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 i16] 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 i16 %v4, i16* %out4, align 2
16 ; SSE2: LV: Found an estimated cost of 22 for VF 2 For instruction: store i16 %v4, i16* %out4, align 2
17 ; SSE2: LV: Found an estimated cost of 43 for VF 4 For instruction: store i16 %v4, i16* %out4, align 2
18 ; SSE2: LV: Found an estimated cost of 85 for VF 8 For instruction: store i16 %v4, i16* %out4, align 2
19 ; SSE2: LV: Found an estimated cost of 170 for VF 16 For instruction: store i16 %v4, i16* %out4, align 2
21 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: store i16 %v4, i16* %out4, align 2
22 ; AVX1: LV: Found an estimated cost of 27 for VF 2 For instruction: store i16 %v4, i16* %out4, align 2
23 ; AVX1: LV: Found an estimated cost of 45 for VF 4 For instruction: store i16 %v4, i16* %out4, align 2
24 ; AVX1: LV: Found an estimated cost of 88 for VF 8 For instruction: store i16 %v4, i16* %out4, align 2
25 ; AVX1: LV: Found an estimated cost of 215 for VF 16 For instruction: store i16 %v4, i16* %out4, align 2
26 ; AVX1: LV: Found an estimated cost of 430 for VF 32 For instruction: store i16 %v4, i16* %out4, align 2
28 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: store i16 %v4, i16* %out4, align 2
29 ; AVX2: LV: Found an estimated cost of 27 for VF 2 For instruction: store i16 %v4, i16* %out4, align 2
30 ; AVX2: LV: Found an estimated cost of 45 for VF 4 For instruction: store i16 %v4, i16* %out4, align 2
31 ; AVX2: LV: Found an estimated cost of 88 for VF 8 For instruction: store i16 %v4, i16* %out4, align 2
32 ; AVX2: LV: Found an estimated cost of 215 for VF 16 For instruction: store i16 %v4, i16* %out4, align 2
33 ; AVX2: LV: Found an estimated cost of 430 for VF 32 For instruction: store i16 %v4, i16* %out4, align 2
35 ; AVX512: LV: Found an estimated cost of 1 for VF 1 For instruction: store i16 %v4, i16* %out4, align 2
36 ; AVX512: LV: Found an estimated cost of 11 for VF 2 For instruction: store i16 %v4, i16* %out4, align 2
37 ; AVX512: LV: Found an estimated cost of 11 for VF 4 For instruction: store i16 %v4, i16* %out4, align 2
38 ; AVX512: LV: Found an estimated cost of 22 for VF 8 For instruction: store i16 %v4, i16* %out4, align 2
39 ; AVX512: LV: Found an estimated cost of 33 for VF 16 For instruction: store i16 %v4, i16* %out4, align 2
40 ; AVX512: LV: Found an estimated cost of 55 for VF 32 For instruction: store i16 %v4, i16* %out4, align 2
41 ; AVX512: LV: Found an estimated cost of 110 for VF 64 For instruction: store i16 %v4, i16* %out4, align 2
43 ; CHECK-NOT: LV: Found an estimated cost of {{[0-9]+}} for VF {{[0-9]+}} For instruction: store i16 %v4, i16* %out4, align 2
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
58 %in = getelementptr inbounds [1024 x i8], [1024 x i8]* @A, i64 0, i64 %iv.0
59 %v.narrow = load i8, i8* %in
61 %v = zext i8 %v.narrow to i16
69 %out0 = getelementptr inbounds [1024 x i16], [1024 x i16]* @B, i64 0, i64 %iv.0
70 %out1 = getelementptr inbounds [1024 x i16], [1024 x i16]* @B, i64 0, i64 %iv.1
71 %out2 = getelementptr inbounds [1024 x i16], [1024 x i16]* @B, i64 0, i64 %iv.2
72 %out3 = getelementptr inbounds [1024 x i16], [1024 x i16]* @B, i64 0, i64 %iv.3
73 %out4 = getelementptr inbounds [1024 x i16], [1024 x i16]* @B, i64 0, i64 %iv.4
75 store i16 %v0, i16* %out0
76 store i16 %v1, i16* %out1
77 store i16 %v2, i16* %out2
78 store i16 %v3, i16* %out3
79 store i16 %v4, i16* %out4
81 %iv.next = add nuw nsw i64 %iv.0, 5
82 %cmp = icmp ult i64 %iv.next, 1024
83 br i1 %cmp, label %for.body, label %for.cond.cleanup