Skip to content

Commit e7b9d21

Browse files
IcohedronIcohedron
committed
Fix tests with incorrect GEPs
A few tests had incorrect GEP indices or types. This commit fixes these GEPs and array types.
1 parent b08bb11 commit e7b9d21

File tree

5 files changed

+38
-42
lines changed

5 files changed

+38
-42
lines changed

‎llvm/test/CodeGen/DirectX/flatten-array.ll

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -159,9 +159,9 @@ define void @global_gep_load_index(i32 %row, i32 %col, i32 %timeIndex) {
159159
define void @global_incomplete_gep_chain(i32 %row, i32 %col) {
160160
; CHECK-LABEL: define void @global_incomplete_gep_chain(
161161
; CHECK-SAME: i32 [[ROW:%.*]], i32 [[COL:%.*]]) {
162-
; CHECK-NEXT: [[TMP1:%.*]] = mul i32 [[COL]], 1
162+
; CHECK-NEXT: [[TMP1:%.*]] = mul i32 [[COL]], 4
163163
; CHECK-NEXT: [[TMP2:%.*]] = add i32 0, [[TMP1]]
164-
; CHECK-NEXT: [[TMP3:%.*]] = mul i32 [[ROW]], 3
164+
; CHECK-NEXT: [[TMP3:%.*]] = mul i32 [[ROW]], 12
165165
; CHECK-NEXT: [[TMP4:%.*]] = add i32 [[TMP2]], [[TMP3]]
166166
; CHECK-NEXT: [[DOTFLAT:%.*]] = getelementptr inbounds [24 x i32], ptr @a.1dim, i32 0, i32 [[TMP4]]
167167
; CHECK-NOT: getelementptr inbounds [2 x [3 x [4 x i32]]]{{.*}}

‎llvm/test/CodeGen/DirectX/flatten-bug-117273.ll

Lines changed: 4 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -8,16 +8,16 @@
88
define internal void @main() {
99
; CHECK-LABEL: define internal void @main() {
1010
; CHECK-NEXT: [[ENTRY:.*:]]
11-
; CHECK-NEXT: [[TMP0:%.*]] = getelementptr [24 x float], ptr @ZerroInitArr.1dim, i32 0, i32 1
11+
; CHECK-NEXT: [[TMP0:%.*]] = getelementptr [6 x float], ptr @ZerroInitArr.1dim, i32 0, i32 3
1212
; CHECK-NEXT: [[DOTI0:%.*]] = load float, ptr [[TMP0]], align 16
13-
; CHECK-NEXT: [[TMP1:%.*]] = getelementptr [24 x float], ptr @ZerroInitArr.1dim, i32 0, i32 2
13+
; CHECK-NEXT: [[TMP1:%.*]] = getelementptr [6 x float], ptr @ZerroInitArr.1dim, i32 0, i32 6
1414
; CHECK-NEXT: [[DOTI03:%.*]] = load float, ptr [[TMP1]], align 16
1515
; CHECK-NEXT: ret void
1616
;
1717
entry:
18-
%0 = getelementptr [8 x [3 x float]], ptr @ZerroInitArr, i32 0, i32 1
18+
%0 = getelementptr [2 x [3 x float]], ptr @ZerroInitArr, i32 0, i32 1
1919
%.i0 = load float, ptr %0, align 16
20-
%1 = getelementptr [8 x [3 x float]], ptr @ZerroInitArr, i32 0, i32 2
20+
%1 = getelementptr [2 x [3 x float]], ptr @ZerroInitArr, i32 0, i32 2
2121
%.i03 = load float, ptr %1, align 16
2222
ret void
2323
}

‎llvm/test/CodeGen/DirectX/llc-vector-load-scalarize.ll

Lines changed: 27 additions & 33 deletions
Original file line numberDiff line numberDiff line change
@@ -3,43 +3,35 @@
33

44
; Make sure we can load groupshared, static vectors and arrays of vectors
55

6-
@"arrayofVecData" = local_unnamed_addr addrspace(3) global [2 x <3 x float>] zeroinitializer, align 16
6+
@"arrayofVecData" = local_unnamed_addr addrspace(3) global [2 x <4 x i32>] zeroinitializer, align 16
77
@"vecData" = external addrspace(3) global <4 x i32>, align 4
88
@staticArrayOfVecData = internal global [3 x <4 x i32>] [<4 x i32> <i32 1, i32 2, i32 3, i32 4>, <4 x i32> <i32 5, i32 6, i32 7, i32 8>, <4 x i32> <i32 9, i32 10, i32 11, i32 12>], align 4
9-
@"groushared2dArrayofVectors" = local_unnamed_addr addrspace(3) global [3 x [ 3 x <4 x i32>]] zeroinitializer, align 16
9+
@"groupshared2dArrayofVectors" = local_unnamed_addr addrspace(3) global [3 x [3 x <4 x i32>]] zeroinitializer, align 16
1010

