llvm-project/llvm/test/CodeGen/AMDGPU/local-stack-alloc-add-references.gfx8.mir
Ivan Kosarev 21c1ba16ed
[TableGen] Complete the support for artificial registers (#183371)
Artificial registers were added in
eb0c510ecde667cd911682cc1e855f73f341d134
as a means of giving super-registers heavier weights than that
of their subregisters, even when they only contain a single
physical subregister.

Artifical registers thus do exist in code and participate in
register unit weight calculations, but are not supposed to be
available for register allocation.

This patch completes the support for artificial registers to:

- Ignore artificial registers when joining register unit uber
  sets. Artificial registers may be members of classes that
  together include registers and their sub-registers, making it
  impossible to compute normalised weights for uber sets they
  belong to.

  We have a use case downstream relying on this being supported,
  which allows to avoid introducing a large number of additional
  register classes.

- Not generate purely artificial register class intersections.
  It is critical not to have such classes, as the common LLVM
  codegen infrastructure will try to use them to constrain
  classes of virtual registers instead of producing COPYs
  whenever both the source and target register classes contain
  the same artificial registers.

- Not generate sub-classes where classes with the same
  non-artificial members already exist. This is mostly for
  convenience. For example, the HI16-capable subset of AMDGPU's
  AV_32 is VGPR_32, except VGPR_32 also contains the artificial
  staging registers. If the staging registers are not ignored,
  we'll end up having an additional generated register class,
  AV_32_with_hi16_in_VGPR_16, -- harmless, but also useless.

Eliminates a few inferred AMDGPU register classes:
    - VS_32_with_hi16
    - VS_32_Lo256_with_hi16
    - VS_32_Lo128_with_hi16
    - VRegOrLds_32_and_VS_32_Lo256
    - VRegOrLds_32_and_VS_32_Lo128
    - SRegOrLds_32_and_VRegOrLds_32

Causes no register class changes for other targets.
2026-03-04 13:33:26 +00:00

1018 lines
67 KiB
YAML

# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py UTC_ARGS: --version 5
# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx803 -run-pass=localstackalloc -verify-machineinstrs -o - %s | FileCheck -check-prefix=GFX803 %s
# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 -run-pass=localstackalloc -verify-machineinstrs -o - %s | FileCheck -check-prefix=GFX900 %s
# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx942 -run-pass=localstackalloc -verify-machineinstrs -o - %s | FileCheck -check-prefix=GFX942 %s
# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1030 -mattr=+wavefrontsize64 -run-pass=localstackalloc -verify-machineinstrs -o - %s | FileCheck -check-prefix=GFX10 %s
# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1200 -mattr=+wavefrontsize64 -run-pass=localstackalloc -verify-machineinstrs -o - %s | FileCheck -check-prefix=GFX12 %s
---
name: local_stack_alloc__v_add_co_u32_e32__literal_offsets
tracksRegLiveness: true
stack:
- { id: 0, size: 4096, alignment: 4 }
machineFunctionInfo:
scratchRSrcReg: '$sgpr0_sgpr1_sgpr2_sgpr3'
frameOffsetReg: '$sgpr33'
stackPtrOffsetReg: '$sgpr32'
body: |
bb.0:
; GFX803-LABEL: name: local_stack_alloc__v_add_co_u32_e32__literal_offsets
; GFX803: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 256
; GFX803-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX803-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX803-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_CO_U32_e64_]]
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[COPY]]
; GFX803-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 256, [[V_ADD_CO_U32_e64_]], implicit-def dead $vcc, implicit $exec
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__v_add_co_u32_e32__literal_offsets
; GFX900: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 256
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX900-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[COPY]]
; GFX900-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 256, [[V_ADD_U32_e64_]], implicit-def dead $vcc, implicit $exec
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_co_u32_e32__literal_offsets
; GFX942: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 256
; GFX942-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 %stack.0
; GFX942-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32_xexec_hi = S_ADD_I32 killed [[S_MOV_B32_]], [[S_MOV_B32_1]], implicit-def dead $scc
; GFX942-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[COPY]]
; GFX942-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX942-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 256, [[COPY1]], implicit-def dead $vcc, implicit $exec
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_co_u32_e32__literal_offsets
; GFX10: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 256
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[COPY]]
; GFX10-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 256, [[V_ADD_U32_e64_]], implicit-def dead $vcc, implicit $exec
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_co_u32_e32__literal_offsets
; GFX12: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 256
; GFX12-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 %stack.0
; GFX12-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32_xexec_hi = S_ADD_I32 killed [[S_MOV_B32_]], [[S_MOV_B32_1]], implicit-def dead $scc
; GFX12-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[COPY]]
; GFX12-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX12-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 256, [[COPY1]], implicit-def dead $vcc, implicit $exec
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX12-NEXT: SI_RETURN
%0:vgpr_32 = V_ADD_CO_U32_e32 256, %stack.0, implicit-def dead $vcc, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_CO_U32_e32 512, %stack.0, implicit-def dead $vcc, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, %1
SI_RETURN
...
