| 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*%v0 = load float, ptr %in0, align 4" --filter "LV: Found an estimated cost of [0-9]+ for VF [0-9]+ For instruction:\s*%v0 = load float, float\* %in0, align 4" |
| 2 |
; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx512bw --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,AVX512 |
| 3 |
; REQUIRES: asserts |
| 4 |
|
| 5 |
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128" |
| 6 |
target triple = "x86_64-unknown-linux-gnu" |
| 7 |
|
| 8 |
@A = global [1024 x float] zeroinitializer, align 128 |
| 9 |
@B = global [1024 x i8] zeroinitializer, align 128 |
| 10 |
|
| 11 |
define void @test() { |
| 12 |
; CHECK-LABEL: 'test' |
| 13 |
; CHECK: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load float, ptr %in0, align 4 |
| 14 |
; CHECK: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load float, ptr %in0, align 4 |
| 15 |
; CHECK: LV: Found an estimated cost of 3 for VF 2 For instruction: %v0 = load float, ptr %in0, align 4 |
| 16 |
; CHECK: LV: Found an estimated cost of 3 for VF 4 For instruction: %v0 = load float, ptr %in0, align 4 |
| 17 |
; CHECK: LV: Found an estimated cost of 3 for VF 8 For instruction: %v0 = load float, ptr %in0, align 4 |
| 18 |
; CHECK: LV: Found an estimated cost of 5 for VF 16 For instruction: %v0 = load float, ptr %in0, align 4 |
| 19 |
; CHECK: LV: Found an estimated cost of 22 for VF 32 For instruction: %v0 = load float, ptr %in0, align 4 |
| 20 |
; CHECK: LV: Found an estimated cost of 92 for VF 64 For instruction: %v0 = load float, ptr %in0, align 4 |
| 21 |
; |
| 22 |
entry: |
| 23 |
br label %for.body |
| 24 |
|
| 25 |
for.body: |
| 26 |
%iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ] |
| 27 |
|
| 28 |
%iv.0 = add nuw nsw i64 %iv, 0 |
| 29 |
%iv.1 = add nuw nsw i64 %iv, 1 |
| 30 |
|
| 31 |
%in0 = getelementptr inbounds [1024 x float], ptr @A, i64 0, i64 %iv.0 |
| 32 |
%in1 = getelementptr inbounds [1024 x float], ptr @A, i64 0, i64 %iv.1 |
| 33 |
|
| 34 |
%v0 = load float, ptr %in0 |
| 35 |
%v1 = load float, ptr %in1 |
| 36 |
|
| 37 |
%reduce.add.0 = fadd float %v0, %v1 |
| 38 |
|
| 39 |
%reduce.add.0.narrow = fptoui float %reduce.add.0 to i8 |
| 40 |
|
| 41 |
%out = getelementptr inbounds [1024 x i8], ptr @B, i64 0, i64 %iv.0 |
| 42 |
store i8 %reduce.add.0.narrow, ptr %out |
| 43 |
|
| 44 |
%iv.next = add nuw nsw i64 %iv.0, 2 |
| 45 |
%cmp = icmp ult i64 %iv.next, 1024 |
| 46 |
br i1 %cmp, label %for.body, label %for.cond.cleanup |
| 47 |
|
| 48 |
for.cond.cleanup: |
| 49 |
ret void |
| 50 |
} |
| 51 |
;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line: |
| 52 |
; AVX512: {{.*}} |