11-
; CHECK: @arrayofVecData.scalarized.1dim = local_unnamed_addr addrspace(3) global [6 x float] zeroinitializer, align 16
11+
; CHECK: @arrayofVecData.scalarized.1dim = local_unnamed_addr addrspace(3) global [8 x i32] zeroinitializer, align 16
1212
; CHECK: @vecData.scalarized = external addrspace(3) global [4 x i32], align 4
1313
; CHECK: @staticArrayOfVecData.scalarized.1dim = internal global [12 x i32] [i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12], align 4
14-
; CHECK: @groushared2dArrayofVectors.scalarized.1dim = local_unnamed_addr addrspace(3) global [36 x i32] zeroinitializer, align 16
14+
; CHECK: @groupshared2dArrayofVectors.scalarized.1dim = local_unnamed_addr addrspace(3) global [36 x i32] zeroinitializer, align 16
1515

1616
; CHECK-NOT: @arrayofVecData
1717
; CHECK-NOT: @arrayofVecData.scalarized
1818
; CHECK-NOT: @vecData
1919
; CHECK-NOT: @staticArrayOfVecData
2020
; CHECK-NOT: @staticArrayOfVecData.scalarized
21-
; CHECK-NOT: @groushared2dArrayofVectors
22-
; CHECK-NOT: @groushared2dArrayofVectors.scalarized
21+
; CHECK-NOT: @groupshared2dArrayofVectors
22+
; CHECK-NOT: @groupshared2dArrayofVectors.scalarized
2323

