81 lines · plain
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*store i64 %v1, ptr %out1"2; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+sse2 --debug-only=loop-vectorize --disable-output < %s 2>&1 | FileCheck %s --check-prefix=SSE23; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx --debug-only=loop-vectorize --disable-output < %s 2>&1 | FileCheck %s --check-prefix=AVX14; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx2 --debug-only=loop-vectorize --disable-output < %s 2>&1 | FileCheck %s --check-prefix=AVX25; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx512vl --debug-only=loop-vectorize --disable-output < %s 2>&1 | FileCheck %s --check-prefix=AVX5126; REQUIRES: asserts7 8target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"9target triple = "x86_64-unknown-linux-gnu"10 11@A = global [1024 x i8] zeroinitializer, align 12812@B = global [1024 x i64] zeroinitializer, align 12813 14define void @test() {15; SSE2-LABEL: 'test'16; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: store i64 %v1, ptr %out1, align 817; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: store i64 %v1, ptr %out1, align 818; SSE2: LV: Found an estimated cost of 14 for VF 2 For instruction: store i64 %v1, ptr %out1, align 819; SSE2: LV: Found an estimated cost of 28 for VF 4 For instruction: store i64 %v1, ptr %out1, align 820; SSE2: LV: Found an estimated cost of 56 for VF 8 For instruction: store i64 %v1, ptr %out1, align 821; SSE2: LV: Found an estimated cost of 112 for VF 16 For instruction: store i64 %v1, ptr %out1, align 822;23; AVX1-LABEL: 'test'24; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: store i64 %v1, ptr %out1, align 825; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: store i64 %v1, ptr %out1, align 826; AVX1: LV: Found an estimated cost of 10 for VF 2 For instruction: store i64 %v1, ptr %out1, align 827; AVX1: LV: Found an estimated cost of 22 for VF 4 For instruction: store i64 %v1, ptr %out1, align 828; AVX1: LV: Found an estimated cost of 44 for VF 8 For instruction: store i64 %v1, ptr %out1, align 829; AVX1: LV: Found an estimated cost of 88 for VF 16 For instruction: store i64 %v1, ptr %out1, align 830; AVX1: LV: Found an estimated cost of 176 for VF 32 For instruction: store i64 %v1, ptr %out1, align 831;32; AVX2-LABEL: 'test'33; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: store i64 %v1, ptr %out1, align 834; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: store i64 %v1, ptr %out1, align 835; AVX2: LV: Found an estimated cost of 3 for VF 2 For instruction: store i64 %v1, ptr %out1, align 836; AVX2: LV: Found an estimated cost of 6 for VF 4 For instruction: store i64 %v1, ptr %out1, align 837; AVX2: LV: Found an estimated cost of 12 for VF 8 For instruction: store i64 %v1, ptr %out1, align 838; AVX2: LV: Found an estimated cost of 24 for VF 16 For instruction: store i64 %v1, ptr %out1, align 839; AVX2: LV: Found an estimated cost of 48 for VF 32 For instruction: store i64 %v1, ptr %out1, align 840;41; AVX512-LABEL: 'test'42; AVX512: LV: Found an estimated cost of 1 for VF 1 For instruction: store i64 %v1, ptr %out1, align 843; AVX512: LV: Found an estimated cost of 1 for VF 1 For instruction: store i64 %v1, ptr %out1, align 844; AVX512: LV: Found an estimated cost of 2 for VF 2 For instruction: store i64 %v1, ptr %out1, align 845; AVX512: LV: Found an estimated cost of 2 for VF 4 For instruction: store i64 %v1, ptr %out1, align 846; AVX512: LV: Found an estimated cost of 5 for VF 8 For instruction: store i64 %v1, ptr %out1, align 847; AVX512: LV: Found an estimated cost of 10 for VF 16 For instruction: store i64 %v1, ptr %out1, align 848; AVX512: LV: Found an estimated cost of 20 for VF 32 For instruction: store i64 %v1, ptr %out1, align 849; AVX512: LV: Found an estimated cost of 40 for VF 64 For instruction: store i64 %v1, ptr %out1, align 850;51entry:52 br label %for.body53 54for.body:55 %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ]56 57 %iv.0 = add nuw nsw i64 %iv, 058 %iv.1 = add nuw nsw i64 %iv, 159 60 %in = getelementptr inbounds [1024 x i8], ptr @A, i64 0, i64 %iv.061 %v.narrow = load i8, ptr %in62 63 %v = zext i8 %v.narrow to i6464 65 %v0 = add i64 %v, 066 %v1 = add i64 %v, 167 68 %out0 = getelementptr inbounds [1024 x i64], ptr @B, i64 0, i64 %iv.069 %out1 = getelementptr inbounds [1024 x i64], ptr @B, i64 0, i64 %iv.170 71 store i64 %v0, ptr %out072 store i64 %v1, ptr %out173 74 %iv.next = add nuw nsw i64 %iv.0, 275 %cmp = icmp ult i64 %iv.next, 102476 br i1 %cmp, label %for.body, label %for.cond.cleanup77 78for.cond.cleanup:79 ret void80}81