgpr_32 = COPY $vgpr9
+    %2:sreg_64 = COPY $sgpr10
+    %3:sreg_64 = COPY $sgpr11
+    %4:vreg_64_align2 = REG_SEQUENCE %0, %subreg.sub0, %1, %subreg.sub1
+    %5:sreg_64 = REG_SEQUENCE %2, %subreg.sub0, %3, %subreg.sub1
+    %6:vreg_64_align2 = V_PK_MOV_B32 12, %4, 8, %5, 0, 0, 0, 0, 0, implicit $exec
+    %7:vgpr_32 = COPY %6.sub1
+    %8:vgpr_32 = COPY %6.sub0
+    $vgpr4 = COPY %7
+    $vgpr5 = COPY %8
+    S_ENDPGM 0, implicit $vgpr4, implicit $vgpr5
+
+...
+
+---
+name: v_pk_mov_b32__reg_sequence_shuffle_sgpr_vgpr_3_0
+tracksRegLiveness: true
+body:             |
+  bb.0:
+    liveins: $sgpr8, $sgpr9, $vgpr10, $vgpr11
+
+    ; CHECK-LABEL: name: v_pk_mov_b32__reg_sequence_shuffle_sgpr_vgpr_3_0
+    ; CHECK: liveins: $sgpr8, $sgpr9, $vgpr10, $vgpr11
+    ; CHECK-NEXT: {{  $}}
+    ; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr8
+    ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr9
+    ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr10
+    ; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr11
+    ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sreg_64 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1
+    ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64_align2 = REG_SEQUENCE [[COPY2]], %subreg.sub0, [[COPY3]], %subreg.sub1
+    ; CHECK-NEXT: [[V_PK_MOV_B32_:%[0-9]+]]:vreg_64_align2 = V_PK_MOV_B32 12, [[REG_SEQUENCE]], 8, [[REG_SEQUENCE1]], 0, 0, 0, 0, 0, implicit $exec
+    ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub1
+    ; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub0
+    ; CHECK-NEXT: $vgpr4 = COPY [[COPY4]]
+    ; CHECK-NEXT: $vgpr5 = COPY [[COPY5]]
+    ; CHECK-NEXT: S_ENDPGM 0, implicit $vgpr4, implicit $vgpr5
+    %0:sreg_32 = COPY $sgpr8
+    %1:sreg_32 = COPY $sgpr9
+    %2:vgpr_32 = COPY $vgpr10
+    %3:vgpr_32 = COPY $vgpr11
+    %4:sreg_64 = REG_SEQUENCE %0, %subreg.sub0, %1, %subreg.sub1
+    %5:vreg_64_align2 = REG_SEQUENCE %2, %subreg.sub0, %3, %subreg.sub1
+    %6:vreg_64_align2 = V_PK_MOV_B32 12, %4, 8, %5, 0, 0, 0, 0, 0, implicit $exec
+    %7:vgpr_32 = COPY %6.sub1
+    %8:vgpr_32 = COPY %6.sub0
+    $vgpr4 = COPY %7
+    $vgpr5 = COPY %8
+    S_ENDPGM 0, implicit $vgpr4, implicit $vgpr5
+
+...
+
+---
+name: v_pk_mov_b32__reg_sequence_shuffle_vgpr_vgpr_3_0_undef_lhs
+tracksRegLiveness: true
+body:             |
+  bb.0:
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+
+    ; CHECK-LABEL: name: v_pk_mov_b32__reg_sequence_shuffle_vgpr_vgpr_3_0_undef_lhs
+    ; CHECK: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+    ; CHECK-NEXT: {{  $}}
+    ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+    ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+    ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+    ; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+    ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64_align2 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1
+    ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64_align2 = REG_SEQUENCE [[COPY2]], %subreg.sub0, [[COPY3]], %subreg.sub1
+    ; CHECK-NEXT: [[V_PK_MOV_B32_:%[0-9]+]]:vreg_64_align2 = V_PK_MOV_B32 12, undef [[REG_SEQUENCE]], 8, [[REG_SEQUENCE1]], 0, 0, 0, 0, 0, implicit $exec
+    ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub1
+    ; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub0
+    ; CHECK-NEXT: $vgpr4 = COPY [[COPY4]]
+    ; CHECK-NEXT: $vgpr5 = COPY [[COPY5]]
+    ; CHECK-NEXT: S_ENDPGM 0, implicit $vgpr4, implicit $vgpr5
+    %0:vgpr_32 = COPY $vgpr0
