Skip to content

AMDGPU: Replace insertelement undef with poison in cases with manual updates #130898

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
Mar 12, 2025
Merged

AMDGPU: Replace insertelement undef with poison in cases with manual updates #130898

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
16 changes: 8 additions & 8 deletions llvm/test/CodeGen/AMDGPU/fcanonicalize.f16.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -238,7 +238,7 @@ define <2 x half> @v_test_canonicalize_build_vector_v2f16(half %lo, half %hi) #1
; GFX11-FAKE16-NEXT: s_delay_alu instid0(VALU_DEP_1)
; GFX11-FAKE16-NEXT: v_pk_max_f16 v0, v0, v0
; GFX11-FAKE16-NEXT: s_setpc_b64 s[30:31]
%ins0 = insertelement <2 x half> undef, half %lo, i32 0
%ins0 = insertelement <2 x half> poison, half %lo, i32 0
%ins1 = insertelement <2 x half> %ins0, half %hi, i32 1
%canonicalized = call <2 x half> @llvm.canonicalize.v2f16(<2 x half> %ins1)
ret <2 x half> %canonicalized
Expand Down Expand Up @@ -2581,7 +2581,7 @@ define <2 x half> @v_test_canonicalize_reg_undef_v2f16(half %val) #1 {
; GFX11-FAKE16-NEXT: s_delay_alu instid0(VALU_DEP_1)
; GFX11-FAKE16-NEXT: v_pack_b32_f16 v0, v0, 0
; GFX11-FAKE16-NEXT: s_setpc_b64 s[30:31]
%vec = insertelement <2 x half> undef, half %val, i32 0
%vec = insertelement <2 x half> poison, half %val, i32 0
%canonicalized = call <2 x half> @llvm.canonicalize.v2f16(<2 x half> %vec)
ret <2 x half> %canonicalized
}
Expand Down Expand Up @@ -2622,7 +2622,7 @@ define <2 x half> @v_test_canonicalize_undef_reg_v2f16(half %val) #1 {
; GFX11-FAKE16-NEXT: s_delay_alu instid0(VALU_DEP_1)
; GFX11-FAKE16-NEXT: v_lshlrev_b32_e32 v0, 16, v0
; GFX11-FAKE16-NEXT: s_setpc_b64 s[30:31]
%vec = insertelement <2 x half> undef, half %val, i32 1
%vec = insertelement <2 x half> poison, half %val, i32 1
%canonicalized = call <2 x half> @llvm.canonicalize.v2f16(<2 x half> %vec)
ret <2 x half> %canonicalized
}
Expand Down Expand Up @@ -2785,7 +2785,7 @@ define <2 x half> @v_test_canonicalize_reg_k_v2f16(half %val) #1 {
; GFX11-FAKE16-NEXT: s_delay_alu instid0(VALU_DEP_1)
; GFX11-FAKE16-NEXT: v_pack_b32_f16 v0, v0, 2.0
; GFX11-FAKE16-NEXT: s_setpc_b64 s[30:31]
%vec0 = insertelement <2 x half> undef, half %val, i32 0
%vec0 = insertelement <2 x half> poison, half %val, i32 0
%vec1 = insertelement <2 x half> %vec0, half 2.0, i32 1
%canonicalized = call <2 x half> @llvm.canonicalize.v2f16(<2 x half> %vec1)
ret <2 x half> %canonicalized
Expand Down Expand Up @@ -2829,7 +2829,7 @@ define <2 x half> @v_test_canonicalize_k_reg_v2f16(half %val) #1 {
; GFX11-FAKE16-NEXT: s_delay_alu instid0(VALU_DEP_1)
; GFX11-FAKE16-NEXT: v_pack_b32_f16 v0, 2.0, v0
; GFX11-FAKE16-NEXT: s_setpc_b64 s[30:31]
%vec0 = insertelement <2 x half> undef, half 2.0, i32 0
%vec0 = insertelement <2 x half> poison, half 2.0, i32 0
%vec1 = insertelement <2 x half> %vec0, half %val, i32 1
%canonicalized = call <2 x half> @llvm.canonicalize.v2f16(<2 x half> %vec1)
ret <2 x half> %canonicalized
Expand Down Expand Up @@ -2925,7 +2925,7 @@ define <4 x half> @v_test_canonicalize_reg_undef_undef_undef_v4f16(half %val) #1
; GFX11-FAKE16-NEXT: s_delay_alu instid0(VALU_DEP_2)
; GFX11-FAKE16-NEXT: v_pack_b32_f16 v0, v0, 0
; GFX11-FAKE16-NEXT: s_setpc_b64 s[30:31]
