1 ; RUN: opt %loadPolly -polly-import-jscop -polly-opt-isl \
2 ; RUN: -polly-target-throughput-vector-fma=1 \
3 ; RUN: -polly-target-latency-vector-fma=8 \
4 ; RUN: -polly-target-1st-cache-level-associativity=8 \
5 ; RUN: -polly-target-2nd-cache-level-associativity=8 \
6 ; RUN: -polly-target-1st-cache-level-size=32768 \
7 ; RUN: -polly-target-vector-register-bitwidth=256 \
8 ; RUN: -polly-target-2nd-cache-level-size=262144 \
9 ; RUN: -polly-import-jscop-postfix=transformed -polly-codegen -S < %s \
12 ; Check that we do not create different alias sets for locations represented by
13 ; different raw pointers.
15 ; Also check that we disable the Loop Vectorizer.
17 ; CHECK-NOT: !76 = distinct !{!76, !5, !"second level alias metadata"}
18 ; CHECK: !{!"llvm.loop.vectorize.enable", i1 false}
20 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
21 target triple = "x86_64-unknown-unknown"
23 define void @kernel_gemm(i32 %ni, i32 %nj, i32 %nk, [1024 x double]* %A, [1024 x double]* %B, [1024 x double]* %C, double* %C1) {
27 entry.split: ; preds = %entry
30 for.body: ; preds = %for.inc22, %entry.split
31 %indvars.iv43 = phi i64 [ 0, %entry.split ], [ %indvars.iv.next44, %for.inc22 ]
34 for.body3: ; preds = %for.inc19, %for.body
35 %indvars.iv40 = phi i64 [ 0, %for.body ], [ %indvars.iv.next41, %for.inc19 ]
38 for.body6: ; preds = %for.body6, %for.body3
39 %indvars.iv = phi i64 [ 0, %for.body3 ], [ %indvars.iv.next, %for.body6 ]
40 %tmp = load double, double* %C1, align 8
41 %arrayidx9 = getelementptr inbounds [1024 x double], [1024 x double]* %A, i64 %indvars.iv43, i64 %indvars.iv
42 %tmp1 = load double, double* %arrayidx9, align 8
43 %arrayidx13 = getelementptr inbounds [1024 x double], [1024 x double]* %B, i64 %indvars.iv, i64 %indvars.iv40
44 %tmp2 = load double, double* %arrayidx13, align 8
45 %mul = fmul double %tmp1, %tmp2
46 %add = fadd double %tmp, %mul
47 %arrayidx17 = getelementptr inbounds [1024 x double], [1024 x double]* %C, i64 %indvars.iv43, i64 %indvars.iv40
48 %tmp3 = load double, double* %arrayidx17, align 8
49 %add18 = fadd double %tmp3, %add
50 store double %add18, double* %arrayidx17, align 8
51 %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
52 %exitcond = icmp ne i64 %indvars.iv.next, 1024
53 br i1 %exitcond, label %for.body6, label %for.inc19
55 for.inc19: ; preds = %for.body6
56 %indvars.iv.next41 = add nuw nsw i64 %indvars.iv40, 1
57 %exitcond42 = icmp ne i64 %indvars.iv.next41, 1024
58 br i1 %exitcond42, label %for.body3, label %for.inc22
60 for.inc22: ; preds = %for.inc19
61 %indvars.iv.next44 = add nuw nsw i64 %indvars.iv43, 1
62 %exitcond45 = icmp ne i64 %indvars.iv.next44, 1024
63 br i1 %exitcond45, label %for.body, label %for.end24
65 for.end24: ; preds = %for.inc22