---
name: local_stack_alloc__v_add_co_u32_e32__literal_offsets_live_vcc
tracksRegLiveness: true
stack:
- { id: 0, size: 4096, alignment: 4 }
machineFunctionInfo:
scratchRSrcReg: '$sgpr0_sgpr1_sgpr2_sgpr3'
frameOffsetReg: '$sgpr33'
stackPtrOffsetReg: '$sgpr32'
body: |
bb.0:
; GFX803-LABEL: name: local_stack_alloc__v_add_co_u32_e32__literal_offsets_live_vcc
; GFX803: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 256, %stack.0, implicit-def dead $vcc, implicit $exec
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX803-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 512, %stack.0, implicit-def $vcc, implicit $exec
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]], implicit $vcc
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__v_add_co_u32_e32__literal_offsets_live_vcc
; GFX900: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 256, %stack.0, implicit-def dead $vcc, implicit $exec
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX900-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 512, %stack.0, implicit-def $vcc, implicit $exec
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]], implicit $vcc
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_co_u32_e32__literal_offsets_live_vcc
; GFX942: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 256, %stack.0, implicit-def dead $vcc, implicit $exec
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX942-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 512, %stack.0, implicit-def $vcc, implicit $exec
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]], implicit $vcc
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_co_u32_e32__literal_offsets_live_vcc
; GFX10: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 256, %stack.0, implicit-def dead $vcc, implicit $exec
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX10-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 512, %stack.0, implicit-def $vcc, implicit $exec
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]], implicit $vcc
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_co_u32_e32__literal_offsets_live_vcc
; GFX12: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 256, %stack.0, implicit-def dead $vcc, implicit $exec
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX12-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 512, %stack.0, implicit-def $vcc, implicit $exec
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]], implicit $vcc
; GFX12-NEXT: SI_RETURN
%0:vgpr_32 = V_ADD_CO_U32_e32 256, %stack.0, implicit-def dead $vcc, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_CO_U32_e32 512, %stack.0, implicit-def $vcc, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, %1, implicit $vcc
SI_RETURN
...
---
name: local_stack_alloc__v_add_co_u32_e32__inline_imm_offsets
tracksRegLiveness: true
stack:
- { id: 0, size: 64, alignment: 4 }
machineFunctionInfo:
scratchRSrcReg: '$sgpr0_sgpr1_sgpr2_sgpr3'
frameOffsetReg: '$sgpr33'
stackPtrOffsetReg: '$sgpr32'
body: |
bb.0:
; GFX803-LABEL: name: local_stack_alloc__v_add_co_u32_e32__inline_imm_offsets
; GFX803: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 8
; GFX803-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX803-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX803-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_CO_U32_e64_]]
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[COPY]]
; GFX803-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 8, [[V_ADD_CO_U32_e64_]], implicit-def dead $vcc, implicit $exec
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__v_add_co_u32_e32__inline_imm_offsets
; GFX900: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 8
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX900-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[COPY]]
; GFX900-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 8, [[V_ADD_U32_e64_]], implicit-def dead $vcc, implicit $exec
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_co_u32_e32__inline_imm_offsets
; GFX942: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 8
; GFX942-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 %stack.0
; GFX942-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32_xexec_hi = S_ADD_I32 killed [[S_MOV_B32_]], [[S_MOV_B32_1]], implicit-def dead $scc
; GFX942-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[COPY]]
; GFX942-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX942-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 8, [[COPY1]], implicit-def dead $vcc, implicit $exec
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_co_u32_e32__inline_imm_offsets
; GFX10: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 8
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[COPY]]
; GFX10-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 8, [[V_ADD_U32_e64_]], implicit-def dead $vcc, implicit $exec
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_co_u32_e32__inline_imm_offsets
; GFX12: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 8
; GFX12-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 %stack.0
; GFX12-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32_xexec_hi = S_ADD_I32 killed [[S_MOV_B32_]], [[S_MOV_B32_1]], implicit-def dead $scc
; GFX12-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[COPY]]
; GFX12-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX12-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 8, [[COPY1]], implicit-def dead $vcc, implicit $exec
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX12-NEXT: SI_RETURN
%0:vgpr_32 = V_ADD_CO_U32_e32 8, %stack.0, implicit-def dead $vcc, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_CO_U32_e32 16, %stack.0, implicit-def dead $vcc, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, %1
SI_RETURN
...
