1 ; RUN: opt %loadPolly -polly-codegen-ppcg -polly-acc-dump-code \
2 ; RUN: -polly-acc-use-private \
3 ; RUN: -disable-output < %s | \
4 ; RUN: FileCheck -check-prefix=CODE %s
6 ; RUN: opt %loadPolly -polly-codegen-ppcg \
7 ; RUN: -polly-acc-use-private \
8 ; RUN: -disable-output -polly-acc-dump-kernel-ir < %s | \
9 ; RUN: FileCheck -check-prefix=KERNEL %s
13 ; void add(float *A) {
14 ; for (long i = 0; i < 32; i++)
15 ; for (long j = 0; j < 10; j++)
22 ; CODE: for (int c3 = 0; c3 <= 9; c3 += 1)
23 ; CODE: Stmt_bb5(t0, c3);
27 ; KERNEL: %private_array = alloca [1 x float]
29 ; KERNEL: %polly.access.cast.private_array = bitcast [1 x float]* %private_array to float*
30 ; KERNEL-NEXT: %polly.access.private_array = getelementptr float, float* %polly.access.cast.private_array, i64 0
31 ; KERNEL-NEXT: %polly.access.cast.MemRef_A = bitcast i8 addrspace(1)* %MemRef_A to float addrspace(1)*
32 ; KERNEL-NEXT: %polly.access.MemRef_A = getelementptr float, float addrspace(1)* %polly.access.cast.MemRef_A, i64 %t0
33 ; KERNEL-NEXT: %shared.read = load float, float addrspace(1)* %polly.access.MemRef_A
34 ; KERNEL-NEXT: store float %shared.read, float* %polly.access.private_array
36 ; KERNEL: %polly.access.cast.private_array5 = bitcast [1 x float]* %private_array to float*
37 ; KERNEL-NEXT: %polly.access.private_array6 = getelementptr float, float* %polly.access.cast.private_array5, i64 0
38 ; KERNEL-NEXT: %polly.access.cast.MemRef_A7 = bitcast i8 addrspace(1)* %MemRef_A to float addrspace(1)*
39 ; KERNEL-NEXT: %polly.access.MemRef_A8 = getelementptr float, float addrspace(1)* %polly.access.cast.MemRef_A7, i64 %t0
40 ; KERNEL-NEXT: %shared.write = load float, float* %polly.access.private_array6
41 ; KERNEL-NEXT: store float %shared.write, float addrspace(1)* %polly.access.MemRef_A8
43 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
45 define void @add(float* %A) {
49 bb2: ; preds = %bb11, %bb
50 %i.0 = phi i64 [ 0, %bb ], [ %tmp12, %bb11 ]
51 %exitcond1 = icmp ne i64 %i.0, 32
52 br i1 %exitcond1, label %bb3, label %bb13
57 bb4: ; preds = %bb8, %bb3
58 %j.0 = phi i64 [ 0, %bb3 ], [ %tmp9, %bb8 ]
59 %exitcond = icmp ne i64 %j.0, 10
60 br i1 %exitcond, label %bb5, label %bb10
63 %tmp = getelementptr inbounds float, float* %A, i64 %i.0
64 %tmp6 = load float, float* %tmp, align 4
65 %tmp7 = fadd float %tmp6, 1.000000e+00
66 store float %tmp7, float* %tmp, align 4
70 %tmp9 = add nuw nsw i64 %j.0, 1
77 %tmp12 = add nuw nsw i64 %i.0, 1