807 lines
36 KiB
YAML
807 lines
36 KiB
YAML
# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
|
|
# RUN: llc -amdgpu-global-isel-new-legality -march=amdgcn -mcpu=hawaii -run-pass=instruction-select -verify-machineinstrs -global-isel-abort=0 -o - %s | FileCheck -check-prefix=GFX7 %s
|
|
# RUN: llc -amdgpu-global-isel-new-legality -march=amdgcn -mcpu=fiji -run-pass=instruction-select -verify-machineinstrs -global-isel-abort=0 -o - %s | FileCheck -check-prefix=GFX8 %s
|
|
# RUN: llc -amdgpu-global-isel-new-legality -march=amdgcn -mcpu=gfx900 -run-pass=instruction-select -verify-machineinstrs -global-isel-abort=0 -o - %s | FileCheck -check-prefix=GFX9 %s
|
|
# RUN: llc -amdgpu-global-isel-new-legality -march=amdgcn -mcpu=gfx1010 -run-pass=instruction-select -verify-machineinstrs -global-isel-abort=0 -o - %s | FileCheck -check-prefix=GFX10 %s
|
|
|
|
---
|
|
|
|
name: store_flat_s32_to_4
|
|
legalized: true
|
|
tracksRegLiveness: true
|
|
regBankSelected: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2
|
|
|
|
; GFX7-LABEL: name: store_flat_s32_to_4
|
|
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX7: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX7: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX7: FLAT_STORE_DWORD [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 4)
|
|
; GFX8-LABEL: name: store_flat_s32_to_4
|
|
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX8: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX8: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX8: FLAT_STORE_DWORD [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 4)
|
|
; GFX9-LABEL: name: store_flat_s32_to_4
|
|
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX9: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX9: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX9: FLAT_STORE_DWORD [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 4)
|
|
; GFX10-LABEL: name: store_flat_s32_to_4
|
|
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX10: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX10: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX10: FLAT_STORE_DWORD [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 4)
|
|
%0:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(s32) = COPY $vgpr2
|
|
G_STORE %1, %0 :: (store 4, align 4, addrspace 0)
|
|
|
|
...
|
|
|
|
---
|
|
name: store_flat_s32_to_2
|
|
legalized: true
|
|
tracksRegLiveness: true
|
|
regBankSelected: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2
|
|
|
|
; GFX7-LABEL: name: store_flat_s32_to_2
|
|
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX7: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX7: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX7: FLAT_STORE_SHORT [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 2)
|
|
; GFX8-LABEL: name: store_flat_s32_to_2
|
|
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX8: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX8: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX8: FLAT_STORE_SHORT [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 2)
|
|
; GFX9-LABEL: name: store_flat_s32_to_2
|
|
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX9: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX9: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX9: FLAT_STORE_SHORT [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 2)
|
|
; GFX10-LABEL: name: store_flat_s32_to_2
|
|
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX10: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX10: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX10: FLAT_STORE_SHORT [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 2)
|
|
%0:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(s32) = COPY $vgpr2
|
|
G_STORE %1, %0 :: (store 2, align 2, addrspace 0)
|
|
|
|
...
|
|
|
|
---
|
|
name: store_flat_s32_to_1
|
|
legalized: true
|
|
tracksRegLiveness: true
|
|
regBankSelected: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2
|
|
|
|
; GFX7-LABEL: name: store_flat_s32_to_1
|
|
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX7: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX7: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX7: FLAT_STORE_BYTE [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 1)
|
|
; GFX8-LABEL: name: store_flat_s32_to_1
|
|
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX8: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX8: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX8: FLAT_STORE_BYTE [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 1)
|
|
; GFX9-LABEL: name: store_flat_s32_to_1
|
|
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX9: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX9: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX9: FLAT_STORE_BYTE [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 1)
|
|
; GFX10-LABEL: name: store_flat_s32_to_1
|
|
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX10: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX10: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX10: FLAT_STORE_BYTE [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 1)
|
|
%0:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(s32) = COPY $vgpr2
|
|
G_STORE %1, %0 :: (store 1, align 1, addrspace 0)
|
|
|
|
...
