1 ; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py UTC_ARGS: --filter "LV: Found an estimated cost of [0-9]+ for VF [0-9]+ For instruction:\s*%valB = load i32, ptr %inB, align 4"
2 ; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+sse2 --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=SSE2
3 ; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+sse4.2 --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=SSE42
4 ; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=AVX1
5 ; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx2,-fast-gather --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=AVX2-SLOWGATHER
6 ; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx2,+fast-gather --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=AVX2-FASTGATHER
7 ; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx512bw --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=AVX512
11 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
12 target triple = "x86_64-unknown-linux-gnu"
14 @A = global [1024 x i8] zeroinitializer, align 128
15 @B = global [1024 x i32] zeroinitializer, align 128
16 @C = global [1024 x i32] zeroinitializer, align 128
20 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %valB = load i32, ptr %inB, align 4
21 ; SSE2: LV: Found an estimated cost of 25 for VF 2 For instruction: %valB = load i32, ptr %inB, align 4
22 ; SSE2: LV: Found an estimated cost of 51 for VF 4 For instruction: %valB = load i32, ptr %inB, align 4
23 ; SSE2: LV: Found an estimated cost of 102 for VF 8 For instruction: %valB = load i32, ptr %inB, align 4
24 ; SSE2: LV: Found an estimated cost of 204 for VF 16 For instruction: %valB = load i32, ptr %inB, align 4
27 ; SSE42: LV: Found an estimated cost of 1 for VF 1 For instruction: %valB = load i32, ptr %inB, align 4
28 ; SSE42: LV: Found an estimated cost of 24 for VF 2 For instruction: %valB = load i32, ptr %inB, align 4
29 ; SSE42: LV: Found an estimated cost of 48 for VF 4 For instruction: %valB = load i32, ptr %inB, align 4
30 ; SSE42: LV: Found an estimated cost of 96 for VF 8 For instruction: %valB = load i32, ptr %inB, align 4
31 ; SSE42: LV: Found an estimated cost of 192 for VF 16 For instruction: %valB = load i32, ptr %inB, align 4
34 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %valB = load i32, ptr %inB, align 4
35 ; AVX1: LV: Found an estimated cost of 24 for VF 2 For instruction: %valB = load i32, ptr %inB, align 4
36 ; AVX1: LV: Found an estimated cost of 48 for VF 4 For instruction: %valB = load i32, ptr %inB, align 4
37 ; AVX1: LV: Found an estimated cost of 97 for VF 8 For instruction: %valB = load i32, ptr %inB, align 4
38 ; AVX1: LV: Found an estimated cost of 194 for VF 16 For instruction: %valB = load i32, ptr %inB, align 4
39 ; AVX1: LV: Found an estimated cost of 388 for VF 32 For instruction: %valB = load i32, ptr %inB, align 4
41 ; AVX2-SLOWGATHER-LABEL: 'test'
42 ; AVX2-SLOWGATHER: LV: Found an estimated cost of 1 for VF 1 For instruction: %valB = load i32, ptr %inB, align 4
43 ; AVX2-SLOWGATHER: LV: Found an estimated cost of 4 for VF 2 For instruction: %valB = load i32, ptr %inB, align 4
44 ; AVX2-SLOWGATHER: LV: Found an estimated cost of 8 for VF 4 For instruction: %valB = load i32, ptr %inB, align 4
45 ; AVX2-SLOWGATHER: LV: Found an estimated cost of 17 for VF 8 For instruction: %valB = load i32, ptr %inB, align 4
46 ; AVX2-SLOWGATHER: LV: Found an estimated cost of 34 for VF 16 For instruction: %valB = load i32, ptr %inB, align 4
47 ; AVX2-SLOWGATHER: LV: Found an estimated cost of 68 for VF 32 For instruction: %valB = load i32, ptr %inB, align 4
49 ; AVX2-FASTGATHER-LABEL: 'test'
50 ; AVX2-FASTGATHER: LV: Found an estimated cost of 1 for VF 1 For instruction: %valB = load i32, ptr %inB, align 4
51 ; AVX2-FASTGATHER: LV: Found an estimated cost of 4 for VF 2 For instruction: %valB = load i32, ptr %inB, align 4
52 ; AVX2-FASTGATHER: LV: Found an estimated cost of 6 for VF 4 For instruction: %valB = load i32, ptr %inB, align 4
53 ; AVX2-FASTGATHER: LV: Found an estimated cost of 12 for VF 8 For instruction: %valB = load i32, ptr %inB, align 4
54 ; AVX2-FASTGATHER: LV: Found an estimated cost of 24 for VF 16 For instruction: %valB = load i32, ptr %inB, align 4
55 ; AVX2-FASTGATHER: LV: Found an estimated cost of 48 for VF 32 For instruction: %valB = load i32, ptr %inB, align 4
57 ; AVX512-LABEL: 'test'
58 ; AVX512: LV: Found an estimated cost of 1 for VF 1 For instruction: %valB = load i32, ptr %inB, align 4
59 ; AVX512: LV: Found an estimated cost of 6 for VF 2 For instruction: %valB = load i32, ptr %inB, align 4
60 ; AVX512: LV: Found an estimated cost of 13 for VF 4 For instruction: %valB = load i32, ptr %inB, align 4
61 ; AVX512: LV: Found an estimated cost of 10 for VF 8 For instruction: %valB = load i32, ptr %inB, align 4
62 ; AVX512: LV: Found an estimated cost of 18 for VF 16 For instruction: %valB = load i32, ptr %inB, align 4
63 ; AVX512: LV: Found an estimated cost of 36 for VF 32 For instruction: %valB = load i32, ptr %inB, align 4
64 ; AVX512: LV: Found an estimated cost of 72 for VF 64 For instruction: %valB = load i32, ptr %inB, align 4
70 %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ]
72 %inA = getelementptr inbounds [1024 x i8], ptr @A, i64 0, i64 %iv
73 %valA = load i8, ptr %inA
74 %valA.ext = sext i8 %valA to i64
76 %inB = getelementptr inbounds [1024 x i32], ptr @B, i64 0, i64 %valA.ext
77 %valB = load i32, ptr %inB
79 %out = getelementptr inbounds [1024 x i32], ptr @C, i64 0, i64 %iv
80 store i32 %valB, ptr %out
82 %iv.next = add nuw nsw i64 %iv, 1
83 %cmp = icmp ult i64 %iv.next, 1024
84 br i1 %cmp, label %for.body, label %for.cond.cleanup