---
name: local_stack_alloc__v_add_co_u32_e64__inline_imm_offsets
tracksRegLiveness: true
stack:
- { id: 0, size: 64, alignment: 4 }
machineFunctionInfo:
scratchRSrcReg: '$sgpr0_sgpr1_sgpr2_sgpr3'
frameOffsetReg: '$sgpr33'
stackPtrOffsetReg: '$sgpr32'
body: |
bb.0:
; GFX803-LABEL: name: local_stack_alloc__v_add_co_u32_e64__inline_imm_offsets
; GFX803: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 8
; GFX803-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX803-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX803-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_CO_U32_e64_]]
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[COPY]]
; GFX803-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 8, [[V_ADD_CO_U32_e64_]], 0, implicit $exec
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__v_add_co_u32_e64__inline_imm_offsets
; GFX900: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 8
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX900-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[COPY]]
; GFX900-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 8, [[V_ADD_U32_e64_]], 0, implicit $exec
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_co_u32_e64__inline_imm_offsets
; GFX942: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 8
; GFX942-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 %stack.0
; GFX942-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32_xexec_hi = S_ADD_I32 killed [[S_MOV_B32_]], [[S_MOV_B32_1]], implicit-def dead $scc
; GFX942-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[COPY]]
; GFX942-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 8, [[S_ADD_I32_]], 0, implicit $exec
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_co_u32_e64__inline_imm_offsets
; GFX10: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 8
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[COPY]]
; GFX10-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 8, [[V_ADD_U32_e64_]], 0, implicit $exec
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_co_u32_e64__inline_imm_offsets
; GFX12: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 8
; GFX12-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 %stack.0
; GFX12-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32_xexec_hi = S_ADD_I32 killed [[S_MOV_B32_]], [[S_MOV_B32_1]], implicit-def dead $scc
; GFX12-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[COPY]]
; GFX12-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 8, [[S_ADD_I32_]], 0, implicit $exec
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX12-NEXT: SI_RETURN
%0:vgpr_32, dead %2:sreg_64_xexec = V_ADD_CO_U32_e64 %stack.0, 8, 0, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, %0
%1:vgpr_32, dead %3:sreg_64_xexec = V_ADD_CO_U32_e64 16, %stack.0, 0, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, %1
SI_RETURN
...
---
name: local_stack_alloc__v_add_co_u32_e64__inline_imm_offsets_live_vcc
tracksRegLiveness: true
stack:
- { id: 0, size: 64, alignment: 4 }
machineFunctionInfo:
scratchRSrcReg: '$sgpr0_sgpr1_sgpr2_sgpr3'
frameOffsetReg: '$sgpr33'
stackPtrOffsetReg: '$sgpr32'
body: |
bb.0:
; GFX803-LABEL: name: local_stack_alloc__v_add_co_u32_e64__inline_imm_offsets_live_vcc
; GFX803: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %stack.0, 8, 0, implicit $exec
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX803-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 16, %stack.0, 0, implicit $exec
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX803-NEXT: SI_RETURN implicit [[V_ADD_CO_U32_e64_1]]
;
; GFX900-LABEL: name: local_stack_alloc__v_add_co_u32_e64__inline_imm_offsets_live_vcc
; GFX900: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %stack.0, 8, 0, implicit $exec
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX900-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 16, %stack.0, 0, implicit $exec
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX900-NEXT: SI_RETURN implicit [[V_ADD_CO_U32_e64_1]]
;
; GFX942-LABEL: name: local_stack_alloc__v_add_co_u32_e64__inline_imm_offsets_live_vcc
; GFX942: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %stack.0, 8, 0, implicit $exec
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX942-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 16, %stack.0, 0, implicit $exec
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX942-NEXT: SI_RETURN implicit [[V_ADD_CO_U32_e64_1]]
;
; GFX10-LABEL: name: local_stack_alloc__v_add_co_u32_e64__inline_imm_offsets_live_vcc
; GFX10: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %stack.0, 8, 0, implicit $exec
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX10-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 16, %stack.0, 0, implicit $exec
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX10-NEXT: SI_RETURN implicit [[V_ADD_CO_U32_e64_1]]
;
; GFX12-LABEL: name: local_stack_alloc__v_add_co_u32_e64__inline_imm_offsets_live_vcc
; GFX12: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %stack.0, 8, 0, implicit $exec
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX12-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 16, %stack.0, 0, implicit $exec
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX12-NEXT: SI_RETURN implicit [[V_ADD_CO_U32_e64_1]]
%0:vgpr_32, %2:sreg_64_xexec = V_ADD_CO_U32_e64 %stack.0, 8, 0, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, %0
%1:vgpr_32, %3:sreg_64_xexec = V_ADD_CO_U32_e64 16, %stack.0, 0, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, %1
SI_RETURN implicit %2
...