|
|
|
|
---
|
|
|
|
name: store_flat_s64
|
|
legalized: true
|
|
tracksRegLiveness: true
|
|
regBankSelected: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
|
|
; GFX7-LABEL: name: store_flat_s64
|
|
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX7: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX7: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
|
|
; GFX7: FLAT_STORE_DWORDX2 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 8)
|
|
; GFX8-LABEL: name: store_flat_s64
|
|
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX8: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX8: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
|
|
; GFX8: FLAT_STORE_DWORDX2 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 8)
|
|
; GFX9-LABEL: name: store_flat_s64
|
|
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX9: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX9: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
|
|
; GFX9: FLAT_STORE_DWORDX2 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 8)
|
|
; GFX10-LABEL: name: store_flat_s64
|
|
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX10: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX10: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
|
|
; GFX10: FLAT_STORE_DWORDX2 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 8)
|
|
%0:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(s64) = COPY $vgpr2_vgpr3
|
|
G_STORE %1, %0 :: (store 8, align 8, addrspace 0)
|
|
|
|
...
|
|
---
|
|
|
|
name: store_flat_s96
|
|
legalized: true
|
|
tracksRegLiveness: true
|
|
regBankSelected: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4
|
|
|
|
; GFX7-LABEL: name: store_flat_s96
|
|
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4
|
|
; GFX7: [[COPY:%[0-9]+]]:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
; GFX7: [[COPY1:%[0-9]+]]:vgpr(s96) = COPY $vgpr2_vgpr3_vgpr4
|
|
; GFX7: G_STORE [[COPY1]](s96), [[COPY]](p1) :: (store 12, align 16)
|
|
; GFX8-LABEL: name: store_flat_s96
|
|
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4
|
|
; GFX8: [[COPY:%[0-9]+]]:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
; GFX8: [[COPY1:%[0-9]+]]:vgpr(s96) = COPY $vgpr2_vgpr3_vgpr4
|
|
; GFX8: G_STORE [[COPY1]](s96), [[COPY]](p1) :: (store 12, align 16)
|
|
; GFX9-LABEL: name: store_flat_s96
|
|
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4
|
|
; GFX9: [[COPY:%[0-9]+]]:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
; GFX9: [[COPY1:%[0-9]+]]:vgpr(s96) = COPY $vgpr2_vgpr3_vgpr4
|
|
; GFX9: G_STORE [[COPY1]](s96), [[COPY]](p1) :: (store 12, align 16)
|
|
; GFX10-LABEL: name: store_flat_s96
|
|
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4
|
|
; GFX10: [[COPY:%[0-9]+]]:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
; GFX10: [[COPY1:%[0-9]+]]:vgpr(s96) = COPY $vgpr2_vgpr3_vgpr4
|
|
; GFX10: G_STORE [[COPY1]](s96), [[COPY]](p1) :: (store 12, align 16)
|
|
%0:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(s96) = COPY $vgpr2_vgpr3_vgpr4
|
|
G_STORE %1, %0 :: (store 12, align 16, addrspace 0)
|
|
|
|
...
|
|
---
|
|
|
|
name: store_flat_s128
|
|
legalized: true
|
|
tracksRegLiveness: true
|
|
regBankSelected: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
|
|
; GFX7-LABEL: name: store_flat_s128
|
|
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX7: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX7: [[COPY1:%[0-9]+]]:vreg_128 = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX7: FLAT_STORE_DWORDX4 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 16)
|
|
; GFX8-LABEL: name: store_flat_s128
|
|
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX8: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX8: [[COPY1:%[0-9]+]]:vreg_128 = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX8: FLAT_STORE_DWORDX4 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 16)
|
|
; GFX9-LABEL: name: store_flat_s128
|
|
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX9: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX9: [[COPY1:%[0-9]+]]:vreg_128 = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX9: FLAT_STORE_DWORDX4 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 16)
|
|
; GFX10-LABEL: name: store_flat_s128
|
|
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX10: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX10: [[COPY1:%[0-9]+]]:vreg_128 = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX10: FLAT_STORE_DWORDX4 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 16)
|
|
%0:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(s128) = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
G_STORE %1, %0 :: (store 16, align 16, addrspace 0)
|
|
|
|
...
