| blob | ccac86f52fe60764fde30181f7f8abf9d98466a6 |
1 ; RUN: opt %loadPolly -polly-opt-isl -analyze -polly-ast -polly-tile-sizes=256,16 < %s | FileCheck %s
2 ; RUN: opt %loadPolly -polly-opt-isl -analyze -polly-tiling=false -polly-ast -polly-tile-sizes=256,16 < %s | FileCheck %s --check-prefix=NOTILING
4 ; RUN: opt %loadPolly -polly-opt-isl -analyze \
5 ; RUN: -polly-2nd-level-tiling -polly-ast \
6 ; RUN: -polly-tile-sizes=256,16 \
7 ; RUN: -polly-2nd-level-tile-sizes=16,8 < %s | \
8 ; RUN: FileCheck %s --check-prefix=TWOLEVEL
10 ; RUN: opt %loadPolly -polly-opt-isl -analyze \
11 ; RUN: -polly-2nd-level-tiling -polly-ast \
12 ; RUN: -polly-tile-sizes=256,16 \
13 ; RUN: -polly-register-tiling \
14 ; RUN: -polly-2nd-level-tile-sizes=16,8 < %s | \
15 ; RUN: FileCheck %s --check-prefix=TWO-PLUS-REGISTER
17 ; RUN: opt %loadPolly -polly-opt-isl -analyze \
18 ; RUN: -polly-2nd-level-tiling -polly-ast \
19 ; RUN: -polly-tile-sizes=256,16 \
20 ; RUN: -polly-register-tiling -polly-register-tile-sizes=2,4 \
21 ; RUN: -polly-vectorizer=polly \
22 ; RUN: -polly-2nd-level-tile-sizes=16,8 < %s | \
23 ; RUN: FileCheck %s --check-prefix=TWO-PLUS-REGISTER-PLUS-VECTORIZATION
25 ; CHECK: // 1st level tiling - Tiles
26 ; CHECK: for (int c0 = 0; c0 <= 3; c0 += 1)
27 ; CHECK: for (int c1 = 0; c1 <= 31; c1 += 1)
28 ; CHECK: // 1st level tiling - Points
29 ; CHECK: for (int c2 = 0; c2 <= 255; c2 += 1)
30 ; CHECK: for (int c3 = 0; c3 <= 15; c3 += 1)
31 ; CHECK: Stmt_for_body3(256 * c0 + c2, 16 * c1 + c3);
33 ; NOTILING: for (int c0 = 0; c0 <= 1023; c0 += 1)
34 ; NOTILING: for (int c1 = 0; c1 <= 511; c1 += 1)
35 ; NOTILING: Stmt_for_body3(c0, c1);
38 ; TWOLEVEL: // 1st level tiling - Tiles
39 ; TWOLEVEL: for (int c0 = 0; c0 <= 3; c0 += 1)
40 ; TWOLEVEL: for (int c1 = 0; c1 <= 31; c1 += 1)
41 ; TWOLEVEL: // 1st level tiling - Points
42 ; TWOLEVEL: // 2nd level tiling - Tiles
43 ; TWOLEVEL: for (int c2 = 0; c2 <= 15; c2 += 1)
44 ; TWOLEVEL: for (int c3 = 0; c3 <= 1; c3 += 1)
45 ; TWOLEVEL: // 2nd level tiling - Points
46 ; TWOLEVEL: for (int c4 = 0; c4 <= 15; c4 += 1)
47 ; TWOLEVEL: for (int c5 = 0; c5 <= 7; c5 += 1)
48 ; TWOLEVEL: Stmt_for_body3(256 * c0 + 16 * c2 + c4, 16 * c1 + 8 * c3 + c5);
51 ; TWO-PLUS-REGISTER: // 1st level tiling - Tiles
52 ; TWO-PLUS-REGISTER: for (int c0 = 0; c0 <= 3; c0 += 1)
53 ; TWO-PLUS-REGISTER: for (int c1 = 0; c1 <= 31; c1 += 1)
54 ; TWO-PLUS-REGISTER: // 1st level tiling - Points
55 ; TWO-PLUS-REGISTER: // 2nd level tiling - Tiles
56 ; TWO-PLUS-REGISTER: for (int c2 = 0; c2 <= 15; c2 += 1)
57 ; TWO-PLUS-REGISTER: for (int c3 = 0; c3 <= 1; c3 += 1)
58 ; TWO-PLUS-REGISTER: // 2nd level tiling - Points
59 ; TWO-PLUS-REGISTER: // Register tiling - Tiles
