2 ; RUN: opt < %s -basicaa -loop-interchange -verify-dom-info -verify-loop-info \
3 ; RUN: -S -debug 2>&1 | FileCheck %s
5 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
6 target triple = "x86_64-unknown-linux-gnu"
8 @A = common global [100 x [100 x i32]] zeroinitializer
9 @B = common global [100 x i32] zeroinitializer
10 @C = common global [100 x [100 x i32]] zeroinitializer
11 @D = common global [100 x [100 x [100 x i32]]] zeroinitializer
13 ;; Test that a flow dependency in outer loop doesn't prevent interchange in
16 ;; for (int k = 0; k < 100; ++k) {
18 ;; for (int i = 0; i < 1000; ++i)
19 ;; for(int j = 1; j < 1000; ++j)
20 ;; Arr[j][i] = Arr[j][i]+k;
24 ; CHECK: Processing Inner Loop Id = 2 and OuterLoopId = 1
25 ; CHECK: Loops interchanged.
27 ; CHECK: Processing Inner Loop Id = 1 and OuterLoopId = 0
28 ; CHECK: Not interchanging loops. Cannot prove legality.
30 @T = internal global [100 x double] zeroinitializer, align 4
31 @Arr = internal global [1000 x [1000 x i32]] zeroinitializer, align 4
33 define void @interchange_09(i32 %k) {
37 for.cond.cleanup: ; preds = %for.cond.cleanup4
40 for.body: ; preds = %for.cond.cleanup4, %entry
41 %indvars.iv45 = phi i64 [ 0, %entry ], [ %indvars.iv.next46, %for.cond.cleanup4 ]
42 %call = call double @fn1()
43 %arrayidx = getelementptr inbounds [100 x double], [100 x double]* @T, i64 0, i64 %indvars.iv45
44 store double %call, double* %arrayidx, align 8
45 br label %for.cond6.preheader
47 for.cond6.preheader: ; preds = %for.cond.cleanup8, %for.body
48 %indvars.iv42 = phi i64 [ 0, %for.body ], [ %indvars.iv.next43, %for.cond.cleanup8 ]
51 for.cond.cleanup4: ; preds = %for.cond.cleanup8
52 %tmp = load double, double* %arrayidx, align 8
53 call void @fn2(double %tmp)
54 %indvars.iv.next46 = add nuw nsw i64 %indvars.iv45, 1
55 %exitcond47 = icmp ne i64 %indvars.iv.next46, 100
56 br i1 %exitcond47, label %for.body, label %for.cond.cleanup
58 for.cond.cleanup8: ; preds = %for.body9
59 %indvars.iv.next43 = add nuw nsw i64 %indvars.iv42, 1
60 %exitcond44 = icmp ne i64 %indvars.iv.next43, 1000
61 br i1 %exitcond44, label %for.cond6.preheader, label %for.cond.cleanup4
63 for.body9: ; preds = %for.body9, %for.cond6.preheader
64 %indvars.iv = phi i64 [ 1, %for.cond6.preheader ], [ %indvars.iv.next, %for.body9 ]
65 %arrayidx13 = getelementptr inbounds [1000 x [1000 x i32]], [1000 x [1000 x i32]]* @Arr, i64 0, i64 %indvars.iv, i64 %indvars.iv42
66 %tmp1 = load i32, i32* %arrayidx13, align 4
67 %tmp2 = trunc i64 %indvars.iv45 to i32
68 %add = add nsw i32 %tmp1, %tmp2
69 store i32 %add, i32* %arrayidx13, align 4
70 %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
71 %exitcond = icmp ne i64 %indvars.iv.next, 1000
72 br i1 %exitcond, label %for.body9, label %for.cond.cleanup8
75 declare double @fn1() readnone
76 declare void @fn2(double) readnone