|
|
|
|
---
|
|
|
|
name: store_flat_v2s32
|
|
legalized: true
|
|
tracksRegLiveness: true
|
|
regBankSelected: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
|
|
; GFX7-LABEL: name: store_flat_v2s32
|
|
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX7: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX7: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
|
|
; GFX7: FLAT_STORE_DWORDX2 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 8)
|
|
; GFX8-LABEL: name: store_flat_v2s32
|
|
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX8: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX8: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
|
|
; GFX8: FLAT_STORE_DWORDX2 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 8)
|
|
; GFX9-LABEL: name: store_flat_v2s32
|
|
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX9: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX9: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
|
|
; GFX9: FLAT_STORE_DWORDX2 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 8)
|
|
; GFX10-LABEL: name: store_flat_v2s32
|
|
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX10: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX10: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
|
|
; GFX10: FLAT_STORE_DWORDX2 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 8)
|
|
%0:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(<2 x s32>) = COPY $vgpr2_vgpr3
|
|
G_STORE %1, %0 :: (store 8, align 8, addrspace 0)
|
|
|
|
...
|
|
---
|
|
|
|
name: store_flat_v3s32
|
|
legalized: true
|
|
tracksRegLiveness: true
|
|
regBankSelected: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4
|
|
|
|
; GFX7-LABEL: name: store_flat_v3s32
|
|
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4
|
|
; GFX7: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX7: [[COPY1:%[0-9]+]]:vreg_96 = COPY $vgpr2_vgpr3_vgpr4
|
|
; GFX7: FLAT_STORE_DWORDX3 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 12, align 16)
|
|
; GFX8-LABEL: name: store_flat_v3s32
|
|
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4
|
|
; GFX8: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX8: [[COPY1:%[0-9]+]]:vreg_96 = COPY $vgpr2_vgpr3_vgpr4
|
|
; GFX8: FLAT_STORE_DWORDX3 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 12, align 16)
|
|
; GFX9-LABEL: name: store_flat_v3s32
|
|
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4
|
|
; GFX9: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX9: [[COPY1:%[0-9]+]]:vreg_96 = COPY $vgpr2_vgpr3_vgpr4
|
|
; GFX9: FLAT_STORE_DWORDX3 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 12, align 16)
|
|
; GFX10-LABEL: name: store_flat_v3s32
|
|
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4
|
|
; GFX10: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX10: [[COPY1:%[0-9]+]]:vreg_96 = COPY $vgpr2_vgpr3_vgpr4
|
|
; GFX10: FLAT_STORE_DWORDX3 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 12, align 16)
|
|
%0:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(<3 x s32>) = COPY $vgpr2_vgpr3_vgpr4
|
|
G_STORE %1, %0 :: (store 12, align 16, addrspace 0)
|
|
|
|
...
|
|
---
|
|
|
|
name: store_flat_v4s32
|
|
legalized: true
|
|
tracksRegLiveness: true
|
|
regBankSelected: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
|
|
; GFX7-LABEL: name: store_flat_v4s32
|
|
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX7: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX7: [[COPY1:%[0-9]+]]:vreg_128 = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX7: FLAT_STORE_DWORDX4 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 16)
|
|
; GFX8-LABEL: name: store_flat_v4s32
|
|
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX8: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX8: [[COPY1:%[0-9]+]]:vreg_128 = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX8: FLAT_STORE_DWORDX4 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 16)
|
|
; GFX9-LABEL: name: store_flat_v4s32
|
|
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX9: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX9: [[COPY1:%[0-9]+]]:vreg_128 = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX9: FLAT_STORE_DWORDX4 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 16)
|
|
; GFX10-LABEL: name: store_flat_v4s32
|
|
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX10: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX10: [[COPY1:%[0-9]+]]:vreg_128 = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX10: FLAT_STORE_DWORDX4 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 16)
|
|
%0:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(<4 x s32>) = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
G_STORE %1, %0 :: (store 16, align 16, addrspace 0)
|
|
|
|
...