60 ; TWO-PLUS-REGISTER: for (int c4 = 0; c4 <= 7; c4 += 1)
61 ; TWO-PLUS-REGISTER: for (int c5 = 0; c5 <= 3; c5 += 1)
62 ; TWO-PLUS-REGISTER: // Register tiling - Points
63 ; TWO-PLUS-REGISTER: {
64 ; TWO-PLUS-REGISTER: Stmt_for_body3(256 * c0 + 16 * c2 + 2 * c4, 16 * c1 + 8 * c3 + 2 * c5);
65 ; TWO-PLUS-REGISTER: Stmt_for_body3(256 * c0 + 16 * c2 + 2 * c4, 16 * c1 + 8 * c3 + 2 * c5 + 1);
66 ; TWO-PLUS-REGISTER: Stmt_for_body3(256 * c0 + 16 * c2 + 2 * c4 + 1, 16 * c1 + 8 * c3 + 2 * c5);
67 ; TWO-PLUS-REGISTER: Stmt_for_body3(256 * c0 + 16 * c2 + 2 * c4 + 1, 16 * c1 + 8 * c3 + 2 * c5 + 1);
68 ; TWO-PLUS-REGISTER: }
70 ; TWO-PLUS-REGISTER-PLUS-VECTORIZATION: #pragma known-parallel
71 ; TWO-PLUS-REGISTER-PLUS-VECTORIZATION: for (int c0 = 0; c0 <= 3; c0 += 1)
72 ; TWO-PLUS-REGISTER-PLUS-VECTORIZATION: for (int c1 = 0; c1 <= 31; c1 += 1)
73 ; TWO-PLUS-REGISTER-PLUS-VECTORIZATION: for (int c2 = 0; c2 <= 15; c2 += 1)
74 ; TWO-PLUS-REGISTER-PLUS-VECTORIZATION: for (int c3 = 0; c3 <= 1; c3 += 1)
75 ; TWO-PLUS-REGISTER-PLUS-VECTORIZATION: for (int c4 = 0; c4 <= 7; c4 += 1)
76 ; TWO-PLUS-REGISTER-PLUS-VECTORIZATION: for (int c5 = 0; c5 <= 1; c5 += 1) {
77 ; TWO-PLUS-REGISTER-PLUS-VECTORIZATION: // SIMD
78 ; TWO-PLUS-REGISTER-PLUS-VECTORIZATION: for (int c8 = 0; c8 <= 3; c8 += 1)
79 ; TWO-PLUS-REGISTER-PLUS-VECTORIZATION: Stmt_for_body3(256 * c0 + 16 * c2 + 2 * c4, 16 * c1 + 8 * c3 + 4 * c5 + c8);
80 ; TWO-PLUS-REGISTER-PLUS-VECTORIZATION: // SIMD
81 ; TWO-PLUS-REGISTER-PLUS-VECTORIZATION: for (int c8 = 0; c8 <= 3; c8 += 1)
82 ; TWO-PLUS-REGISTER-PLUS-VECTORIZATION: Stmt_for_body3(256 * c0 + 16 * c2 + 2 * c4 + 1, 16 * c1 + 8 * c3 + 4 * c5 + c8);
83 ; TWO-PLUS-REGISTER-PLUS-VECTORIZATION: }
85 target datalayout = "e-m:e-p:32:32-i64:64-v128:64:128-n32-S64"
87 ; Function Attrs: nounwind
88 define void @rect([512 x i32]* %A) {
89 entry:
90 br label %entry.split
92 entry.split: ; preds = %entry
93 br label %for.body3.lr.ph
95 for.body3.lr.ph: ; preds = %for.inc5, %entry.split
96 %i.0 = phi i32 [ 0, %entry.split ], [ %inc6, %for.inc5 ]
97 br label %for.body3
99 for.body3: ; preds = %for.body3.lr.ph, %for.body3
100 %j.0 = phi i32 [ 0, %for.body3.lr.ph ], [ %inc, %for.body3 ]
101 %mul = mul nsw i32 %j.0, %i.0
102 %rem = srem i32 %mul, 42
103 %arrayidx4 = getelementptr inbounds [512 x i32], [512 x i32]* %A, i32 %i.0, i32 %j.0
104 store i32 %rem, i32* %arrayidx4, align 4
105 %inc = add nsw i32 %j.0, 1
106 %cmp2 = icmp slt i32 %inc, 512
107 br i1 %cmp2, label %for.body3, label %for.inc5
109 for.inc5: ; preds = %for.body3
110 %inc6 = add nsw i32 %i.0, 1
111 %cmp = icmp slt i32 %inc6, 1024
112 br i1 %cmp, label %for.body3.lr.ph, label %for.end7
114 for.end7: ; preds = %for.inc5
115 ret void
116 }