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 float] 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 float %v5, float* %out5, align 4
16 ; SSE2: LV: Found an estimated cost of 21 for VF 2 For instruction: store float %v5, float* %out5, align 4
17 ; SSE2: LV: Found an estimated cost of 48 for VF 4 For instruction: store float %v5, float* %out5, align 4
18 ; SSE2: LV: Found an estimated cost of 96 for VF 8 For instruction: store float %v5, float* %out5, align 4
20 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: store float %v5, float* %out5, align 4
21 ; AVX1: LV: Found an estimated cost of 20 for VF 2 For instruction: store float %v5, float* %out5, align 4
22 ; AVX1: LV: Found an estimated cost of 45 for VF 4 For instruction: store float %v5, float* %out5, align 4
23 ; AVX1: LV: Found an estimated cost of 114 for VF 8 For instruction: store float %v5, float* %out5, align 4
24 ; AVX1: LV: Found an estimated cost of 228 for VF 16 For instruction: store float %v5, float* %out5, align 4
26 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: store float %v5, float* %out5, align 4
27 ; AVX2: LV: Found an estimated cost of 11 for VF 2 For instruction: store float %v5, float* %out5, align 4
28 ; AVX2: LV: Found an estimated cost of 15 for VF 4 For instruction: store float %v5, float* %out5, align 4
29 ; AVX2: LV: Found an estimated cost of 39 for VF 8 For instruction: store float %v5, float* %out5, align 4
30 ; AVX2: LV: Found an estimated cost of 78 for VF 16 For instruction: store float %v5, float* %out5, align 4
32 ; AVX512: LV: Found an estimated cost of 1 for VF 1 For instruction: store float %v5, float* %out5, align 4
33 ; AVX512: LV: Found an estimated cost of 8 for VF 2 For instruction: store float %v5, float* %out5, align 4
34 ; AVX512: LV: Found an estimated cost of 17 for VF 4 For instruction: store float %v5, float* %out5, align 4
35 ; AVX512: LV: Found an estimated cost of 25 for VF 8 For instruction: store float %v5, float* %out5, align 4
36 ; AVX512: LV: Found an estimated cost of 51 for VF 16 For instruction: store float %v5, float* %out5, align 4
37 ; AVX512: LV: Found an estimated cost of 102 for VF 32 For instruction: store float %v5, float* %out5, align 4
38 ; AVX512: LV: Found an estimated cost of 204 for VF 64 For instruction: store float %v5, float* %out5, align 4
40 ; CHECK-NOT: LV: Found an estimated cost of {{[0-9]+}} for VF {{[0-9]+}} For instruction: store float %v5, float* %out5, align 4
47 %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ]
49 %iv.0 = add nuw nsw i64 %iv, 0
50 %iv.1 = add nuw nsw i64 %iv, 1
51 %iv.2 = add nuw nsw i64 %iv, 2
52 %iv.3 = add nuw nsw i64 %iv, 3
53 %iv.4 = add nuw nsw i64 %iv, 4
54 %iv.5 = add nuw nsw i64 %iv, 5
56 %in = getelementptr inbounds [1024 x i8], [1024 x i8]* @A, i64 0, i64 %iv.0
57 %v.narrow = load i8, i8* %in
59 %v = uitofp i8 %v.narrow to float
61 %v0 = fadd float %v, 0.0
62 %v1 = fadd float %v, 1.0
63 %v2 = fadd float %v, 2.0
64 %v3 = fadd float %v, 3.0
65 %v4 = fadd float %v, 4.0
66 %v5 = fadd float %v, 5.0
68 %out0 = getelementptr inbounds [1024 x float], [1024 x float]* @B, i64 0, i64 %iv.0
69 %out1 = getelementptr inbounds [1024 x float], [1024 x float]* @B, i64 0, i64 %iv.1
70 %out2 = getelementptr inbounds [1024 x float], [1024 x float]* @B, i64 0, i64 %iv.2
71 %out3 = getelementptr inbounds [1024 x float], [1024 x float]* @B, i64 0, i64 %iv.3
72 %out4 = getelementptr inbounds [1024 x float], [1024 x float]* @B, i64 0, i64 %iv.4
73 %out5 = getelementptr inbounds [1024 x float], [1024 x float]* @B, i64 0, i64 %iv.5
75 store float %v0, float* %out0
76 store float %v1, float* %out1
77 store float %v2, float* %out2
78 store float %v3, float* %out3
79 store float %v4, float* %out4
80 store float %v5, float* %out5
82 %iv.next = add nuw nsw i64 %iv.0, 6
83 %cmp = icmp ult i64 %iv.next, 1024
84 br i1 %cmp, label %for.body, label %for.cond.cleanup