|
|
|
|
---
|
|
|
|
name: store_flat_v2s16
|
|
legalized: true
|
|
tracksRegLiveness: true
|
|
regBankSelected: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2
|
|
|
|
; GFX7-LABEL: name: store_flat_v2s16
|
|
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX7: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX7: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX7: FLAT_STORE_DWORD [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 4)
|
|
; GFX8-LABEL: name: store_flat_v2s16
|
|
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX8: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX8: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX8: FLAT_STORE_DWORD [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 4)
|
|
; GFX9-LABEL: name: store_flat_v2s16
|
|
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX9: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX9: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX9: FLAT_STORE_DWORD [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 4)
|
|
; GFX10-LABEL: name: store_flat_v2s16
|
|
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX10: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX10: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX10: FLAT_STORE_DWORD [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 4)
|
|
%0:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(<2 x s16>) = COPY $vgpr2
|
|
G_STORE %1, %0 :: (store 4, align 4, addrspace 0)
|
|
|
|
...
|
|
|
|
---
|
|
|
|
name: store_flat_v4s16
|
|
legalized: true
|
|
tracksRegLiveness: true
|
|
regBankSelected: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
|
|
; GFX7-LABEL: name: store_flat_v4s16
|
|
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX7: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX7: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
|
|
; GFX7: FLAT_STORE_DWORDX2 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 8)
|
|
; GFX8-LABEL: name: store_flat_v4s16
|
|
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX8: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX8: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
|
|
; GFX8: FLAT_STORE_DWORDX2 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 8)
|
|
; GFX9-LABEL: name: store_flat_v4s16
|
|
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX9: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX9: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
|
|
; GFX9: FLAT_STORE_DWORDX2 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 8)
|
|
; GFX10-LABEL: name: store_flat_v4s16
|
|
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX10: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX10: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
|
|
; GFX10: FLAT_STORE_DWORDX2 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 8)
|
|
%0:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(<4 x s16>) = COPY $vgpr2_vgpr3
|
|
G_STORE %1, %0 :: (store 8, align 8, addrspace 0)
|
|
|
|
...
|
|
|
|
---
|
|
|
|
name: store_flat_v6s16
|
|
legalized: true
|
|
tracksRegLiveness: true
|
|
regBankSelected: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4
|
|
|
|
; GFX7-LABEL: name: store_flat_v6s16
|
|
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4
|
|
; GFX7: [[COPY:%[0-9]+]]:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
; GFX7: [[COPY1:%[0-9]+]]:vgpr(<6 x s16>) = COPY $vgpr2_vgpr3_vgpr4
|
|
; GFX7: G_STORE [[COPY1]](<6 x s16>), [[COPY]](p1) :: (store 12, align 16)
|
|
; GFX8-LABEL: name: store_flat_v6s16
|
|
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4
|
|
; GFX8: [[COPY:%[0-9]+]]:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
; GFX8: [[COPY1:%[0-9]+]]:vgpr(<6 x s16>) = COPY $vgpr2_vgpr3_vgpr4
|
|
; GFX8: G_STORE [[COPY1]](<6 x s16>), [[COPY]](p1) :: (store 12, align 16)
|
|
; GFX9-LABEL: name: store_flat_v6s16
|
|
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4
|
|
; GFX9: [[COPY:%[0-9]+]]:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
; GFX9: [[COPY1:%[0-9]+]]:vgpr(<6 x s16>) = COPY $vgpr2_vgpr3_vgpr4
|
|
; GFX9: G_STORE [[COPY1]](<6 x s16>), [[COPY]](p1) :: (store 12, align 16)
|
|
; GFX10-LABEL: name: store_flat_v6s16
|
|
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4
|
|
; GFX10: [[COPY:%[0-9]+]]:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
; GFX10: [[COPY1:%[0-9]+]]:vgpr(<6 x s16>) = COPY $vgpr2_vgpr3_vgpr4
|
|
; GFX10: G_STORE [[COPY1]](<6 x s16>), [[COPY]](p1) :: (store 12, align 16)
|
|
%0:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(<6 x s16>) = COPY $vgpr2_vgpr3_vgpr4
|
|
G_STORE %1, %0 :: (store 12, align 16, addrspace 0)
|
|
|
|
...
