Skip to content

[AMDGPU] Add aperture classes to VS_64 #158823

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Sep 16, 2025
Merged

[AMDGPU] Add aperture classes to VS_64 #158823

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 5 additions & 3 deletions llvm/lib/Target/AMDGPU/SIRegisterInfo.td
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1131,15 +1131,16 @@ def VS_32_Lo256 : SIRegisterClass<"AMDGPU", [i32, f32, i16, f16, bf16, v2i16, v2
let Size = 32;
}

def VS_64 : SIRegisterClass<"AMDGPU", VReg_64.RegTypes, 32, (add VReg_64, SReg_64)> {
def VS_64 : SIRegisterClass<"AMDGPU", VReg_64.RegTypes, 32,
(add VReg_64, SReg_64_Encodable)> {
let isAllocatable = 0;
let HasVGPR = 1;
let HasSGPR = 1;
let Size = 64;
}

def VS_64_Align2 : SIRegisterClass<"AMDGPU", VReg_64.RegTypes, 32,
(add VReg_64_Align2, SReg_64)> {
(add VReg_64_Align2, SReg_64_Encodable)> {
let isAllocatable = 0;
let HasVGPR = 1;
let HasSGPR = 1;
Expand All @@ -1153,7 +1154,8 @@ def AV_32 : SIRegisterClass<"AMDGPU", VGPR_32.RegTypes, 32, (add VGPR_32, AGPR_3
let Size = 32;
}

def VS_64_Lo256 : SIRegisterClass<"AMDGPU", VReg_64.RegTypes, 32, (add VReg_64_Lo256_Align2, SReg_64)> {
def VS_64_Lo256 : SIRegisterClass<"AMDGPU", VReg_64.RegTypes, 32,
(add VReg_64_Lo256_Align2, SReg_64_Encodable)> {
let isAllocatable = 0;
let HasVGPR = 1;
let HasSGPR = 1;
Expand Down
2 changes: 1 addition & 1 deletion llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-inline-asm.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -136,7 +136,7 @@ define float @test_multiple_register_outputs_same() #0 {
define double @test_multiple_register_outputs_mixed() #0 {
; CHECK-LABEL: name: test_multiple_register_outputs_mixed
; CHECK: bb.1 (%ir-block.0):
; CHECK-NEXT: INLINEASM &"v_mov_b32 $0, 0; v_add_f64 $1, 0, 0", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def %8, 3670026 /* regdef:VReg_64 */, def %9
; CHECK-NEXT: INLINEASM &"v_mov_b32 $0, 0; v_add_f64 $1, 0, 0", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def %8, 3735562 /* regdef:VReg_64 */, def %9
; CHECK-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY %8
; CHECK-NEXT: [[COPY1:%[0-9]+]]:_(s64) = COPY %9
; CHECK-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s64)
Expand Down
230 changes: 115 additions & 115 deletions llvm/test/CodeGen/AMDGPU/coalesce-copy-to-agpr-to-av-registers.mir
View file
Open in desktop

Large diffs are not rendered by default.

12 changes: 6 additions & 6 deletions llvm/test/CodeGen/AMDGPU/inflate-reg-class-vgpr-mfma-to-av-with-load-source.mir
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -486,7 +486,7 @@ body: |
; CHECK-NEXT: S_NOP 0, implicit-def $vgpr48_vgpr49_vgpr50_vgpr51_vgpr52_vgpr53_vgpr54_vgpr55
; CHECK-NEXT: S_NOP 0, implicit-def $vgpr56_vgpr57_vgpr58_vgpr59_vgpr60_vgpr61_vgpr62_vgpr63
; CHECK-NEXT: renamable $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15 = COPY killed renamable $agpr0_agpr1_agpr2_agpr3_agpr4_agpr5_agpr6_agpr7_agpr8_agpr9_agpr10_agpr11_agpr12_agpr13_agpr14_agpr15
; CHECK-NEXT: INLINEASM &"; use $0 ", 1 /* sideeffect attdialect */, 38797321 /* reguse:VReg_512_Align2 */, killed renamable $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15
; CHECK-NEXT: INLINEASM &"; use $0 ", 1 /* sideeffect attdialect */, 39190537 /* reguse:VReg_512_Align2 */, killed renamable $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15
; CHECK-NEXT: S_ENDPGM 0
bb.0:
S_NOP 0, implicit-def $agpr0
Expand Down Expand Up @@ -516,7 +516,7 @@ body: |
S_NOP 0, implicit-def $vgpr40_vgpr41_vgpr42_vgpr43_vgpr44_vgpr45_vgpr46_vgpr47
S_NOP 0, implicit-def $vgpr48_vgpr49_vgpr50_vgpr51_vgpr52_vgpr53_vgpr54_vgpr55
S_NOP 0, implicit-def $vgpr56_vgpr57_vgpr58_vgpr59_vgpr60_vgpr61_vgpr62_vgpr63
INLINEASM &"; use $0 ", 1 /* sideeffect attdialect */, 38797321 /* reguse:VReg_512_Align2 */, %0:vreg_512_align2
INLINEASM &"; use $0 ", 1 /* sideeffect attdialect */, 39190537 /* reguse:VReg_512_Align2 */, %0:vreg_512_align2
S_ENDPGM 0