%vec = insertelement <4 x half> undef, half %val, i32 0
%vec = insertelement <4 x half> poison, half %val, i32 0
%canonicalized = call <4 x half> @llvm.canonicalize.v4f16(<4 x half> %vec)
ret <4 x half> %canonicalized
}
Expand Down Expand Up @@ -2977,7 +2977,7 @@ define <4 x half> @v_test_canonicalize_reg_reg_undef_undef_v4f16(half %val0, hal
; GFX11-FAKE16-NEXT: s_delay_alu instid0(VALU_DEP_2)
; GFX11-FAKE16-NEXT: v_pk_max_f16 v0, v0, v0
; GFX11-FAKE16-NEXT: s_setpc_b64 s[30:31]
%vec0 = insertelement <4 x half> undef, half %val0, i32 0
%vec0 = insertelement <4 x half> poison, half %val0, i32 0
%vec1 = insertelement <4 x half> %vec0, half %val1, i32 1
%canonicalized = call <4 x half> @llvm.canonicalize.v4f16(<4 x half> %vec1)
ret <4 x half> %canonicalized
Expand Down Expand Up @@ -3035,7 +3035,7 @@ define <4 x half> @v_test_canonicalize_reg_undef_reg_reg_v4f16(half %val0, half
; GFX11-FAKE16-NEXT: v_pack_b32_f16 v0, v0, 0
; GFX11-FAKE16-NEXT: v_pk_max_f16 v1, v1, v1
; GFX11-FAKE16-NEXT: s_setpc_b64 s[30:31]
%vec0 = insertelement <4 x half> undef, half %val0, i32 0
%vec0 = insertelement <4 x half> poison, half %val0, i32 0
%vec1 = insertelement <4 x half> %vec0, half %val1, i32 2
%vec2 = insertelement <4 x half> %vec1, half %val2, i32 3
%canonicalized = call <4 x half> @llvm.canonicalize.v4f16(<4 x half> %vec2)
Expand Down
6 changes: 3 additions & 3 deletions llvm/test/CodeGen/AMDGPU/promote-alloca-array-aggregate.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -134,7 +134,7 @@ define amdgpu_vs void @promote_load_from_store_aggr_varoff(<4 x i32> %input) {
; CHECK-NEXT: [[FOO3_UNPACK2:%.*]] = load i32, ptr addrspace(1) getelementptr inbounds (i8, ptr addrspace(1) @block4, i64 8), align 4
; CHECK-NEXT: [[TMP1:%.*]] = insertelement <3 x i32> undef, i32 [[FOO3_UNPACK2]], i32 2
; CHECK-NEXT: [[TMP2:%.*]] = extractelement <3 x i32> [[TMP1]], i32 [[FOO3_UNPACK2]]
; CHECK-NEXT: [[FOO12:%.*]] = insertelement <4 x i32> %input, i32 [[TMP2]], i64 3
; CHECK-NEXT: [[FOO12:%.*]] = insertelement <4 x i32> [[INPUT:%.*]], i32 [[TMP2]], i64 3
; CHECK-NEXT: store <4 x i32> [[FOO12]], ptr addrspace(1) @pv1, align 16
; CHECK-NEXT: ret void
;
Expand Down Expand Up @@ -344,7 +344,7 @@ define amdgpu_ps void @promote_double_aggr() #0 {
; CHECK-NEXT: [[FOO10:%.*]] = fadd double [[FOO5_FCA_1_EXTRACT]], [[FOO5_FCA_1_EXTRACT]]
; CHECK-NEXT: [[FOO16:%.*]] = fadd double [[FOO10]], [[FOO5_FCA_1_EXTRACT]]
; CHECK-NEXT: [[FOO17:%.*]] = fptrunc double [[FOO16]] to float
; CHECK-NEXT: [[FOO18:%.*]] = insertelement <4 x float> undef, float [[FOO17]], i32 0
; CHECK-NEXT: [[FOO18:%.*]] = insertelement <4 x float> poison, float [[FOO17]], i32 0
; CHECK-NEXT: [[FOO19:%.*]] = insertelement <4 x float> [[FOO18]], float [[FOO17]], i32 1
; CHECK-NEXT: [[FOO20:%.*]] = insertelement <4 x float> [[FOO19]], float [[FOO17]], i32 2
; CHECK-NEXT: [[FOO21:%.*]] = insertelement <4 x float> [[FOO20]], float [[FOO17]], i32 3
Expand All @@ -370,7 +370,7 @@ define amdgpu_ps void @promote_double_aggr() #0 {
%foo15 = load double, ptr addrspace(5) %foo14
%foo16 = fadd double %foo13, %foo15
%foo17 = fptrunc double %foo16 to float
%foo18 = insertelement <4 x float> undef, float %foo17, i32 0
%foo18 = insertelement <4 x float> poison, float %foo17, i32 0
%foo19 = insertelement <4 x float> %foo18, float %foo17, i32 1