|
|
---
|
|
|
|
name: store_flat_v8s16
|
|
legalized: true
|
|
tracksRegLiveness: true
|
|
regBankSelected: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
|
|
; GFX7-LABEL: name: store_flat_v8s16
|
|
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX7: [[COPY:%[0-9]+]]:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
; GFX7: [[COPY1:%[0-9]+]]:vgpr(<8 x s16>) = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX7: G_STORE [[COPY1]](<8 x s16>), [[COPY]](p1) :: (store 16)
|
|
; GFX8-LABEL: name: store_flat_v8s16
|
|
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX8: [[COPY:%[0-9]+]]:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
; GFX8: [[COPY1:%[0-9]+]]:vgpr(<8 x s16>) = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX8: G_STORE [[COPY1]](<8 x s16>), [[COPY]](p1) :: (store 16)
|
|
; GFX9-LABEL: name: store_flat_v8s16
|
|
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX9: [[COPY:%[0-9]+]]:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
; GFX9: [[COPY1:%[0-9]+]]:vgpr(<8 x s16>) = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX9: G_STORE [[COPY1]](<8 x s16>), [[COPY]](p1) :: (store 16)
|
|
; GFX10-LABEL: name: store_flat_v8s16
|
|
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX10: [[COPY:%[0-9]+]]:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
; GFX10: [[COPY1:%[0-9]+]]:vgpr(<8 x s16>) = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX10: G_STORE [[COPY1]](<8 x s16>), [[COPY]](p1) :: (store 16)
|
|
%0:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(<8 x s16>) = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
G_STORE %1, %0 :: (store 16, align 16, addrspace 0)
|
|
|
|
...
|
|
|
|
---
|
|
|
|
name: store_flat_v2s64
|
|
legalized: true
|
|
tracksRegLiveness: true
|
|
regBankSelected: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
|
|
; GFX7-LABEL: name: store_flat_v2s64
|
|
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX7: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX7: [[COPY1:%[0-9]+]]:vreg_128 = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX7: FLAT_STORE_DWORDX4 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 16)
|
|
; GFX8-LABEL: name: store_flat_v2s64
|
|
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX8: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX8: [[COPY1:%[0-9]+]]:vreg_128 = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX8: FLAT_STORE_DWORDX4 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 16)
|
|
; GFX9-LABEL: name: store_flat_v2s64
|
|
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX9: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX9: [[COPY1:%[0-9]+]]:vreg_128 = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX9: FLAT_STORE_DWORDX4 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 16)
|
|
; GFX10-LABEL: name: store_flat_v2s64
|
|
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX10: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX10: [[COPY1:%[0-9]+]]:vreg_128 = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX10: FLAT_STORE_DWORDX4 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 16)
|
|
%0:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(<2 x s64>) = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
G_STORE %1, %0 :: (store 16, align 16, addrspace 0)
|
|
|
|
...
|
|
|
|
---
|
|
|
|
name: store_flat_p1
|
|
legalized: true
|
|
tracksRegLiveness: true
|
|
regBankSelected: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
|
|
; GFX7-LABEL: name: store_flat_p1
|
|
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX7: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX7: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
|
|
; GFX7: FLAT_STORE_DWORDX2 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 8)
|
|
; GFX8-LABEL: name: store_flat_p1
|
|
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX8: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX8: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
|
|
; GFX8: FLAT_STORE_DWORDX2 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 8)
|
|
; GFX9-LABEL: name: store_flat_p1
|
|
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX9: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX9: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
|
|
; GFX9: FLAT_STORE_DWORDX2 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 8)
|
|
; GFX10-LABEL: name: store_flat_p1
|
|
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX10: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX10: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
|
|
; GFX10: FLAT_STORE_DWORDX2 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 8)
|
|
%0:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(p1) = COPY $vgpr2_vgpr3
|
|
G_STORE %1, %0 :: (store 8, align 8, addrspace 0)
|
|
|
|
...