...
Expand Down Expand Up @@ -1368,7 +1368,7 @@ body: |
; CHECK-NEXT: renamable $vgpr0_vgpr1 = GLOBAL_LOAD_DWORDX2 undef renamable $vgpr0_vgpr1, 0, 0, implicit $exec :: (load (s64), addrspace 1)
; CHECK-NEXT: early-clobber renamable $vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31_vgpr32_vgpr33 = V_MFMA_F32_32X32X8F16_vgprcd_e64 $vgpr16_vgpr17, $vgpr16_vgpr17, $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15, 0, 0, 0, implicit $mode, implicit $exec
; CHECK-NEXT: early-clobber renamable $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15 = V_MFMA_F32_32X32X8F16_vgprcd_e64 $vgpr16_vgpr17, $vgpr16_vgpr17, $vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31_vgpr32_vgpr33, 0, 0, 0, implicit $mode, implicit $exec
; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 38797321 /* reguse:VReg_512_Align2 */, killed renamable $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15
; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 39190537 /* reguse:VReg_512_Align2 */, killed renamable $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15
; CHECK-NEXT: S_CBRANCH_VCCNZ %bb.1, implicit $vcc
; CHECK-NEXT: S_BRANCH %bb.2
; CHECK-NEXT: {{ $}}
Expand Down Expand Up @@ -1408,7 +1408,7 @@ body: |
undef %2.sub0_sub1:vreg_512_align2 = GLOBAL_LOAD_DWORDX2 undef %3:vreg_64_align2, 0, 0, implicit $exec :: (load (s64), addrspace 1)
early-clobber %0:vreg_512_align2 = V_MFMA_F32_32X32X8F16_vgprcd_e64 %1, %1, %2, 0, 0, 0, implicit $mode, implicit $exec
early-clobber %4:vreg_512_align2 = V_MFMA_F32_32X32X8F16_vgprcd_e64 %1, %1, %0, 0, 0, 0, implicit $mode, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 38797321 /* reguse:VReg_512_Align2 */, %4
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 39190537 /* reguse:VReg_512_Align2 */, %4
S_CBRANCH_VCCNZ %bb.1, implicit $vcc
S_BRANCH %bb.2

Expand Down Expand Up @@ -1726,7 +1726,7 @@ body: |
; CHECK-NEXT: renamable $vgpr0_vgpr1 = GLOBAL_LOAD_DWORDX2 undef renamable $vgpr0_vgpr1, 0, 0, implicit $exec :: (load (s64), addrspace 1)
; CHECK-NEXT: renamable $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15 = V_MFMA_F32_32X32X8F16_mac_vgprcd_e64 $vgpr16_vgpr17, $vgpr16_vgpr17, $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15, 0, 0, 0, implicit $mode, implicit $exec
; CHECK-NEXT: early-clobber renamable $vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31_vgpr32_vgpr33 = V_MFMA_F32_32X32X8F16_vgprcd_e64 $vgpr16_vgpr17, $vgpr16_vgpr17, killed $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15, 0, 0, 0, implicit $mode, implicit $exec
; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 38797321 /* reguse:VReg_512_Align2 */, renamable $vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31_vgpr32_vgpr33
; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 39190537 /* reguse:VReg_512_Align2 */, renamable $vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31_vgpr32_vgpr33
; CHECK-NEXT: S_CBRANCH_VCCNZ %bb.1, implicit $vcc
; CHECK-NEXT: S_BRANCH %bb.2
; CHECK-NEXT: {{ $}}
Expand Down Expand Up @@ -1763,7 +1763,7 @@ body: |
undef %0.sub0_sub1:vreg_512_align2 = GLOBAL_LOAD_DWORDX2 undef %3:vreg_64_align2, 0, 0, implicit $exec :: (load (s64), addrspace 1)
%0:vreg_512_align2 = V_MFMA_F32_32X32X8F16_mac_vgprcd_e64 %1, %1, %0, 0, 0, 0, implicit $mode, implicit $exec
%4:vreg_512_align2 = V_MFMA_F32_32X32X8F16_vgprcd_e64 %1, %1, %0, 0, 0, 0, implicit $mode, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 38797321 /* reguse:VReg_512_Align2 */, %4
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 39190537 /* reguse:VReg_512_Align2 */, %4
S_CBRANCH_VCCNZ %bb.1, implicit $vcc
S_BRANCH %bb.2

