1 ; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+sse2 --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,SSE2
2 ; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,AVX1
3 ; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx2 --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,AVX2
4 ; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx512bw,+avx512vl --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,AVX512
7 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
8 target triple = "x86_64-unknown-linux-gnu"
10 @A = global [1024 x i8] zeroinitializer, align 128
11 @B = global [1024 x i8] zeroinitializer, align 128
13 ; CHECK: LV: Checking a loop in "test"
15 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: store i8 %v5, i8* %out5, align 1
16 ; SSE2: LV: Found an estimated cost of 49 for VF 2 For instruction: store i8 %v5, i8* %out5, align 1
17 ; SSE2: LV: Found an estimated cost of 98 for VF 4 For instruction: store i8 %v5, i8* %out5, align 1
18 ; SSE2: LV: Found an estimated cost of 201 for VF 8 For instruction: store i8 %v5, i8* %out5, align 1
19 ; SSE2: LV: Found an estimated cost of 408 for VF 16 For instruction: store i8 %v5, i8* %out5, align 1
21 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: store i8 %v5, i8* %out5, align 1
22 ; AVX1: LV: Found an estimated cost of 27 for VF 2 For instruction: store i8 %v5, i8* %out5, align 1
23 ; AVX1: LV: Found an estimated cost of 54 for VF 4 For instruction: store i8 %v5, i8* %out5, align 1
24 ; AVX1: LV: Found an estimated cost of 101 for VF 8 For instruction: store i8 %v5, i8* %out5, align 1
25 ; AVX1: LV: Found an estimated cost of 201 for VF 16 For instruction: store i8 %v5, i8* %out5, align 1
26 ; AVX1: LV: Found an estimated cost of 498 for VF 32 For instruction: store i8 %v5, i8* %out5, align 1
28 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: store i8 %v5, i8* %out5, align 1
29 ; AVX2: LV: Found an estimated cost of 10 for VF 2 For instruction: store i8 %v5, i8* %out5, align 1
30 ; AVX2: LV: Found an estimated cost of 12 for VF 4 For instruction: store i8 %v5, i8* %out5, align 1
31 ; AVX2: LV: Found an estimated cost of 18 for VF 8 For instruction: store i8 %v5, i8* %out5, align 1
32 ; AVX2: LV: Found an estimated cost of 30 for VF 16 For instruction: store i8 %v5, i8* %out5, align 1
33 ; AVX2: LV: Found an estimated cost of 96 for VF 32 For instruction: store i8 %v5, i8* %out5, align 1
35 ; AVX512: LV: Found an estimated cost of 1 for VF 1 For instruction: store i8 %v5, i8* %out5, align 1
36 ; AVX512: LV: Found an estimated cost of 18 for VF 2 For instruction: store i8 %v5, i8* %out5, align 1
37 ; AVX512: LV: Found an estimated cost of 38 for VF 4 For instruction: store i8 %v5, i8* %out5, align 1
38 ; AVX512: LV: Found an estimated cost of 98 for VF 8 For instruction: store i8 %v5, i8* %out5, align 1
39 ; AVX512: LV: Found an estimated cost of 197 for VF 16 For instruction: store i8 %v5, i8* %out5, align 1
40 ; AVX512: LV: Found an estimated cost of 295 for VF 32 For instruction: store i8 %v5, i8* %out5, align 1
41 ; AVX512: LV: Found an estimated cost of 591 for VF 64 For instruction: store i8 %v5, i8* %out5, align 1
43 ; CHECK-NOT: LV: Found an estimated cost of {{[0-9]+}} for VF {{[0-9]+}} For instruction: store i8 %v5, i8* %out5, align 1
50 %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ]
52 %iv.0 = add nuw nsw i64 %iv, 0
53 %iv.1 = add nuw nsw i64 %iv, 1
54 %iv.2 = add nuw nsw i64 %iv, 2
55 %iv.3 = add nuw nsw i64 %iv, 3
56 %iv.4 = add nuw nsw i64 %iv, 4
57 %iv.5 = add nuw nsw i64 %iv, 5
59 %in = getelementptr inbounds [1024 x i8], [1024 x i8]* @A, i64 0, i64 %iv.0
69 %out0 = getelementptr inbounds [1024 x i8], [1024 x i8]* @B, i64 0, i64 %iv.0
70 %out1 = getelementptr inbounds [1024 x i8], [1024 x i8]* @B, i64 0, i64 %iv.1
71 %out2 = getelementptr inbounds [1024 x i8], [1024 x i8]* @B, i64 0, i64 %iv.2
72 %out3 = getelementptr inbounds [1024 x i8], [1024 x i8]* @B, i64 0, i64 %iv.3
73 %out4 = getelementptr inbounds [1024 x i8], [1024 x i8]* @B, i64 0, i64 %iv.4
74 %out5 = getelementptr inbounds [1024 x i8], [1024 x i8]* @B, i64 0, i64 %iv.5
76 store i8 %v0, i8* %out0
77 store i8 %v1, i8* %out1
78 store i8 %v2, i8* %out2
79 store i8 %v3, i8* %out3
80 store i8 %v4, i8* %out4
81 store i8 %v5, i8* %out5
83 %iv.next = add nuw nsw i64 %iv.0, 6
84 %cmp = icmp ult i64 %iv.next, 1024
85 br i1 %cmp, label %for.body, label %for.cond.cleanup