|
|
|
|
---
|
|
|
|
name: store_flat_v2p1
|
|
legalized: true
|
|
tracksRegLiveness: true
|
|
regBankSelected: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
|
|
; GFX7-LABEL: name: store_flat_v2p1
|
|
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX7: [[COPY:%[0-9]+]]:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
; GFX7: [[COPY1:%[0-9]+]]:vgpr(<2 x p1>) = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX7: G_STORE [[COPY1]](<2 x p1>), [[COPY]](p1) :: (store 16)
|
|
; GFX8-LABEL: name: store_flat_v2p1
|
|
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX8: [[COPY:%[0-9]+]]:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
; GFX8: [[COPY1:%[0-9]+]]:vgpr(<2 x p1>) = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX8: G_STORE [[COPY1]](<2 x p1>), [[COPY]](p1) :: (store 16)
|
|
; GFX9-LABEL: name: store_flat_v2p1
|
|
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX9: [[COPY:%[0-9]+]]:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
; GFX9: [[COPY1:%[0-9]+]]:vgpr(<2 x p1>) = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX9: G_STORE [[COPY1]](<2 x p1>), [[COPY]](p1) :: (store 16)
|
|
; GFX10-LABEL: name: store_flat_v2p1
|
|
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX10: [[COPY:%[0-9]+]]:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
; GFX10: [[COPY1:%[0-9]+]]:vgpr(<2 x p1>) = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
; GFX10: G_STORE [[COPY1]](<2 x p1>), [[COPY]](p1) :: (store 16)
|
|
%0:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(<2 x p1>) = COPY $vgpr2_vgpr3_vgpr4_vgpr5
|
|
G_STORE %1, %0 :: (store 16, align 16, addrspace 0)
|
|
|
|
...
|
|
|
|
---
|
|
|
|
name: store_flat_p3
|
|
legalized: true
|
|
tracksRegLiveness: true
|
|
regBankSelected: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2
|
|
|
|
; GFX7-LABEL: name: store_flat_p3
|
|
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX7: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX7: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX7: FLAT_STORE_DWORD [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 4)
|
|
; GFX8-LABEL: name: store_flat_p3
|
|
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX8: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX8: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX8: FLAT_STORE_DWORD [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 4)
|
|
; GFX9-LABEL: name: store_flat_p3
|
|
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX9: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX9: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX9: FLAT_STORE_DWORD [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 4)
|
|
; GFX10-LABEL: name: store_flat_p3
|
|
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX10: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX10: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX10: FLAT_STORE_DWORD [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 4)
|
|
%0:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(p3) = COPY $vgpr2
|
|
G_STORE %1, %0 :: (store 4, align 4, addrspace 0)
|
|
|
|
...