---
name: local_stack_alloc__s_add_i32__literal_offsets
tracksRegLiveness: true
stack:
- { id: 0, size: 4096, alignment: 4 }
machineFunctionInfo:
scratchRSrcReg: '$sgpr0_sgpr1_sgpr2_sgpr3'
frameOffsetReg: '$sgpr33'
stackPtrOffsetReg: '$sgpr32'
body: |
bb.0:
; GFX803-LABEL: name: local_stack_alloc__s_add_i32__literal_offsets
; GFX803: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def dead $scc
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX803-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def dead $scc
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__s_add_i32__literal_offsets
; GFX900: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def dead $scc
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX900-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def dead $scc
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__s_add_i32__literal_offsets
; GFX942: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def dead $scc
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX942-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def dead $scc
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__s_add_i32__literal_offsets
; GFX10: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def dead $scc
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX10-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def dead $scc
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__s_add_i32__literal_offsets
; GFX12: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def dead $scc
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX12-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def dead $scc
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX12-NEXT: SI_RETURN
%0:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def dead $scc
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, %0
%1:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def dead $scc
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, %1
SI_RETURN
...
---
name: local_stack_alloc__s_add_i32__inline_imm_offsets
tracksRegLiveness: true
stack:
- { id: 0, size: 64, alignment: 4 }
machineFunctionInfo:
scratchRSrcReg: '$sgpr0_sgpr1_sgpr2_sgpr3'
frameOffsetReg: '$sgpr33'
stackPtrOffsetReg: '$sgpr32'
body: |
bb.0:
; GFX803-LABEL: name: local_stack_alloc__s_add_i32__inline_imm_offsets
; GFX803: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 8, %stack.0, implicit-def dead $scc
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX803-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 16, %stack.0, implicit-def dead $scc
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__s_add_i32__inline_imm_offsets
; GFX900: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 8, %stack.0, implicit-def dead $scc
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX900-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 16, %stack.0, implicit-def dead $scc
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__s_add_i32__inline_imm_offsets
; GFX942: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 8, %stack.0, implicit-def dead $scc
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX942-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 16, %stack.0, implicit-def dead $scc
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__s_add_i32__inline_imm_offsets
; GFX10: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 8, %stack.0, implicit-def dead $scc
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX10-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 16, %stack.0, implicit-def dead $scc
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__s_add_i32__inline_imm_offsets
; GFX12: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 8, %stack.0, implicit-def dead $scc
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX12-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 16, %stack.0, implicit-def dead $scc
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX12-NEXT: SI_RETURN
%0:sreg_32 = S_ADD_I32 8, %stack.0, implicit-def dead $scc
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, %0
%1:sreg_32 = S_ADD_I32 16, %stack.0, implicit-def dead $scc
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, %1
SI_RETURN
...