%foo20 = insertelement <4 x float> %foo19, float %foo17, i32 2
%foo21 = insertelement <4 x float> %foo20, float %foo17, i32 3
Expand Down
22 changes: 11 additions & 11 deletions llvm/test/CodeGen/AMDGPU/promote-alloca-loadstores.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,15 +9,15 @@ define amdgpu_kernel void @test_overwrite(i64 %val, i1 %cond) {
; CHECK-NEXT: entry:
; CHECK-NEXT: br i1 [[COND]], label [[LOOP:%.*]], label [[END:%.*]]
; CHECK: loop:
; CHECK-NEXT: [[PROMOTEALLOCA:%.*]] = phi <3 x i64> [ [[TMP2:%.*]], [[LOOP]] ], [ <i64 43, i64 undef, i64 undef>, [[ENTRY:%.*]] ]
; CHECK-NEXT: [[TMP0:%.*]] = extractelement <3 x i64> [[PROMOTEALLOCA]], i32 0
; CHECK-NEXT: [[TMP1:%.*]] = insertelement <3 x i64> [[PROMOTEALLOCA]], i64 68, i32 0
; CHECK-NEXT: [[PROMOTEALLOCA1:%.*]] = phi <3 x i64> [ [[TMP2:%.*]], [[LOOP]] ], [ <i64 43, i64 undef, i64 undef>, [[ENTRY:%.*]] ]
; CHECK-NEXT: [[TMP0:%.*]] = extractelement <3 x i64> [[PROMOTEALLOCA1]], i32 0
; CHECK-NEXT: [[TMP1:%.*]] = insertelement <3 x i64> [[PROMOTEALLOCA1]], i64 68, i32 0
; CHECK-NEXT: [[TMP2]] = insertelement <3 x i64> [[TMP1]], i64 32, i32 0
; CHECK-NEXT: [[LOOP_CC:%.*]] = icmp ne i64 [[TMP0]], 68
; CHECK-NEXT: br i1 [[LOOP_CC]], label [[LOOP]], label [[END]]
; CHECK: end:
; CHECK-NEXT: [[PROMOTEALLOCA1:%.*]] = phi <3 x i64> [ [[TMP2]], [[LOOP]] ], [ <i64 43, i64 undef, i64 undef>, [[ENTRY]] ]
; CHECK-NEXT: [[TMP3:%.*]] = extractelement <3 x i64> [[PROMOTEALLOCA1]], i32 0
; CHECK-NEXT: [[PROMOTEALLOCA:%.*]] = phi <3 x i64> [ [[TMP2]], [[LOOP]] ], [ <i64 43, i64 undef, i64 undef>, [[ENTRY]] ]
; CHECK-NEXT: [[TMP3:%.*]] = extractelement <3 x i64> [[PROMOTEALLOCA]], i32 0
; CHECK-NEXT: ret void
;
entry:
Expand Down Expand Up @@ -64,15 +64,15 @@ define amdgpu_kernel void @test_no_overwrite(i64 %val, i1 %cond) {
; CHECK-NEXT: entry:
; CHECK-NEXT: br i1 [[COND]], label [[LOOP:%.*]], label [[END:%.*]]
; CHECK: loop:
; CHECK-NEXT: [[PROMOTEALLOCA:%.*]] = phi <3 x i64> [ [[TMP1:%.*]], [[LOOP]] ], [ <i64 43, i64 undef, i64 undef>, [[ENTRY:%.*]] ]
; CHECK-NEXT: [[TMP0:%.*]] = extractelement <3 x i64> [[PROMOTEALLOCA]], i32 0
; CHECK-NEXT: [[TMP1]] = insertelement <3 x i64> [[PROMOTEALLOCA]], i64 32, i32 1
; CHECK-NEXT: [[PROMOTEALLOCA1:%.*]] = phi <3 x i64> [ [[TMP1:%.*]], [[LOOP]] ], [ <i64 43, i64 undef, i64 undef>, [[ENTRY:%.*]] ]
; CHECK-NEXT: [[TMP0:%.*]] = extractelement <3 x i64> [[PROMOTEALLOCA1]], i32 0
; CHECK-NEXT: [[TMP1]] = insertelement <3 x i64> [[PROMOTEALLOCA1]], i64 32, i32 1
; CHECK-NEXT: [[LOOP_CC:%.*]] = icmp ne i64 [[TMP0]], 32
; CHECK-NEXT: br i1 [[LOOP_CC]], label [[LOOP]], label [[END]]
; CHECK: end:
; CHECK-NEXT: [[PROMOTEALLOCA1:%.*]] = phi <3 x i64> [ [[TMP1]], [[LOOP]] ], [ <i64 43, i64 undef, i64 undef>, [[ENTRY]] ]
; CHECK-NEXT: [[TMP2:%.*]] = extractelement <3 x i64> [[PROMOTEALLOCA1]], i32 0
; CHECK-NEXT: [[TMP3:%.*]] = extractelement <3 x i64> [[PROMOTEALLOCA1]], i32 1
; CHECK-NEXT: [[PROMOTEALLOCA:%.*]] = phi <3 x i64> [ [[TMP1]], [[LOOP]] ], [ <i64 43, i64 undef, i64 undef>, [[ENTRY]] ]
; CHECK-NEXT: [[TMP2:%.*]] = extractelement <3 x i64> [[PROMOTEALLOCA]], i32 0
; CHECK-NEXT: [[TMP3:%.*]] = extractelement <3 x i64> [[PROMOTEALLOCA]], i32 1
; CHECK-NEXT: ret void
;
entry:
Expand Down
Loading