|
|
|
|
---
|
|
|
|
name: store_flat_v2p3
|
|
legalized: true
|
|
tracksRegLiveness: true
|
|
regBankSelected: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
|
|
; GFX7-LABEL: name: store_flat_v2p3
|
|
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX7: [[COPY:%[0-9]+]]:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
; GFX7: [[COPY1:%[0-9]+]]:vgpr(<2 x p3>) = COPY $vgpr2_vgpr3
|
|
; GFX7: G_STORE [[COPY1]](<2 x p3>), [[COPY]](p1) :: (store 8)
|
|
; GFX8-LABEL: name: store_flat_v2p3
|
|
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX8: [[COPY:%[0-9]+]]:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
; GFX8: [[COPY1:%[0-9]+]]:vgpr(<2 x p3>) = COPY $vgpr2_vgpr3
|
|
; GFX8: G_STORE [[COPY1]](<2 x p3>), [[COPY]](p1) :: (store 8)
|
|
; GFX9-LABEL: name: store_flat_v2p3
|
|
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX9: [[COPY:%[0-9]+]]:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
; GFX9: [[COPY1:%[0-9]+]]:vgpr(<2 x p3>) = COPY $vgpr2_vgpr3
|
|
; GFX9: G_STORE [[COPY1]](<2 x p3>), [[COPY]](p1) :: (store 8)
|
|
; GFX10-LABEL: name: store_flat_v2p3
|
|
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX10: [[COPY:%[0-9]+]]:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
; GFX10: [[COPY1:%[0-9]+]]:vgpr(<2 x p3>) = COPY $vgpr2_vgpr3
|
|
; GFX10: G_STORE [[COPY1]](<2 x p3>), [[COPY]](p1) :: (store 8)
|
|
%0:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(<2 x p3>) = COPY $vgpr2_vgpr3
|
|
G_STORE %1, %0 :: (store 8, align 8, addrspace 0)
|
|
|
|
...
|
|
---
|
|
|
|
name: store_atomic_flat_s32
|
|
legalized: true
|
|
tracksRegLiveness: true
|
|
regBankSelected: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2
|
|
|
|
; GFX7-LABEL: name: store_atomic_flat_s32
|
|
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX7: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX7: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX7: FLAT_STORE_DWORD [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store monotonic 4)
|
|
; GFX8-LABEL: name: store_atomic_flat_s32
|
|
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX8: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX8: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX8: FLAT_STORE_DWORD [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store monotonic 4)
|
|
; GFX9-LABEL: name: store_atomic_flat_s32
|
|
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX9: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX9: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX9: FLAT_STORE_DWORD [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store monotonic 4)
|
|
; GFX10-LABEL: name: store_atomic_flat_s32
|
|
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX10: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX10: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX10: FLAT_STORE_DWORD [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store monotonic 4)
|
|
%0:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(s32) = COPY $vgpr2
|
|
G_STORE %1, %0 :: (store monotonic 4, align 4, addrspace 0)
|
|
|
|
...
|
|
|
|
---
|
|
|
|
name: store_atomic_flat_s64
|
|
legalized: true
|
|
tracksRegLiveness: true
|
|
regBankSelected: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
|
|
; GFX7-LABEL: name: store_atomic_flat_s64
|
|
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX7: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX7: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
|
|
; GFX7: FLAT_STORE_DWORDX2 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store monotonic 8)
|
|
; GFX8-LABEL: name: store_atomic_flat_s64
|
|
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX8: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX8: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
|
|
; GFX8: FLAT_STORE_DWORDX2 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store monotonic 8)
|
|
; GFX9-LABEL: name: store_atomic_flat_s64
|
|
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX9: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX9: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
|
|
; GFX9: FLAT_STORE_DWORDX2 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store monotonic 8)
|
|
; GFX10-LABEL: name: store_atomic_flat_s64
|
|
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; GFX10: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX10: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
|
|
; GFX10: FLAT_STORE_DWORDX2 [[COPY]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store monotonic 8)
|
|
%0:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(s64) = COPY $vgpr2_vgpr3
|
|
G_STORE %1, %0 :: (store monotonic 8, align 8, addrspace 0)
|
|
|
|
...