2424
define <4 x i32> @load_array_vec_test() #0 {
2525
; CHECK-LABEL: define <4 x i32> @load_array_vec_test(
2626
; CHECK-SAME: ) #[[ATTR0:[0-9]+]] {
27-
; CHECK-NEXT: [[TMP1:%.*]] = bitcast ptr addrspace(3) @arrayofVecData.scalarized.1dim to ptr addrspace(3)
28-
; CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(3) [[TMP1]], align 4
29-
; CHECK-NEXT: [[TMP3:%.*]] = bitcast ptr addrspace(3) getelementptr (i32, ptr addrspace(3) @arrayofVecData.scalarized.1dim, i32 1) to ptr addrspace(3)
30-
; CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(3) [[TMP3]], align 4
31-
; CHECK-NEXT: [[TMP5:%.*]] = bitcast ptr addrspace(3) getelementptr (i32, ptr addrspace(3) @arrayofVecData.scalarized.1dim, i32 2) to ptr addrspace(3)
32-
; CHECK-NEXT: [[TMP6:%.*]] = load i32, ptr addrspace(3) [[TMP5]], align 4
33-
; CHECK-NEXT: [[TMP7:%.*]] = bitcast ptr addrspace(3) getelementptr (i32, ptr addrspace(3) @arrayofVecData.scalarized.1dim, i32 3) to ptr addrspace(3)
34-
; CHECK-NEXT: [[TMP8:%.*]] = load i32, ptr addrspace(3) [[TMP7]], align 4
35-
; CHECK-NEXT: [[TMP9:%.*]] = bitcast ptr addrspace(3) getelementptr inbounds ([6 x float], ptr addrspace(3) @arrayofVecData.scalarized.1dim, i32 0, i32 1) to ptr addrspace(3)
36-
; CHECK-NEXT: [[TMP10:%.*]] = load i32, ptr addrspace(3) [[TMP9]], align 4
37-
; CHECK-NEXT: [[TMP11:%.*]] = bitcast ptr addrspace(3) getelementptr (i32, ptr addrspace(3) getelementptr inbounds ([6 x float], ptr addrspace(3) @arrayofVecData.scalarized.1dim, i32 0, i32 1), i32 1) to ptr addrspace(3)
38-
; CHECK-NEXT: [[TMP12:%.*]] = load i32, ptr addrspace(3) [[TMP11]], align 4
39-
; CHECK-NEXT: [[TMP13:%.*]] = bitcast ptr addrspace(3) getelementptr (i32, ptr addrspace(3) getelementptr inbounds ([6 x float], ptr addrspace(3) @arrayofVecData.scalarized.1dim, i32 0, i32 1), i32 2) to ptr addrspace(3)
40-
; CHECK-NEXT: [[TMP14:%.*]] = load i32, ptr addrspace(3) [[TMP13]], align 4
41-
; CHECK-NEXT: [[TMP15:%.*]] = bitcast ptr addrspace(3) getelementptr (i32, ptr addrspace(3) getelementptr inbounds ([6 x float], ptr addrspace(3) @arrayofVecData.scalarized.1dim, i32 0, i32 1), i32 3) to ptr addrspace(3)
42-
; CHECK-NEXT: [[TMP16:%.*]] = load i32, ptr addrspace(3) [[TMP15]], align 4
27+
; CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(3) @arrayofVecData.scalarized.1dim, align 4
28+
; CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(3) getelementptr (i32, ptr addrspace(3) @arrayofVecData.scalarized.1dim, i32 1), align 4
29+
; CHECK-NEXT: [[TMP6:%.*]] = load i32, ptr addrspace(3) getelementptr (i32, ptr addrspace(3) @arrayofVecData.scalarized.1dim, i32 2), align 4
30+
; CHECK-NEXT: [[TMP8:%.*]] = load i32, ptr addrspace(3) getelementptr (i32, ptr addrspace(3) @arrayofVecData.scalarized.1dim, i32 3), align 4
31+
; CHECK-NEXT: [[TMP10:%.*]] = load i32, ptr addrspace(3) getelementptr inbounds ([8 x i32], ptr addrspace(3) @arrayofVecData.scalarized.1dim, i32 0, i32 4), align 4
32+
; CHECK-NEXT: [[TMP12:%.*]] = load i32, ptr addrspace(3) getelementptr (i32, ptr addrspace(3) getelementptr inbounds ([8 x i32], ptr addrspace(3) @arrayofVecData.scalarized.1dim, i32 0, i32 4), i32 1), align 4
33+
; CHECK-NEXT: [[TMP14:%.*]] = load i32, ptr addrspace(3) getelementptr (i32, ptr addrspace(3) getelementptr inbounds ([8 x i32], ptr addrspace(3) @arrayofVecData.scalarized.1dim, i32 0, i32 4), i32 2), align 4
34+
; CHECK-NEXT: [[TMP16:%.*]] = load i32, ptr addrspace(3) getelementptr (i32, ptr addrspace(3) getelementptr inbounds ([8 x i32], ptr addrspace(3) @arrayofVecData.scalarized.1dim, i32 0, i32 4), i32 3), align 4
4335
; CHECK-NEXT: [[DOTI05:%.*]] = add i32 [[TMP2]], [[TMP10]]
4436
; CHECK-NEXT: [[DOTI16:%.*]] = add i32 [[TMP4]], [[TMP12]]
4537
; CHECK-NEXT: [[DOTI27:%.*]] = add i32 [[TMP6]], [[TMP14]]
@@ -77,7 +69,9 @@ define <4 x i32> @load_vec_test() #0 {
7769
define <4 x i32> @load_static_array_of_vec_test(i32 %index) #0 {
7870
; CHECK-LABEL: define <4 x i32> @load_static_array_of_vec_test(
7971
; CHECK-SAME: i32 [[INDEX:%.*]]) #[[ATTR0]] {
80-
; CHECK-NEXT: [[DOTFLAT:%.*]] = getelementptr inbounds [12 x i32], ptr @staticArrayOfVecData.scalarized.1dim, i32 0, i32 [[INDEX]]
72+
; CHECK-NEXT: [[TMP3:%.*]] = mul i32 [[INDEX]], 4
73+
; CHECK-NEXT: [[TMP2:%.*]] = add i32 0, [[TMP3]]
74+
; CHECK-NEXT: [[DOTFLAT:%.*]] = getelementptr inbounds [12 x i32], ptr @staticArrayOfVecData.scalarized.1dim, i32 0, i32 [[TMP2]]
8175
; CHECK-NEXT: [[DOTI0:%.*]] = load i32, ptr [[DOTFLAT]], align 4
8276
; CHECK-NEXT: [[DOTFLAT_I1:%.*]] = getelementptr i32, ptr [[DOTFLAT]], i32 1