---
name: local_stack_alloc__s_add_i32__reg_offsets
tracksRegLiveness: true
stack:
- { id: 0, size: 64, alignment: 4 }
machineFunctionInfo:
scratchRSrcReg: '$sgpr0_sgpr1_sgpr2_sgpr3'
frameOffsetReg: '$sgpr33'
stackPtrOffsetReg: '$sgpr32'
body: |
bb.0:
liveins: $sgpr4, $sgpr5
; GFX803-LABEL: name: local_stack_alloc__s_add_i32__reg_offsets
; GFX803: liveins: $sgpr4, $sgpr5
; GFX803-NEXT: {{ $}}
; GFX803-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX803-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX803-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], %stack.0, implicit-def dead $scc
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX803-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], %stack.0, implicit-def dead $scc
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__s_add_i32__reg_offsets
; GFX900: liveins: $sgpr4, $sgpr5
; GFX900-NEXT: {{ $}}
; GFX900-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX900-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX900-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], %stack.0, implicit-def dead $scc
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX900-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], %stack.0, implicit-def dead $scc
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__s_add_i32__reg_offsets
; GFX942: liveins: $sgpr4, $sgpr5
; GFX942-NEXT: {{ $}}
; GFX942-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX942-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX942-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], %stack.0, implicit-def dead $scc
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX942-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], %stack.0, implicit-def dead $scc
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__s_add_i32__reg_offsets
; GFX10: liveins: $sgpr4, $sgpr5
; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX10-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], %stack.0, implicit-def dead $scc
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX10-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], %stack.0, implicit-def dead $scc
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__s_add_i32__reg_offsets
; GFX12: liveins: $sgpr4, $sgpr5
; GFX12-NEXT: {{ $}}
; GFX12-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX12-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX12-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], %stack.0, implicit-def dead $scc
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX12-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], %stack.0, implicit-def dead $scc
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX12-NEXT: SI_RETURN
%0:sreg_32 = COPY $sgpr4
%1:sreg_32 = COPY $sgpr5
%2:sreg_32 = S_ADD_I32 %0, %stack.0, implicit-def dead $scc
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, %2
%3:sreg_32 = S_ADD_I32 %1, %stack.0, implicit-def dead $scc
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, %3
SI_RETURN
...
---
name: local_stack_alloc__s_add_i32__reg_offsets_commute
tracksRegLiveness: true
stack:
- { id: 0, size: 64, alignment: 4 }
machineFunctionInfo:
scratchRSrcReg: '$sgpr0_sgpr1_sgpr2_sgpr3'
frameOffsetReg: '$sgpr33'
stackPtrOffsetReg: '$sgpr32'
body: |
bb.0:
liveins: $sgpr4, $sgpr5
; GFX803-LABEL: name: local_stack_alloc__s_add_i32__reg_offsets_commute
; GFX803: liveins: $sgpr4, $sgpr5
; GFX803-NEXT: {{ $}}
; GFX803-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX803-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX803-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 %stack.0, [[COPY]], implicit-def dead $scc
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX803-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 %stack.0, [[COPY1]], implicit-def dead $scc
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__s_add_i32__reg_offsets_commute
; GFX900: liveins: $sgpr4, $sgpr5
; GFX900-NEXT: {{ $}}
; GFX900-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX900-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX900-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 %stack.0, [[COPY]], implicit-def dead $scc
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX900-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 %stack.0, [[COPY1]], implicit-def dead $scc
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__s_add_i32__reg_offsets_commute
; GFX942: liveins: $sgpr4, $sgpr5
; GFX942-NEXT: {{ $}}
; GFX942-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX942-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX942-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 %stack.0, [[COPY]], implicit-def dead $scc
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX942-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 %stack.0, [[COPY1]], implicit-def dead $scc
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__s_add_i32__reg_offsets_commute
; GFX10: liveins: $sgpr4, $sgpr5
; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX10-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 %stack.0, [[COPY]], implicit-def dead $scc
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX10-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 %stack.0, [[COPY1]], implicit-def dead $scc
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__s_add_i32__reg_offsets_commute
; GFX12: liveins: $sgpr4, $sgpr5
; GFX12-NEXT: {{ $}}
; GFX12-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX12-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX12-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 %stack.0, [[COPY]], implicit-def dead $scc
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX12-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 %stack.0, [[COPY1]], implicit-def dead $scc
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX12-NEXT: SI_RETURN
%0:sreg_32 = COPY $sgpr4
%1:sreg_32 = COPY $sgpr5
%2:sreg_32 = S_ADD_I32 %stack.0, %0, implicit-def dead $scc
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, %2
%3:sreg_32 = S_ADD_I32 %stack.0, %1, implicit-def dead $scc
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, %3
SI_RETURN
...