Expand Down
24 changes: 12 additions & 12 deletions llvm/test/CodeGen/AMDGPU/inline-asm.i128.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,16 +8,16 @@
define amdgpu_kernel void @s_input_output_i128() {
; GFX908-LABEL: name: s_input_output_i128
; GFX908: bb.0 (%ir-block.0):
; GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 9240586 /* regdef:SGPR_128 */, def %13
; GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 9633802 /* regdef:SGPR_128 */, def %13
; GFX908-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY %13
; GFX908-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9240585 /* reguse:SGPR_128 */, [[COPY]]
; GFX908-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9633801 /* reguse:SGPR_128 */, [[COPY]]
; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: s_input_output_i128
; GFX90A: bb.0 (%ir-block.0):
; GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 9240586 /* regdef:SGPR_128 */, def %11
; GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 9633802 /* regdef:SGPR_128 */, def %11
; GFX90A-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY %11
; GFX90A-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9240585 /* reguse:SGPR_128 */, [[COPY]]
; GFX90A-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9633801 /* reguse:SGPR_128 */, [[COPY]]
; GFX90A-NEXT: S_ENDPGM 0
%val = tail call i128 asm sideeffect "; def $0", "=s"()
call void asm sideeffect "; use $0", "s"(i128 %val)
Expand All @@ -27,16 +27,16 @@ define amdgpu_kernel void @s_input_output_i128() {
define amdgpu_kernel void @v_input_output_i128() {
; GFX908-LABEL: name: v_input_output_i128
; GFX908: bb.0 (%ir-block.0):
; GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7405578 /* regdef:VReg_128 */, def %13
; GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7798794 /* regdef:VReg_128 */, def %13
; GFX908-NEXT: [[COPY:%[0-9]+]]:vreg_128 = COPY %13
; GFX908-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 7405577 /* reguse:VReg_128 */, [[COPY]]
; GFX908-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 7798793 /* reguse:VReg_128 */, [[COPY]]
; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: v_input_output_i128
; GFX90A: bb.0 (%ir-block.0):
; GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7602186 /* regdef:VReg_128_Align2 */, def %11
; GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7995402 /* regdef:VReg_128_Align2 */, def %11
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vreg_128_align2 = COPY %11
; GFX90A-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 7602185 /* reguse:VReg_128_Align2 */, [[COPY]]
; GFX90A-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 7995401 /* reguse:VReg_128_Align2 */, [[COPY]]
; GFX90A-NEXT: S_ENDPGM 0
%val = tail call i128 asm sideeffect "; def $0", "=v"()
call void asm sideeffect "; use $0", "v"(i128 %val)
Expand All @@ -47,16 +47,16 @@ define amdgpu_kernel void @a_input_output_i128() {