+    %1:vgpr_32 = COPY $vgpr1
+    %2:vgpr_32 = COPY $vgpr2
+    %3:vgpr_32 = COPY $vgpr3
+    %4:vreg_64_align2 = REG_SEQUENCE %0, %subreg.sub0, %1, %subreg.sub1
+    %5:vreg_64_align2 = REG_SEQUENCE %2, %subreg.sub0, %3, %subreg.sub1
+    %6:vreg_64_align2 = V_PK_MOV_B32 12, undef %4, 8, %5, 0, 0, 0, 0, 0, implicit $exec
+    %7:vgpr_32 = COPY %6.sub1
+    %8:vgpr_32 = COPY %6.sub0
+    $vgpr4 = COPY %7
+    $vgpr5 = COPY %8
+    S_ENDPGM 0, implicit $vgpr4, implicit $vgpr5
+
+...
+
+---
+name: v_pk_mov_b32__reg_sequence_shuffle_vgpr_vgpr_3_0_undef_rhs
+tracksRegLiveness: true
+body:             |
+  bb.0:
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+
+    ; CHECK-LABEL: name: v_pk_mov_b32__reg_sequence_shuffle_vgpr_vgpr_3_0_undef_rhs
+    ; CHECK: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+    ; CHECK-NEXT: {{  $}}
+    ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+    ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+    ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+    ; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+    ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64_align2 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1
+    ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64_align2 = REG_SEQUENCE [[COPY2]], %subreg.sub0, [[COPY3]], %subreg.sub1
+    ; CHECK-NEXT: [[V_PK_MOV_B32_:%[0-9]+]]:vreg_64_align2 = V_PK_MOV_B32 12, [[REG_SEQUENCE]], 8, undef [[REG_SEQUENCE1]], 0, 0, 0, 0, 0, implicit $exec
+    ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub1
+    ; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub0
+    ; CHECK-NEXT: $vgpr4 = COPY [[COPY4]]
+    ; CHECK-NEXT: $vgpr5 = COPY [[COPY5]]
+    ; CHECK-NEXT: S_ENDPGM 0, implicit $vgpr4, implicit $vgpr5
+    %0:vgpr_32 = COPY $vgpr0
+    %1:vgpr_32 = COPY $vgpr1
+    %2:vgpr_32 = COPY $vgpr2
+    %3:vgpr_32 = COPY $vgpr3
+    %4:vreg_64_align2 = REG_SEQUENCE %0, %subreg.sub0, %1, %subreg.sub1
+    %5:vreg_64_align2 = REG_SEQUENCE %2, %subreg.sub0, %3, %subreg.sub1
+    %6:vreg_64_align2 = V_PK_MOV_B32 12, %4, 8, undef %5, 0, 0, 0, 0, 0, implicit $exec
+    %7:vgpr_32 = COPY %6.sub1
+    %8:vgpr_32 = COPY %6.sub0
+    $vgpr4 = COPY %7
+    $vgpr5 = COPY %8
+    S_ENDPGM 0, implicit $vgpr4, implicit $vgpr5
+
+...
+
+---
+name: v_pk_mov_b32__reg_sequence_shuffle_vgpr_vgpr_3_0_undef_undef
+tracksRegLiveness: true
+body:             |
+  bb.0:
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+
+    ; CHECK-LABEL: name: v_pk_mov_b32__reg_sequence_shuffle_vgpr_vgpr_3_0_undef_undef
+    ; CHECK: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+    ; CHECK-NEXT: {{  $}}
+    ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+    ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+    ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+    ; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+    ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64_align2 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1
+    ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64_align2 = REG_SEQUENCE [[COPY2]], %subreg.sub0, [[COPY3]], %subreg.sub1
+    ; CHECK-NEXT: [[V_PK_MOV_B32_:%[0-9]+]]:vreg_64_align2 = V_PK_MOV_B32 12, undef [[REG_SEQUENCE]], 8, undef [[REG_SEQUENCE1]], 0, 0, 0, 0, 0, implicit $exec
+    ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub1
+    ; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub0
+    ; CHECK-NEXT: $vgpr4 = COPY [[COPY4]]
+    ; CHECK-NEXT: $vgpr5 = COPY [[COPY5]]