---
name: local_stack_alloc__s_add_i32__literal_offsets_live_scc
tracksRegLiveness: true
stack:
- { id: 0, size: 4096, alignment: 4 }
machineFunctionInfo:
scratchRSrcReg: '$sgpr0_sgpr1_sgpr2_sgpr3'
frameOffsetReg: '$sgpr33'
stackPtrOffsetReg: '$sgpr32'
body: |
bb.0:
; GFX803-LABEL: name: local_stack_alloc__s_add_i32__literal_offsets_live_scc
; GFX803: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def $scc
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX803-NEXT: S_NOP 0, implicit $scc
; GFX803-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def $scc
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX803-NEXT: SI_RETURN implicit $scc
;
; GFX900-LABEL: name: local_stack_alloc__s_add_i32__literal_offsets_live_scc
; GFX900: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def $scc
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX900-NEXT: S_NOP 0, implicit $scc
; GFX900-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def $scc
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX900-NEXT: SI_RETURN implicit $scc
;
; GFX942-LABEL: name: local_stack_alloc__s_add_i32__literal_offsets_live_scc
; GFX942: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def $scc
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX942-NEXT: S_NOP 0, implicit $scc
; GFX942-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def $scc
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX942-NEXT: SI_RETURN implicit $scc
;
; GFX10-LABEL: name: local_stack_alloc__s_add_i32__literal_offsets_live_scc
; GFX10: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def $scc
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX10-NEXT: S_NOP 0, implicit $scc
; GFX10-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def $scc
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX10-NEXT: SI_RETURN implicit $scc
;
; GFX12-LABEL: name: local_stack_alloc__s_add_i32__literal_offsets_live_scc
; GFX12: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def $scc
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX12-NEXT: S_NOP 0, implicit $scc
; GFX12-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def $scc
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX12-NEXT: SI_RETURN implicit $scc
%0:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def $scc
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, %0
S_NOP 0, implicit $scc
%1:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def $scc
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1572873 /* reguse:SReg_32 */, %1
SI_RETURN implicit $scc
...
---
name: local_stack_alloc__v_add_co_u32_e32__vgpr_offsets
tracksRegLiveness: true
stack:
- { id: 0, size: 4096, alignment: 4 }
machineFunctionInfo:
scratchRSrcReg: '$sgpr0_sgpr1_sgpr2_sgpr3'
frameOffsetReg: '$sgpr33'
stackPtrOffsetReg: '$sgpr32'
body: |
bb.0:
liveins: $vgpr0
; GFX803-LABEL: name: local_stack_alloc__v_add_co_u32_e32__vgpr_offsets
; GFX803: liveins: $vgpr0
; GFX803-NEXT: {{ $}}
; GFX803-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX803-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX803-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %vgpr_offset, [[V_MOV_B32_e32_]], implicit-def dead $vcc, implicit $exec
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX803-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %vgpr_offset, [[V_MOV_B32_e32_]], implicit-def dead $vcc, implicit $exec
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__v_add_co_u32_e32__vgpr_offsets
; GFX900: liveins: $vgpr0
; GFX900-NEXT: {{ $}}
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX900-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %vgpr_offset, [[V_MOV_B32_e32_]], implicit-def dead $vcc, implicit $exec
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX900-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %vgpr_offset, [[V_MOV_B32_e32_]], implicit-def dead $vcc, implicit $exec
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_co_u32_e32__vgpr_offsets
; GFX942: liveins: $vgpr0
; GFX942-NEXT: {{ $}}
; GFX942-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xexec_hi = S_MOV_B32 %stack.0
; GFX942-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX942-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[S_MOV_B32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX942-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[S_MOV_B32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_co_u32_e32__vgpr_offsets
; GFX10: liveins: $vgpr0
; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %vgpr_offset, [[V_MOV_B32_e32_]], implicit-def dead $vcc, implicit $exec
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX10-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %vgpr_offset, [[V_MOV_B32_e32_]], implicit-def dead $vcc, implicit $exec
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_co_u32_e32__vgpr_offsets
; GFX12: liveins: $vgpr0
; GFX12-NEXT: {{ $}}
; GFX12-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xexec_hi = S_MOV_B32 %stack.0
; GFX12-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX12-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[S_MOV_B32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX12-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[S_MOV_B32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX12-NEXT: SI_RETURN
%vgpr_offset:vgpr_32 = COPY $vgpr0
%0:vgpr_32 = V_ADD_CO_U32_e32 %vgpr_offset, %stack.0, implicit-def dead $vcc, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_CO_U32_e32 %vgpr_offset, %stack.0, implicit-def dead $vcc, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, %1
SI_RETURN
...
