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=+sse42 --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,SSE42
3 ; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,AVX1
4 ; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx2,-fast-gather --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,AVX2-SLOWGATHER
5 ; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx2,+fast-gather --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,AVX2-FASTGATHER
6 ; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx512bw --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,AVX512
10 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
11 target triple = "x86_64-unknown-linux-gnu"
13 @A = global [1024 x i8] zeroinitializer, align 128
14 @B = global [1024 x i8] zeroinitializer, align 128
16 ; CHECK: LV: Checking a loop in "test"
18 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: store i8 %valB, i8* %out, align 1
19 ; SSE2: LV: Found an estimated cost of 2 for VF 2 For instruction: store i8 %valB, i8* %out, align 1
20 ; SSE2: LV: Found an estimated cost of 5 for VF 4 For instruction: store i8 %valB, i8* %out, align 1
21 ; SSE2: LV: Found an estimated cost of 11 for VF 8 For instruction: store i8 %valB, i8* %out, align 1
22 ; SSE2: LV: Found an estimated cost of 23 for VF 16 For instruction: store i8 %valB, i8* %out, align 1
24 ; SSE42: LV: Found an estimated cost of 1 for VF 1 For instruction: store i8 %valB, i8* %out, align 1
25 ; SSE42: LV: Found an estimated cost of 2 for VF 2 For instruction: store i8 %valB, i8* %out, align 1
26 ; SSE42: LV: Found an estimated cost of 5 for VF 4 For instruction: store i8 %valB, i8* %out, align 1
27 ; SSE42: LV: Found an estimated cost of 11 for VF 8 For instruction: store i8 %valB, i8* %out, align 1
28 ; SSE42: LV: Found an estimated cost of 23 for VF 16 For instruction: store i8 %valB, i8* %out, align 1
30 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: store i8 %valB, i8* %out, align 1
31 ; AVX1: LV: Found an estimated cost of 2 for VF 2 For instruction: store i8 %valB, i8* %out, align 1
32 ; AVX1: LV: Found an estimated cost of 4 for VF 4 For instruction: store i8 %valB, i8* %out, align 1
33 ; AVX1: LV: Found an estimated cost of 8 for VF 8 For instruction: store i8 %valB, i8* %out, align 1
34 ; AVX1: LV: Found an estimated cost of 16 for VF 16 For instruction: store i8 %valB, i8* %out, align 1
35 ; AVX1: LV: Found an estimated cost of 40 for VF 32 For instruction: store i8 %valB, i8* %out, align 1
37 ; AVX2-SLOWGATHER: LV: Found an estimated cost of 1 for VF 1 For instruction: store i8 %valB, i8* %out, align 1
38 ; AVX2-SLOWGATHER: LV: Found an estimated cost of 2 for VF 2 For instruction: store i8 %valB, i8* %out, align 1
39 ; AVX2-SLOWGATHER: LV: Found an estimated cost of 4 for VF 4 For instruction: store i8 %valB, i8* %out, align 1
40 ; AVX2-SLOWGATHER: LV: Found an estimated cost of 8 for VF 8 For instruction: store i8 %valB, i8* %out, align 1
41 ; AVX2-SLOWGATHER: LV: Found an estimated cost of 16 for VF 16 For instruction: store i8 %valB, i8* %out, align 1
42 ; AVX2-SLOWGATHER: LV: Found an estimated cost of 40 for VF 32 For instruction: store i8 %valB, i8* %out, align 1
44 ; AVX2-FASTGATHER: LV: Found an estimated cost of 1 for VF 1 For instruction: store i8 %valB, i8* %out, align 1
45 ; AVX2-FASTGATHER: LV: Found an estimated cost of 2 for VF 2 For instruction: store i8 %valB, i8* %out, align 1
46 ; AVX2-FASTGATHER: LV: Found an estimated cost of 4 for VF 4 For instruction: store i8 %valB, i8* %out, align 1
47 ; AVX2-FASTGATHER: LV: Found an estimated cost of 8 for VF 8 For instruction: store i8 %valB, i8* %out, align 1
48 ; AVX2-FASTGATHER: LV: Found an estimated cost of 16 for VF 16 For instruction: store i8 %valB, i8* %out, align 1
49 ; AVX2-FASTGATHER: LV: Found an estimated cost of 40 for VF 32 For instruction: store i8 %valB, i8* %out, align 1
51 ; AVX512: LV: Found an estimated cost of 1 for VF 1 For instruction: store i8 %valB, i8* %out, align 1
52 ; AVX512: LV: Found an estimated cost of 2 for VF 2 For instruction: store i8 %valB, i8* %out, align 1
53 ; AVX512: LV: Found an estimated cost of 2 for VF 4 For instruction: store i8 %valB, i8* %out, align 1
54 ; AVX512: LV: Found an estimated cost of 2 for VF 8 For instruction: store i8 %valB, i8* %out, align 1
55 ; AVX512: LV: Found an estimated cost of 1 for VF 16 For instruction: store i8 %valB, i8* %out, align 1
56 ; AVX512: LV: Found an estimated cost of 1 for VF 32 For instruction: store i8 %valB, i8* %out, align 1
57 ; AVX512: LV: Found an estimated cost of 1 for VF 64 For instruction: store i8 %valB, i8* %out, align 1
59 ; CHECK-NOT: LV: Found an estimated cost of {{[0-9]+}} for VF {{[0-9]+}} For instruction: store i8 %valB, i8* %out, align 1
60 define void @test([1024 x i8]* %C) {
65 %iv = phi i64 [ 0, %entry ], [ %iv.next, %end ]
67 %inB = getelementptr inbounds [1024 x i8], [1024 x i8]* @B, i64 0, i64 %iv
68 %valB = load i8, i8* %inB
70 %inA = getelementptr inbounds [1024 x i8], [1024 x i8]* @A, i64 0, i64 %iv
71 %valA = load i8, i8* %inA
72 %canStore = icmp ne i8 %valA, 0
73 br i1 %canStore, label %store, label %mask
76 %out = getelementptr inbounds [1024 x i8], [1024 x i8]* %C, i64 0, i64 %iv
77 store i8 %valB, i8* %out
84 %iv.next = add nuw nsw i64 %iv, 1
85 %cmp = icmp ult i64 %iv.next, 1024
86 br i1 %cmp, label %for.body, label %for.cond.cleanup