+    ; CHECK-NEXT: S_ENDPGM 0, implicit $vgpr4, implicit $vgpr5
+    %0:vgpr_32 = COPY $vgpr0
+    %1:vgpr_32 = COPY $vgpr1
+    %2:vgpr_32 = COPY $vgpr2
+    %3:vgpr_32 = COPY $vgpr3
+    %4:vreg_64_align2 = REG_SEQUENCE %0, %subreg.sub0, %1, %subreg.sub1
+    %5:vreg_64_align2 = REG_SEQUENCE %2, %subreg.sub0, %3, %subreg.sub1
+    %6:vreg_64_align2 = V_PK_MOV_B32 12, undef %4, 8, undef %5, 0, 0, 0, 0, 0, implicit $exec
+    %7:vgpr_32 = COPY %6.sub1
+    %8:vgpr_32 = COPY %6.sub0
+    $vgpr4 = COPY %7
+    $vgpr5 = COPY %8
+    S_ENDPGM 0, implicit $vgpr4, implicit $vgpr5
+
+...
+
+---
+name: v_pk_mov_b32__reg_sequence_shuffle_vgpr_vgpr_3_0_compose_src_subregs_0
+tracksRegLiveness: true
+body:             |
+  bb.0:
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
+
+    ; CHECK-LABEL: name: v_pk_mov_b32__reg_sequence_shuffle_vgpr_vgpr_3_0_compose_src_subregs_0
+    ; CHECK: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
+    ; CHECK-NEXT: {{  $}}
+    ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+    ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+    ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+    ; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+    ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr4
+    ; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr5
+    ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_96_align2 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2
+    ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_96_align2 = REG_SEQUENCE [[COPY3]], %subreg.sub0, [[COPY4]], %subreg.sub1, [[COPY5]], %subreg.sub2
+    ; CHECK-NEXT: [[V_PK_MOV_B32_:%[0-9]+]]:vreg_64_align2 = V_PK_MOV_B32 12, [[REG_SEQUENCE]].sub0_sub1, 8, [[REG_SEQUENCE1]].sub0_sub1, 0, 0, 0, 0, 0, implicit $exec
+    ; CHECK-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub1
+    ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub0
+    ; CHECK-NEXT: $vgpr4 = COPY [[COPY6]]
+    ; CHECK-NEXT: $vgpr5 = COPY [[COPY7]]
+    ; CHECK-NEXT: S_ENDPGM 0, implicit $vgpr4, implicit $vgpr5
+    %0:vgpr_32 = COPY $vgpr0
+    %1:vgpr_32 = COPY $vgpr1
+    %2:vgpr_32 = COPY $vgpr2
+    %3:vgpr_32 = COPY $vgpr3
+    %4:vgpr_32 = COPY $vgpr4
+    %5:vgpr_32 = COPY $vgpr5
+    %6:vreg_96_align2 = REG_SEQUENCE %0, %subreg.sub0, %1, %subreg.sub1, %2, %subreg.sub2
+    %7:vreg_96_align2 = REG_SEQUENCE %3, %subreg.sub0, %4, %subreg.sub1, %5, %subreg.sub2
+    %8:vreg_64_align2 = V_PK_MOV_B32 12, %6.sub0_sub1, 8, %7.sub0_sub1, 0, 0, 0, 0, 0, implicit $exec
+    %9:vgpr_32 = COPY %8.sub1
+    %10:vgpr_32 = COPY %8.sub0
+    $vgpr4 = COPY %9
+    $vgpr5 = COPY %10
+    S_ENDPGM 0, implicit $vgpr4, implicit $vgpr5
+
+...
+
+---
+name: v_pk_mov_b32__reg_sequence_shuffle_vgpr_vgpr_0_3_compose_src_subregs_0
+tracksRegLiveness: true
+body:             |
+  bb.0:
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
+
+    ; CHECK-LABEL: name: v_pk_mov_b32__reg_sequence_shuffle_vgpr_vgpr_0_3_compose_src_subregs_0
+    ; CHECK: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
+    ; CHECK-NEXT: {{  $}}
+    ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+    ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+    ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+    ; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+    ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr4
+    ; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr5
+    ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_96_align2 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2