8377
; CHECK-NEXT: [[DOTI1:%.*]] = load i32, ptr [[DOTFLAT_I1]], align 4
@@ -99,14 +93,14 @@ define <4 x i32> @load_static_array_of_vec_test(i32 %index) #0 {
9993
define <4 x i32> @multid_load_test() #0 {
10094
; CHECK-LABEL: define <4 x i32> @multid_load_test(
10195
; CHECK-SAME: ) #[[ATTR0]] {
102-
; CHECK-NEXT: [[TMP1:%.*]] = load i32, ptr addrspace(3) @groushared2dArrayofVectors.scalarized.1dim, align 4
103-
; CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(3) getelementptr (i32, ptr addrspace(3) @groushared2dArrayofVectors.scalarized.1dim, i32 1), align 4
104-
; CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(3) getelementptr (i32, ptr addrspace(3) @groushared2dArrayofVectors.scalarized.1dim, i32 2), align 4
105-
; CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(3) getelementptr (i32, ptr addrspace(3) @groushared2dArrayofVectors.scalarized.1dim, i32 3), align 4
106-
; CHECK-NEXT: [[TMP5:%.*]] = load i32, ptr addrspace(3) getelementptr inbounds ([36 x i32], ptr addrspace(3) @groushared2dArrayofVectors.scalarized.1dim, i32 0, i32 4), align 4
107-
; CHECK-NEXT: [[DOTI13:%.*]] = load i32, ptr addrspace(3) getelementptr (i32, ptr addrspace(3) getelementptr inbounds ([36 x i32], ptr addrspace(3) @groushared2dArrayofVectors.scalarized.1dim, i32 0, i32 4), i32 1), align 4
108-
; CHECK-NEXT: [[DOTI25:%.*]] = load i32, ptr addrspace(3) getelementptr (i32, ptr addrspace(3) getelementptr inbounds ([36 x i32], ptr addrspace(3) @groushared2dArrayofVectors.scalarized.1dim, i32 0, i32 4), i32 2), align 4
109-
; CHECK-NEXT: [[DOTI37:%.*]] = load i32, ptr addrspace(3) getelementptr (i32, ptr addrspace(3) getelementptr inbounds ([36 x i32], ptr addrspace(3) @groushared2dArrayofVectors.scalarized.1dim, i32 0, i32 4), i32 3), align 4
96+
; CHECK-NEXT: [[TMP1:%.*]] = load i32, ptr addrspace(3) @groupshared2dArrayofVectors.scalarized.1dim, align 4
97+
; CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(3) getelementptr (i32, ptr addrspace(3) @groupshared2dArrayofVectors.scalarized.1dim, i32 1), align 4
98+
; CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(3) getelementptr (i32, ptr addrspace(3) @groupshared2dArrayofVectors.scalarized.1dim, i32 2), align 4
99+
; CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(3) getelementptr (i32, ptr addrspace(3) @groupshared2dArrayofVectors.scalarized.1dim, i32 3), align 4
100+
; CHECK-NEXT: [[TMP5:%.*]] = load i32, ptr addrspace(3) getelementptr inbounds ([36 x i32], ptr addrspace(3) @groupshared2dArrayofVectors.scalarized.1dim, i32 0, i32 16), align 4
101+
; CHECK-NEXT: [[DOTI13:%.*]] = load i32, ptr addrspace(3) getelementptr (i32, ptr addrspace(3) getelementptr inbounds ([36 x i32], ptr addrspace(3) @groupshared2dArrayofVectors.scalarized.1dim, i32 0, i32 16), i32 1), align 4
102+
; CHECK-NEXT: [[DOTI25:%.*]] = load i32, ptr addrspace(3) getelementptr (i32, ptr addrspace(3) getelementptr inbounds ([36 x i32], ptr addrspace(3) @groupshared2dArrayofVectors.scalarized.1dim, i32 0, i32 16), i32 2), align 4
103+
; CHECK-NEXT: [[DOTI37:%.*]] = load i32, ptr addrspace(3) getelementptr (i32, ptr addrspace(3) getelementptr inbounds ([36 x i32], ptr addrspace(3) @groupshared2dArrayofVectors.scalarized.1dim, i32 0, i32 16), i32 3), align 4
110104
; CHECK-NEXT: [[DOTI08:%.*]] = add i32 [[TMP1]], [[TMP5]]
111105
; CHECK-NEXT: [[DOTI19:%.*]] = add i32 [[TMP2]], [[DOTI13]]
112106
; CHECK-NEXT: [[DOTI210:%.*]] = add i32 [[TMP3]], [[DOTI25]]
@@ -117,8 +111,8 @@ define <4 x i32> @multid_load_test() #0 {
117111
; CHECK-NEXT: [[TMP6:%.*]] = insertelement <4 x i32> [[DOTUPTO217]], i32 [[DOTI311]], i32 3
118112
; CHECK-NEXT: ret <4 x i32> [[TMP6]]
119113
;
120-
%1 = load <4 x i32>, <4 x i32> addrspace(3)* getelementptr inbounds ([3 x [3 x <4 x i32>]], [3 x [3 x <4 x i32>]] addrspace(3)* @"groushared2dArrayofVectors", i32 0, i32 0, i32 0), align 4
121-
%2 = load <4 x i32>, <4 x i32> addrspace(3)* getelementptr inbounds ([3 x [3 x <4 x i32>]], [3 x [3 x <4 x i32>]] addrspace(3)* @"groushared2dArrayofVectors", i32 0, i32 1, i32 1), align 4
114+
%1 = load <4 x i32>, <4 x i32> addrspace(3)* getelementptr inbounds ([3 x [3 x <4 x i32>]], [3 x [3 x <4 x i32>]] addrspace(3)* @"groupshared2dArrayofVectors", i32 0, i32 0, i32 0), align 4
115+
%2 = load <4 x i32>, <4 x i32> addrspace(3)* getelementptr inbounds ([3 x [3 x <4 x i32>]], [3 x [3 x <4 x i32>]] addrspace(3)* @"groupshared2dArrayofVectors", i32 0, i32 1, i32 1), align 4
122116
%3 = add <4 x i32> %1, %2
123117
ret <4 x i32> %3
124118
}