; GFX908-LABEL: name: a_input_output_i128
; GFX908: bb.0 (%ir-block.0):
; GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7929866 /* regdef:AReg_128 */, def %13
; GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 8323082 /* regdef:AReg_128 */, def %13
; GFX908-NEXT: [[COPY:%[0-9]+]]:areg_128 = COPY %13
; GFX908-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 7929865 /* reguse:AReg_128 */, [[COPY]]
; GFX908-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 8323081 /* reguse:AReg_128 */, [[COPY]]
; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: a_input_output_i128
; GFX90A: bb.0 (%ir-block.0):
; GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 8257546 /* regdef:AReg_128_Align2 */, def %11
; GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 8650762 /* regdef:AReg_128_Align2 */, def %11
; GFX90A-NEXT: [[COPY:%[0-9]+]]:areg_128_align2 = COPY %11
; GFX90A-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 8257545 /* reguse:AReg_128_Align2 */, [[COPY]]
; GFX90A-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 8650761 /* reguse:AReg_128_Align2 */, [[COPY]]
; GFX90A-NEXT: S_ENDPGM 0
%val = call i128 asm sideeffect "; def $0", "=a"()
call void asm sideeffect "; use $0", "a"(i128 %val)
Expand Down
16 changes: 8 additions & 8 deletions llvm/test/CodeGen/AMDGPU/partial-regcopy-and-spill-missed-at-regalloc.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,9 +12,9 @@ define amdgpu_kernel void @partial_copy(<4 x i32> %arg) #0 {
; REGALLOC-GFX908-NEXT: liveins: $sgpr4_sgpr5
; REGALLOC-GFX908-NEXT: {{ $}}
; REGALLOC-GFX908-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:AGPR_32 */, undef %6:agpr_32
; REGALLOC-GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7405578 /* regdef:VReg_128 */, def %25
; REGALLOC-GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7798794 /* regdef:VReg_128 */, def %25
; REGALLOC-GFX908-NEXT: [[COPY:%[0-9]+]]:av_128 = COPY %25
; REGALLOC-GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3670026 /* regdef:VReg_64 */, def %27
; REGALLOC-GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3735562 /* regdef:VReg_64 */, def %27
; REGALLOC-GFX908-NEXT: SI_SPILL_AV64_SAVE %27, %stack.0, $sgpr32, 0, implicit $exec :: (store (s64) into %stack.0, align 4, addrspace 5)
; REGALLOC-GFX908-NEXT: [[COPY1:%[0-9]+]]:vreg_128 = COPY [[COPY]]
; REGALLOC-GFX908-NEXT: GLOBAL_STORE_DWORDX4 undef %15:vreg_64, [[COPY1]], 0, 0, implicit $exec :: (volatile store (s128) into `ptr addrspace(1) poison`, addrspace 1)
Expand All @@ -37,9 +37,9 @@ define amdgpu_kernel void @partial_copy(<4 x i32> %arg) #0 {
; PEI-GFX908-NEXT: $sgpr12 = S_ADD_U32 $sgpr12, $sgpr9, implicit-def $scc, implicit-def $sgpr12_sgpr13_sgpr14_sgpr15
; PEI-GFX908-NEXT: $sgpr13 = S_ADDC_U32 $sgpr13, 0, implicit-def dead $scc, implicit $scc, implicit-def $sgpr12_sgpr13_sgpr14_sgpr15
; PEI-GFX908-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:AGPR_32 */, undef renamable $agpr0
; PEI-GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7405578 /* regdef:VReg_128 */, def renamable $vgpr0_vgpr1_vgpr2_vgpr3
; PEI-GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7798794 /* regdef:VReg_128 */, def renamable $vgpr0_vgpr1_vgpr2_vgpr3
; PEI-GFX908-NEXT: renamable $agpr0_agpr1_agpr2_agpr3 = COPY killed renamable $vgpr0_vgpr1_vgpr2_vgpr3, implicit $exec
; PEI-GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3670026 /* regdef:VReg_64 */, def renamable $vgpr0_vgpr1
; PEI-GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3735562 /* regdef:VReg_64 */, def renamable $vgpr0_vgpr1
; PEI-GFX908-NEXT: BUFFER_STORE_DWORD_OFFSET killed $vgpr0, $sgpr12_sgpr13_sgpr14_sgpr15, 0, 0, 0, 0, implicit $exec, implicit-def $vgpr0_vgpr1, implicit $vgpr0_vgpr1 :: (store (s32) into %stack.0, addrspace 5)
; PEI-GFX908-NEXT: $agpr4 = V_ACCVGPR_WRITE_B32_e64 killed $vgpr1, implicit $exec, implicit killed $vgpr0_vgpr1
; PEI-GFX908-NEXT: renamable $vgpr0_vgpr1_vgpr2_vgpr3 = COPY killed renamable $agpr0_agpr1_agpr2_agpr3, implicit $exec
Expand All @@ -61,9 +61,9 @@ define amdgpu_kernel void @partial_copy(<4 x i32> %arg) #0 {
; REGALLOC-GFX90A-NEXT: liveins: $sgpr4_sgpr5
; REGALLOC-GFX90A-NEXT: {{ $}}