+    ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_96_align2 = REG_SEQUENCE [[COPY3]], %subreg.sub0, [[COPY4]], %subreg.sub1, [[COPY5]], %subreg.sub2
+    ; CHECK-NEXT: [[V_PK_MOV_B32_:%[0-9]+]]:vreg_64_align2 = V_PK_MOV_B32 8, [[REG_SEQUENCE]].sub0_sub1, 12, [[REG_SEQUENCE1]].sub0_sub1, 0, 0, 0, 0, 0, implicit $exec
+    ; CHECK-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub1
+    ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub0
+    ; CHECK-NEXT: $vgpr4 = COPY [[COPY6]]
+    ; CHECK-NEXT: $vgpr5 = COPY [[COPY7]]
+    ; CHECK-NEXT: S_ENDPGM 0, implicit $vgpr4, implicit $vgpr5
+    %0:vgpr_32 = COPY $vgpr0
+    %1:vgpr_32 = COPY $vgpr1
+    %2:vgpr_32 = COPY $vgpr2
+    %3:vgpr_32 = COPY $vgpr3
+    %4:vgpr_32 = COPY $vgpr4
+    %5:vgpr_32 = COPY $vgpr5
+    %6:vreg_96_align2 = REG_SEQUENCE %0, %subreg.sub0, %1, %subreg.sub1, %2, %subreg.sub2
+    %7:vreg_96_align2 = REG_SEQUENCE %3, %subreg.sub0, %4, %subreg.sub1, %5, %subreg.sub2
+    %8:vreg_64_align2 = V_PK_MOV_B32 8, %6.sub0_sub1, 12, %7.sub0_sub1, 0, 0, 0, 0, 0, implicit $exec
+    %9:vgpr_32 = COPY %8.sub1
+    %10:vgpr_32 = COPY %8.sub0
+    $vgpr4 = COPY %9
+    $vgpr5 = COPY %10
+    S_ENDPGM 0, implicit $vgpr4, implicit $vgpr5
+
+...
+
+---
+name: v_pk_mov_b32__reg_sequence_shuffle_vgpr_vgpr_compose_src_subregs_1
+tracksRegLiveness: true
+body:             |
+  bb.0:
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7
+
+    ; CHECK-LABEL: name: v_pk_mov_b32__reg_sequence_shuffle_vgpr_vgpr_compose_src_subregs_1
+    ; CHECK: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7
+    ; CHECK-NEXT: {{  $}}
+    ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+    ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+    ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+    ; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+    ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr4
+    ; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr5
+    ; CHECK-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr6
+    ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY $vgpr7
+    ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128_align2 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
+    ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_128_align2 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1, [[COPY6]], %subreg.sub2, [[COPY7]], %subreg.sub3
+    ; CHECK-NEXT: [[V_PK_MOV_B32_:%[0-9]+]]:vreg_64_align2 = V_PK_MOV_B32 12, [[REG_SEQUENCE]].sub2_sub3, 12, [[REG_SEQUENCE1]].sub0_sub1, 0, 0, 0, 0, 0, implicit $exec
+    ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub1
+    ; CHECK-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[V_PK_MOV_B32_]].sub0
+    ; CHECK-NEXT: $vgpr4 = COPY [[COPY8]]
+    ; CHECK-NEXT: $vgpr5 = COPY [[COPY9]]
+    ; CHECK-NEXT: S_ENDPGM 0, implicit $vgpr4, implicit $vgpr5
+    %0:vgpr_32 = COPY $vgpr0
+    %1:vgpr_32 = COPY $vgpr1
+    %2:vgpr_32 = COPY $vgpr2
+    %3:vgpr_32 = COPY $vgpr3
+    %4:vgpr_32 = COPY $vgpr4
+    %5:vgpr_32 = COPY $vgpr5
+    %6:vgpr_32 = COPY $vgpr6
+    %7:vgpr_32 = COPY $vgpr7
+    %8:vreg_128_align2 = REG_SEQUENCE %0, %subreg.sub0, %1, %subreg.sub1, %2, %subreg.sub2, %3, %subreg.sub3
+    %9:vreg_128_align2 = REG_SEQUENCE %4, %subreg.sub0, %5, %subreg.sub1, %6, %subreg.sub2, %7, %subreg.sub3
+    %10:vreg_64_align2 = V_PK_MOV_B32 12, %8.sub2_sub3, 12, %9.sub0_sub1, 0, 0, 0, 0, 0, implicit $exec
+    %11:vgpr_32 = COPY %10.sub1
+    %12:vgpr_32 = COPY %10.s...
[truncated]