|
|
|
|
---
|
|
|
|
name: store_flat_s32_gep_2047
|
|
legalized: true
|
|
tracksRegLiveness: true
|
|
regBankSelected: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2
|
|
|
|
; GFX7-LABEL: name: store_flat_s32_gep_2047
|
|
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX7: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX7: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX7: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 2047, implicit $exec
|
|
; GFX7: [[V_MOV_B32_e32_1:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
|
|
; GFX7: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[V_MOV_B32_e32_]], %subreg.sub0, [[V_MOV_B32_e32_1]], %subreg.sub1
|
|
; GFX7: [[COPY2:%[0-9]+]]:vgpr_32 = COPY [[COPY]].sub0
|
|
; GFX7: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE]].sub0
|
|
; GFX7: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[COPY]].sub1
|
|
; GFX7: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE]].sub1
|
|
; GFX7: [[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 [[COPY2]], [[COPY3]], 0, implicit $exec
|
|
; GFX7: %9:vgpr_32, dead %11:sreg_64_xexec = V_ADDC_U32_e64 [[COPY4]], [[COPY5]], killed [[V_ADD_CO_U32_e64_1]], 0, implicit $exec
|
|
; GFX7: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[V_ADD_CO_U32_e64_]], %subreg.sub0, %9, %subreg.sub1
|
|
; GFX7: FLAT_STORE_DWORD [[REG_SEQUENCE1]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 4)
|
|
; GFX8-LABEL: name: store_flat_s32_gep_2047
|
|
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX8: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX8: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX8: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 2047, implicit $exec
|
|
; GFX8: [[V_MOV_B32_e32_1:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
|
|
; GFX8: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[V_MOV_B32_e32_]], %subreg.sub0, [[V_MOV_B32_e32_1]], %subreg.sub1
|
|
; GFX8: [[COPY2:%[0-9]+]]:vgpr_32 = COPY [[COPY]].sub0
|
|
; GFX8: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE]].sub0
|
|
; GFX8: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[COPY]].sub1
|
|
; GFX8: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE]].sub1
|
|
; GFX8: [[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 [[COPY2]], [[COPY3]], 0, implicit $exec
|
|
; GFX8: %9:vgpr_32, dead %11:sreg_64_xexec = V_ADDC_U32_e64 [[COPY4]], [[COPY5]], killed [[V_ADD_CO_U32_e64_1]], 0, implicit $exec
|
|
; GFX8: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[V_ADD_CO_U32_e64_]], %subreg.sub0, %9, %subreg.sub1
|
|
; GFX8: FLAT_STORE_DWORD [[REG_SEQUENCE1]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 4)
|
|
; GFX9-LABEL: name: store_flat_s32_gep_2047
|
|
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX9: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX9: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX9: FLAT_STORE_DWORD [[COPY]], [[COPY1]], 2047, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 4)
|
|
; GFX10-LABEL: name: store_flat_s32_gep_2047
|
|
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2
|
|
; GFX10: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
|
|
; GFX10: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX10: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 2047, implicit $exec
|
|
; GFX10: [[V_MOV_B32_e32_1:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
|
|
; GFX10: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[V_MOV_B32_e32_]], %subreg.sub0, [[V_MOV_B32_e32_1]], %subreg.sub1
|
|
; GFX10: [[COPY2:%[0-9]+]]:vgpr_32 = COPY [[COPY]].sub0
|
|
; GFX10: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE]].sub0
|
|
; GFX10: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[COPY]].sub1
|
|
; GFX10: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE]].sub1
|
|
; GFX10: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_ADD_CO_U32_e64 [[COPY2]], [[COPY3]], 0, implicit $exec
|
|
; GFX10: %9:vgpr_32, dead %11:sreg_32_xm0_xexec = V_ADDC_U32_e64 [[COPY4]], [[COPY5]], killed [[V_ADD_CO_U32_e64_1]], 0, implicit $exec
|
|
; GFX10: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[V_ADD_CO_U32_e64_]], %subreg.sub0, %9, %subreg.sub1
|
|
; GFX10: FLAT_STORE_DWORD [[REG_SEQUENCE1]], [[COPY1]], 0, 0, 0, 0, implicit $exec, implicit $flat_scr :: (store 4)
|
|
%0:vgpr(p1) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(s32) = COPY $vgpr2
|
|
%2:vgpr(s64) = G_CONSTANT i64 2047
|
|
%3:vgpr(p1) = G_PTR_ADD %0, %2
|
|
G_STORE %1, %3 :: (store 4, align 4, addrspace 0)
|
|
|
|
...
|