---
name: local_stack_alloc__v_add_co_u32_e32__vgpr_offsets_commute
tracksRegLiveness: true
stack:
- { id: 0, size: 4096, alignment: 4 }
machineFunctionInfo:
scratchRSrcReg: '$sgpr0_sgpr1_sgpr2_sgpr3'
frameOffsetReg: '$sgpr33'
stackPtrOffsetReg: '$sgpr32'
body: |
bb.0:
liveins: $vgpr0
; GFX803-LABEL: name: local_stack_alloc__v_add_co_u32_e32__vgpr_offsets_commute
; GFX803: liveins: $vgpr0
; GFX803-NEXT: {{ $}}
; GFX803-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX803-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX803-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[V_MOV_B32_e32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX803-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[V_MOV_B32_e32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__v_add_co_u32_e32__vgpr_offsets_commute
; GFX900: liveins: $vgpr0
; GFX900-NEXT: {{ $}}
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX900-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[V_MOV_B32_e32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX900-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[V_MOV_B32_e32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_co_u32_e32__vgpr_offsets_commute
; GFX942: liveins: $vgpr0
; GFX942-NEXT: {{ $}}
; GFX942-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xexec_hi = S_MOV_B32 %stack.0
; GFX942-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX942-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[S_MOV_B32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX942-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[S_MOV_B32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_co_u32_e32__vgpr_offsets_commute
; GFX10: liveins: $vgpr0
; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[V_MOV_B32_e32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX10-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[V_MOV_B32_e32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_co_u32_e32__vgpr_offsets_commute
; GFX12: liveins: $vgpr0
; GFX12-NEXT: {{ $}}
; GFX12-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xexec_hi = S_MOV_B32 %stack.0
; GFX12-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX12-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[S_MOV_B32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX12-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[S_MOV_B32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX12-NEXT: SI_RETURN
%vgpr_offset:vgpr_32 = COPY $vgpr0
%0:vgpr_32 = V_ADD_CO_U32_e32 %stack.0, %vgpr_offset, implicit-def dead $vcc, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_CO_U32_e32 %stack.0, %vgpr_offset, implicit-def dead $vcc, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, %1
SI_RETURN
...
---
name: local_stack_alloc__v_add_co_u32_e32__sgpr_offsets
tracksRegLiveness: true
stack:
- { id: 0, size: 4096, alignment: 4 }
machineFunctionInfo:
scratchRSrcReg: '$sgpr0_sgpr1_sgpr2_sgpr3'
frameOffsetReg: '$sgpr33'
stackPtrOffsetReg: '$sgpr32'
body: |
bb.0:
liveins: $sgpr8
; GFX803-LABEL: name: local_stack_alloc__v_add_co_u32_e32__sgpr_offsets
; GFX803: liveins: $sgpr8
; GFX803-NEXT: {{ $}}
; GFX803-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX803-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, %stack.0, implicit-def dead $vcc, implicit $exec
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX803-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, %stack.0, implicit-def dead $vcc, implicit $exec
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__v_add_co_u32_e32__sgpr_offsets
; GFX900: liveins: $sgpr8
; GFX900-NEXT: {{ $}}
; GFX900-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX900-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, %stack.0, implicit-def dead $vcc, implicit $exec
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX900-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, %stack.0, implicit-def dead $vcc, implicit $exec
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_co_u32_e32__sgpr_offsets
; GFX942: liveins: $sgpr8
; GFX942-NEXT: {{ $}}
; GFX942-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX942-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, %stack.0, implicit-def dead $vcc, implicit $exec
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX942-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, %stack.0, implicit-def dead $vcc, implicit $exec
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_co_u32_e32__sgpr_offsets
; GFX10: liveins: $sgpr8
; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX10-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, [[V_MOV_B32_e32_]], implicit-def dead $vcc, implicit $exec
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX10-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, [[V_MOV_B32_e32_]], implicit-def dead $vcc, implicit $exec
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_co_u32_e32__sgpr_offsets
; GFX12: liveins: $sgpr8
; GFX12-NEXT: {{ $}}
; GFX12-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xexec_hi = S_MOV_B32 %stack.0
; GFX12-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX12-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; GFX12-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, [[COPY]], implicit-def dead $vcc, implicit $exec
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX12-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; GFX12-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, [[COPY1]], implicit-def dead $vcc, implicit $exec
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX12-NEXT: SI_RETURN
%sgpr_offset:sreg_32 = COPY $sgpr8
%0:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, %stack.0, implicit-def dead $vcc, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, %stack.0, implicit-def dead $vcc, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, %1
SI_RETURN
...