github-actions · 2025-02-04T09:30:52Z

⚠️ undef deprecator found issues in your code. ⚠️

You can test this locally with the following command:

git diff -U0 --pickaxe-regex -S '([^a-zA-Z0-9#_-]undef[^a-zA-Z0-9_-]|UndefValue::get)' d21fc58aeeaa7f0369a24dbe70a0360e0edbf76f 436447d162d2d4429cbef4ccb82e3a9cab13836d

The following files introduce new uses of undef:

llvm/test/CodeGen/AMDGPU/reg-sequence-like-v-pk-mov-b32.mir

Undef is now deprecated and should only be used in the rare cases where no replacement is possible. For example, a load of uninitialized memory yields undef. You should use poison values for placeholders instead.

In tests, avoid using undef and having tests that trigger undefined behavior. If you need an operand with some unimportant value, you can add a new argument to the function and use that instead.

For example, this is considered a bad practice:

define void @fn() {
  ...
  br i1 undef, ...
}

Please use the following instead:

define void @fn(i1 %cond) {
  ...
  br i1 %cond, ...
}

Please refer to the Undefined Behavior Manual for more information.

…llvm#125656)

arsenm added the backend:AMDGPU label Feb 4, 2025 — with Graphite App

arsenm requested review from jayfoad, kparzysz, qcolombet, RKSimon, sdesmalen-arm and wangpc-pp February 4, 2025 09:27

This was referenced Feb 4, 2025

X86: Remove hack in shouldRewriteCopySrc for subregister handling #125224

Merged

PeepholeOpt: Fix looking for def of current copy to coalesce #125533

Merged

This was referenced Feb 4, 2025

AMDGPU: Use default shouldRewriteCopySrc #125535

Merged

WIP: AMDGPU: Implement getRegSequenceLikeInputs for v_pk_mov_b32 #125657

Open

arsenm requested review from Sisyph and rampitec and removed request for RKSimon, qcolombet, sdesmalen-arm and wangpc-pp February 4, 2025 09:27

arsenm marked this pull request as ready for review February 4, 2025 09:27

arsenm mentioned this pull request Feb 4, 2025

PeepholeOpt: Do not skip reg_sequence sources with subregs #125667

Open

rampitec approved these changes Feb 4, 2025

View reviewed changes

arsenm force-pushed the users/arsenm/amdgpu/use-default-shouldRewriteCopySrc branch from e7b88d2 to bbb7726 Compare February 5, 2025 16:33

arsenm force-pushed the users/arsenm/amdgpu/add-baseline-test-reg-sequence-like-v-pk-mov-b32 branch from 10c3b75 to d538f04 Compare February 5, 2025 16:35

Base automatically changed from users/arsenm/amdgpu/use-default-shouldRewriteCopySrc to main February 7, 2025 05:31

AMDGPU: Add baseline test for treating v_pk_mov_b32 like reg_sequence

436447d

arsenm force-pushed the users/arsenm/amdgpu/add-baseline-test-reg-sequence-like-v-pk-mov-b32 branch from d538f04 to 436447d Compare February 7, 2025 05:35

arsenm merged commit eef0205 into main Feb 13, 2025
7 of 8 checks passed

arsenm deleted the users/arsenm/amdgpu/add-baseline-test-reg-sequence-like-v-pk-mov-b32 branch February 13, 2025 11:12

flovent pushed a commit to flovent/llvm-project that referenced this pull request Feb 13, 2025

AMDGPU: Add baseline test for treating v_pk_mov_b32 like reg_sequence (…

6912175

…llvm#125656)

joaosaffran pushed a commit to joaosaffran/llvm-project that referenced this pull request Feb 14, 2025

AMDGPU: Add baseline test for treating v_pk_mov_b32 like reg_sequence (…

c71df06

…llvm#125656)

sivan-shani pushed a commit to sivan-shani/llvm-project that referenced this pull request Feb 24, 2025

AMDGPU: Add baseline test for treating v_pk_mov_b32 like reg_sequence (…

86aa9cc

…llvm#125656)

YutongZhuu pushed a commit to YutongZhuu/llvm-project that referenced this pull request Mar 8, 2025

AMDGPU: Add baseline test for treating v_pk_mov_b32 like reg_sequence (…

1d56e2a

…llvm#125656)

YutongZhuu pushed a commit to YutongZhuu/llvm-project that referenced this pull request Mar 8, 2025

AMDGPU: Add baseline test for treating v_pk_mov_b32 like reg_sequence (…

8a37e4c

…llvm#125656)

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

AMDGPU: Add baseline test for treating v_pk_mov_b32 like reg_sequence #125656

AMDGPU: Add baseline test for treating v_pk_mov_b32 like reg_sequence #125656

arsenm commented Feb 4, 2025

arsenm commented Feb 4, 2025 •

edited

Loading

llvmbot commented Feb 4, 2025

github-actions bot commented Feb 4, 2025 •

edited

Loading

AMDGPU: Add baseline test for treating v_pk_mov_b32 like reg_sequence #125656

AMDGPU: Add baseline test for treating v_pk_mov_b32 like reg_sequence #125656

Conversation

arsenm commented Feb 4, 2025

arsenm commented Feb 4, 2025 • edited Loading

llvmbot commented Feb 4, 2025

github-actions bot commented Feb 4, 2025 • edited Loading

arsenm commented Feb 4, 2025 •

edited

Loading

github-actions bot commented Feb 4, 2025 •

edited

Loading