‎llvm/test/CodeGen/DirectX/scalar-bug-117273.ll

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -8,13 +8,13 @@
88
define internal void @main() #1 {
99
; CHECK-LABEL: define internal void @main() {
1010
; CHECK-NEXT: [[ENTRY:.*:]]
11-
; CHECK-NEXT: [[TMP0:%.*]] = getelementptr inbounds [24 x float], ptr @StaticArr.scalarized.1dim, i32 0, i32 1
11+
; CHECK-NEXT: [[TMP0:%.*]] = getelementptr inbounds [24 x float], ptr @StaticArr.scalarized.1dim, i32 0, i32 3
1212
; CHECK-NEXT: [[DOTI0:%.*]] = load float, ptr [[TMP0]], align 16
1313
; CHECK-NEXT: [[DOTI1:%.*]] = getelementptr float, ptr [[TMP0]], i32 1
1414
; CHECK-NEXT: [[DOTI11:%.*]] = load float, ptr [[DOTI1]], align 4
1515
; CHECK-NEXT: [[DOTI2:%.*]] = getelementptr float, ptr [[TMP0]], i32 2
1616
; CHECK-NEXT: [[DOTI22:%.*]] = load float, ptr [[DOTI2]], align 8
17-
; CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds [24 x float], ptr @StaticArr.scalarized.1dim, i32 0, i32 2
17+
; CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds [24 x float], ptr @StaticArr.scalarized.1dim, i32 0, i32 6
1818
; CHECK-NEXT: [[DOTI03:%.*]] = load float, ptr [[TMP1]], align 16
1919
; CHECK-NEXT: [[DOTI14:%.*]] = getelementptr float, ptr [[TMP1]], i32 1
2020
; CHECK-NEXT: [[DOTI15:%.*]] = load float, ptr [[DOTI14]], align 4

‎llvm/test/CodeGen/DirectX/scalarize-alloca.ll

Lines changed: 3 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -33,7 +33,9 @@ define void @alloca_2d_gep_test() {
3333
; FCHECK: [[alloca_val:%.*]] = alloca [4 x i32], align 16
3434
; CHECK: [[tid:%.*]] = tail call i32 @llvm.dx.thread.id(i32 0)
3535
; SCHECK: [[gep:%.*]] = getelementptr inbounds nuw [2 x [2 x i32]], ptr [[alloca_val]], i32 0, i32 [[tid]]
36-
; FCHECK: [[gep:%.*]] = getelementptr inbounds nuw [4 x i32], ptr [[alloca_val]], i32 0, i32 [[tid]]
36+
; FCHECK: [[flatidx_mul:%.*]] = mul i32 [[tid]], 2
37+
; FCHECK: [[flatidx:%.*]] = add i32 0, [[flatidx_mul]]
38+
; FCHECK: [[gep:%.*]] = getelementptr inbounds nuw [4 x i32], ptr [[alloca_val]], i32 0, i32 [[flatidx]]
3739
; CHECK: ret void
3840
%1 = alloca [2 x <2 x i32>], align 16
3941
%2 = tail call i32 @llvm.dx.thread.id(i32 0)

0 commit comments

Comments
 (0)