---
name: local_stack_alloc__v_add_co_u32_e64__sgpr_offsets
tracksRegLiveness: true
stack:
- { id: 0, size: 4096, alignment: 4 }
machineFunctionInfo:
scratchRSrcReg: '$sgpr0_sgpr1_sgpr2_sgpr3'
frameOffsetReg: '$sgpr33'
stackPtrOffsetReg: '$sgpr32'
body: |
bb.0:
liveins: $sgpr8
; GFX803-LABEL: name: local_stack_alloc__v_add_co_u32_e64__sgpr_offsets
; GFX803: liveins: $sgpr8
; GFX803-NEXT: {{ $}}
; GFX803-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX803-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX803-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX803-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__v_add_co_u32_e64__sgpr_offsets
; GFX900: liveins: $sgpr8
; GFX900-NEXT: {{ $}}
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX900-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX900-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_co_u32_e64__sgpr_offsets
; GFX942: liveins: $sgpr8
; GFX942-NEXT: {{ $}}
; GFX942-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xexec_hi = S_MOV_B32 %stack.0
; GFX942-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX942-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; GFX942-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, [[COPY]], 0, implicit $exec
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX942-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; GFX942-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, [[COPY1]], 0, implicit $exec
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_co_u32_e64__sgpr_offsets
; GFX10: liveins: $sgpr8
; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX10-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX10-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_co_u32_e64__sgpr_offsets
; GFX12: liveins: $sgpr8
; GFX12-NEXT: {{ $}}
; GFX12-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xexec_hi = S_MOV_B32 %stack.0
; GFX12-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX12-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, [[S_MOV_B32_]], 0, implicit $exec
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX12-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, [[S_MOV_B32_]], 0, implicit $exec
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX12-NEXT: SI_RETURN
%sgpr_offset:sreg_32 = COPY $sgpr8
%0:vgpr_32, dead %2:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, %stack.0, 0, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, %0
%1:vgpr_32, dead %3:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, %stack.0, 0, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, %1
SI_RETURN
...
---
name: local_stack_alloc__v_add_co_u32_e64__sgpr_offsets_commute
tracksRegLiveness: true
stack:
- { id: 0, size: 4096, alignment: 4 }
machineFunctionInfo:
scratchRSrcReg: '$sgpr0_sgpr1_sgpr2_sgpr3'
frameOffsetReg: '$sgpr33'
stackPtrOffsetReg: '$sgpr32'
body: |
bb.0:
liveins: $sgpr8
; GFX803-LABEL: name: local_stack_alloc__v_add_co_u32_e64__sgpr_offsets_commute
; GFX803: liveins: $sgpr8
; GFX803-NEXT: {{ $}}
; GFX803-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX803-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX803-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[V_MOV_B32_e32_]], %sgpr_offset, 0, implicit $exec
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX803-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[V_MOV_B32_e32_]], %sgpr_offset, 0, implicit $exec
; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__v_add_co_u32_e64__sgpr_offsets_commute
; GFX900: liveins: $sgpr8
; GFX900-NEXT: {{ $}}
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX900-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[V_MOV_B32_e32_]], %sgpr_offset, 0, implicit $exec
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX900-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[V_MOV_B32_e32_]], %sgpr_offset, 0, implicit $exec
; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_co_u32_e64__sgpr_offsets_commute
; GFX942: liveins: $sgpr8
; GFX942-NEXT: {{ $}}
; GFX942-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xexec_hi = S_MOV_B32 %stack.0
; GFX942-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX942-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY %sgpr_offset
; GFX942-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[S_MOV_B32_]], [[COPY]], 0, implicit $exec
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX942-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY %sgpr_offset
; GFX942-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[S_MOV_B32_]], [[COPY1]], 0, implicit $exec
; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_co_u32_e64__sgpr_offsets_commute
; GFX10: liveins: $sgpr8
; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX10-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[V_MOV_B32_e32_]], %sgpr_offset, 0, implicit $exec
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX10-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[V_MOV_B32_e32_]], %sgpr_offset, 0, implicit $exec
; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_co_u32_e64__sgpr_offsets_commute
; GFX12: liveins: $sgpr8
; GFX12-NEXT: {{ $}}
; GFX12-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xexec_hi = S_MOV_B32 %stack.0
; GFX12-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX12-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[S_MOV_B32_]], %sgpr_offset, 0, implicit $exec
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX12-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[S_MOV_B32_]], %sgpr_offset, 0, implicit $exec
; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX12-NEXT: SI_RETURN
%sgpr_offset:sreg_32 = COPY $sgpr8
%0:vgpr_32, dead %2:sreg_64_xexec = V_ADD_CO_U32_e64 %stack.0, %sgpr_offset, 0, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, %0
%1:vgpr_32, dead %3:sreg_64_xexec = V_ADD_CO_U32_e64 %stack.0, %sgpr_offset, 0, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1179657 /* reguse:VGPR_32 */, %1
SI_RETURN
...