; REGALLOC-GFX90A-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:AGPR_32 */, undef %6:agpr_32
; REGALLOC-GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7602186 /* regdef:VReg_128_Align2 */, def %23
; REGALLOC-GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7995402 /* regdef:VReg_128_Align2 */, def %23
; REGALLOC-GFX90A-NEXT: [[COPY:%[0-9]+]]:av_128_align2 = COPY %23
; REGALLOC-GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3866634 /* regdef:VReg_64_Align2 */, def %21
; REGALLOC-GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3997706 /* regdef:VReg_64_Align2 */, def %21
; REGALLOC-GFX90A-NEXT: [[COPY1:%[0-9]+]]:av_64_align2 = COPY %21
; REGALLOC-GFX90A-NEXT: GLOBAL_STORE_DWORDX4 undef %15:vreg_64_align2, [[COPY]], 0, 0, implicit $exec :: (volatile store (s128) into `ptr addrspace(1) poison`, addrspace 1)
; REGALLOC-GFX90A-NEXT: renamable $sgpr0_sgpr1_sgpr2_sgpr3 = S_LOAD_DWORDX4_IMM killed renamable $sgpr4_sgpr5, 0, 0 :: (dereferenceable invariant load (s128) from %ir.arg.kernarg.offset1, addrspace 4)
Expand All @@ -80,9 +80,9 @@ define amdgpu_kernel void @partial_copy(<4 x i32> %arg) #0 {
; PEI-GFX90A-NEXT: liveins: $sgpr4_sgpr5
; PEI-GFX90A-NEXT: {{ $}}
; PEI-GFX90A-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:AGPR_32 */, undef renamable $agpr0
; PEI-GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7602186 /* regdef:VReg_128_Align2 */, def renamable $vgpr0_vgpr1_vgpr2_vgpr3
; PEI-GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7995402 /* regdef:VReg_128_Align2 */, def renamable $vgpr0_vgpr1_vgpr2_vgpr3
; PEI-GFX90A-NEXT: renamable $agpr0_agpr1_agpr2_agpr3 = COPY killed renamable $vgpr0_vgpr1_vgpr2_vgpr3, implicit $exec
; PEI-GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3866634 /* regdef:VReg_64_Align2 */, def renamable $vgpr2_vgpr3
; PEI-GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3997706 /* regdef:VReg_64_Align2 */, def renamable $vgpr2_vgpr3
; PEI-GFX90A-NEXT: GLOBAL_STORE_DWORDX4 undef renamable $vgpr0_vgpr1, killed renamable $agpr0_agpr1_agpr2_agpr3, 0, 0, implicit $exec :: (volatile store (s128) into `ptr addrspace(1) poison`, addrspace 1)
; PEI-GFX90A-NEXT: renamable $sgpr0_sgpr1_sgpr2_sgpr3 = S_LOAD_DWORDX4_IMM killed renamable $sgpr4_sgpr5, 0, 0 :: (dereferenceable invariant load (s128) from %ir.arg.kernarg.offset1, addrspace 4)
; PEI-GFX90A-NEXT: renamable $agpr0_agpr1_agpr2_agpr3 = COPY killed renamable $sgpr0_sgpr1_sgpr2_sgpr3, implicit $exec
Expand Down
8 changes: 4 additions & 4 deletions llvm/test/CodeGen/AMDGPU/rewrite-vgpr-mfma-to-agpr-copy-from.mir
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -41,17 +41,17 @@ body: |
; CHECK-NEXT: [[COPY1:%[0-9]+]]:av_64_align2 = COPY $vgpr0_vgpr1
; CHECK-NEXT: [[COPY2:%[0-9]+]]:av_64_align2 = COPY $vgpr2_vgpr3
; CHECK-NEXT: [[GLOBAL_LOAD_DWORDX4_:%[0-9]+]]:areg_128_align2 = GLOBAL_LOAD_DWORDX4 [[COPY]], 0, 0, implicit $exec :: (load (s128), addrspace 1)
; CHECK-NEXT: [[COPY3:%[0-9]+]]:vreg_128_align2 = COPY [[GLOBAL_LOAD_DWORDX4_]]
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_vgprcd_e64_:%[0-9]+]]:vreg_64_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 [[COPY1]], [[COPY2]], [[COPY3]].sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 3866633 /* reguse:VReg_64_Align2 */, [[V_MFMA_F64_4X4X4F64_vgprcd_e64_]]
; CHECK-NEXT: [[COPY3:%[0-9]+]]:areg_128_align2 = COPY [[GLOBAL_LOAD_DWORDX4_]]
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]]:areg_64_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[COPY3]].sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 3801097 /* reguse:AV_64_Align2 */, [[V_MFMA_F64_4X4X4F64_e64_]]
; CHECK-NEXT: SI_RETURN
%0:vreg_64_align2 = COPY $vgpr4_vgpr5
%1:av_64_align2 = COPY $vgpr0_vgpr1
%2:av_64_align2 = COPY $vgpr2_vgpr3
%3:areg_128_align2 = GLOBAL_LOAD_DWORDX4 %0, 0, 0, implicit $exec :: (load (s128), addrspace 1)
%4:vreg_128_align2 = COPY %3
%5:vreg_64_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %4.sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 3866633 /* reguse:VReg_64_Align2 */, %5
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 3801097 /* reguse:VReg_64_Align2 */, %5
SI_RETURN
...

Expand Down
Loading