1 ; RUN: opt %loadPolly -polly-detect -stats < %s 2>&1 | FileCheck %s
5 ; CHECK-DAG: 4 polly-detect - Maximal number of loops in scops (profitable scops only)
6 ; CHECK-DAG: 4 polly-detect - Maximal number of loops in scops
7 ; CHECK-DAG: 10 polly-detect - Number of loops in scops (profitable scops only)
8 ; CHECK-DAG: 10 polly-detect - Number of loops in scops
9 ; CHECK-DAG: 10 polly-detect - Number of total loops
10 ; CHECK-DAG: 4 polly-detect - Number of scops (profitable scops only)
11 ; CHECK-DAG: 1 polly-detect - Number of scops with maximal loop depth 4 (profitable scops only)
12 ; CHECK-DAG: 1 polly-detect - Number of scops with maximal loop depth 1 (profitable scops only)
13 ; CHECK-DAG: 1 polly-detect - Number of scops with maximal loop depth 3 (profitable scops only)
14 ; CHECK-DAG: 1 polly-detect - Number of scops with maximal loop depth 2 (profitable scops only)
15 ; CHECK-DAG: 4 polly-detect - Number of scops
16 ; CHECK-DAG: 1 polly-detect - Number of scops with maximal loop depth 4
17 ; CHECK-DAG: 1 polly-detect - Number of scops with maximal loop depth 1
18 ; CHECK-DAG: 1 polly-detect - Number of scops with maximal loop depth 3
19 ; CHECK-DAG: 1 polly-detect - Number of scops with maximal loop depth 2
21 ; void foo_1d(float *A) {
22 ; for (long i = 0; i < 1024; i++)
26 ; void foo_2d(float *A) {
27 ; for (long i = 0; i < 1024; i++)
28 ; for (long j = 0; j < 1024; j++)
32 ; void foo_3d(float *A) {
33 ; for (long i = 0; i < 1024; i++)
34 ; for (long j = 0; j < 1024; j++)
35 ; for (long k = 0; k < 1024; k++)
36 ; A[i + j + k] += i + j + k;
39 ; void foo_4d(float *A) {
40 ; for (long i = 0; i < 1024; i++)
41 ; for (long j = 0; j < 1024; j++)
42 ; for (long k = 0; k < 1024; k++)
43 ; for (long l = 0; l < 1024; l++)
44 ; A[i + j + k + l] += i + j + k + l;
47 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
49 define void @foo_1d(float* %A) {
53 bb1: ; preds = %bb6, %bb
54 %i.0 = phi i64 [ 0, %bb ], [ %tmp7, %bb6 ]
55 %exitcond = icmp ne i64 %i.0, 1024
56 br i1 %exitcond, label %bb2, label %bb8
59 %tmp = sitofp i64 %i.0 to float
60 %tmp3 = getelementptr inbounds float, float* %A, i64 %i.0
61 %tmp4 = load float, float* %tmp3, align 4
62 %tmp5 = fadd float %tmp4, %tmp
63 store float %tmp5, float* %tmp3, align 4
67 %tmp7 = add nuw nsw i64 %i.0, 1
74 define void @foo_2d(float* %A) {
78 bb2: ; preds = %bb14, %bb
79 %i.0 = phi i64 [ 0, %bb ], [ %tmp15, %bb14 ]
80 %exitcond1 = icmp ne i64 %i.0, 1024
81 br i1 %exitcond1, label %bb3, label %bb16
86 bb4: ; preds = %bb11, %bb3
87 %j.0 = phi i64 [ 0, %bb3 ], [ %tmp12, %bb11 ]
88 %exitcond = icmp ne i64 %j.0, 1024
89 br i1 %exitcond, label %bb5, label %bb13
92 %tmp = add nuw nsw i64 %i.0, %j.0
93 %tmp6 = sitofp i64 %tmp to float
94 %tmp7 = add nuw nsw i64 %i.0, %j.0
95 %tmp8 = getelementptr inbounds float, float* %A, i64 %tmp7
96 %tmp9 = load float, float* %tmp8, align 4
97 %tmp10 = fadd float %tmp9, %tmp6
98 store float %tmp10, float* %tmp8, align 4
102 %tmp12 = add nuw nsw i64 %j.0, 1
108 bb14: ; preds = %bb13
109 %tmp15 = add nuw nsw i64 %i.0, 1
116 define void @foo_3d(float* %A) {
120 bb3: ; preds = %bb22, %bb
121 %i.0 = phi i64 [ 0, %bb ], [ %tmp23, %bb22 ]
122 %exitcond2 = icmp ne i64 %i.0, 1024
123 br i1 %exitcond2, label %bb4, label %bb24
128 bb5: ; preds = %bb19, %bb4
129 %j.0 = phi i64 [ 0, %bb4 ], [ %tmp20, %bb19 ]
130 %exitcond1 = icmp ne i64 %j.0, 1024
131 br i1 %exitcond1, label %bb6, label %bb21
136 bb7: ; preds = %bb16, %bb6
137 %k.0 = phi i64 [ 0, %bb6 ], [ %tmp17, %bb16 ]
138 %exitcond = icmp ne i64 %k.0, 1024
139 br i1 %exitcond, label %bb8, label %bb18
142 %tmp = add nuw nsw i64 %i.0, %j.0
143 %tmp9 = add nuw nsw i64 %tmp, %k.0
144 %tmp10 = sitofp i64 %tmp9 to float
145 %tmp11 = add nuw nsw i64 %i.0, %j.0
146 %tmp12 = add nuw nsw i64 %tmp11, %k.0
147 %tmp13 = getelementptr inbounds float, float* %A, i64 %tmp12
148 %tmp14 = load float, float* %tmp13, align 4
149 %tmp15 = fadd float %tmp14, %tmp10
150 store float %tmp15, float* %tmp13, align 4
154 %tmp17 = add nuw nsw i64 %k.0, 1
160 bb19: ; preds = %bb18
161 %tmp20 = add nuw nsw i64 %j.0, 1
167 bb22: ; preds = %bb21
168 %tmp23 = add nuw nsw i64 %i.0, 1
175 define void @foo_4d(float* %A) {
179 bb4: ; preds = %bb30, %bb
180 %i.0 = phi i64 [ 0, %bb ], [ %tmp31, %bb30 ]
181 %exitcond3 = icmp ne i64 %i.0, 1024
182 br i1 %exitcond3, label %bb5, label %bb32
187 bb6: ; preds = %bb27, %bb5
188 %j.0 = phi i64 [ 0, %bb5 ], [ %tmp28, %bb27 ]
189 %exitcond2 = icmp ne i64 %j.0, 1024
190 br i1 %exitcond2, label %bb7, label %bb29
195 bb8: ; preds = %bb24, %bb7
196 %k.0 = phi i64 [ 0, %bb7 ], [ %tmp25, %bb24 ]
197 %exitcond1 = icmp ne i64 %k.0, 1024
198 br i1 %exitcond1, label %bb9, label %bb26
203 bb10: ; preds = %bb21, %bb9
204 %l.0 = phi i64 [ 0, %bb9 ], [ %tmp22, %bb21 ]
205 %exitcond = icmp ne i64 %l.0, 1024
206 br i1 %exitcond, label %bb11, label %bb23
208 bb11: ; preds = %bb10
209 %tmp = add nuw nsw i64 %i.0, %j.0
210 %tmp12 = add nuw nsw i64 %tmp, %k.0
211 %tmp13 = add nuw nsw i64 %tmp12, %l.0
212 %tmp14 = sitofp i64 %tmp13 to float
213 %tmp15 = add nuw nsw i64 %i.0, %j.0
214 %tmp16 = add nuw nsw i64 %tmp15, %k.0
215 %tmp17 = add nuw nsw i64 %tmp16, %l.0
216 %tmp18 = getelementptr inbounds float, float* %A, i64 %tmp17
217 %tmp19 = load float, float* %tmp18, align 4
218 %tmp20 = fadd float %tmp19, %tmp14
219 store float %tmp20, float* %tmp18, align 4
222 bb21: ; preds = %bb11
223 %tmp22 = add nuw nsw i64 %l.0, 1
226 bb23: ; preds = %bb10
229 bb24: ; preds = %bb23
230 %tmp25 = add nuw nsw i64 %k.0, 1
236 bb27: ; preds = %bb26
237 %tmp28 = add nuw nsw i64 %j.0, 1
243 bb30: ; preds = %bb29
244 %tmp31 = add nuw nsw i64 %i.0, 1