89 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 i32 %v3, ptr %out3"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 i32] 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 i32 %v3, ptr %out3, align 417; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: store i32 %v3, ptr %out3, align 418; SSE2: LV: Found an estimated cost of 28 for VF 2 For instruction: store i32 %v3, ptr %out3, align 419; SSE2: LV: Found an estimated cost of 60 for VF 4 For instruction: store i32 %v3, ptr %out3, align 420; SSE2: LV: Found an estimated cost of 120 for VF 8 For instruction: store i32 %v3, ptr %out3, align 421; SSE2: LV: Found an estimated cost of 240 for VF 16 For instruction: store i32 %v3, ptr %out3, align 422;23; AVX1-LABEL: 'test'24; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: store i32 %v3, ptr %out3, align 425; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: store i32 %v3, ptr %out3, align 426; AVX1: LV: Found an estimated cost of 18 for VF 2 For instruction: store i32 %v3, ptr %out3, align 427; AVX1: LV: Found an estimated cost of 36 for VF 4 For instruction: store i32 %v3, ptr %out3, align 428; AVX1: LV: Found an estimated cost of 76 for VF 8 For instruction: store i32 %v3, ptr %out3, align 429; AVX1: LV: Found an estimated cost of 152 for VF 16 For instruction: store i32 %v3, ptr %out3, align 430; AVX1: LV: Found an estimated cost of 304 for VF 32 For instruction: store i32 %v3, ptr %out3, align 431;32; AVX2-LABEL: 'test'33; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: store i32 %v3, ptr %out3, align 434; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: store i32 %v3, ptr %out3, align 435; AVX2: LV: Found an estimated cost of 6 for VF 2 For instruction: store i32 %v3, ptr %out3, align 436; AVX2: LV: Found an estimated cost of 8 for VF 4 For instruction: store i32 %v3, ptr %out3, align 437; AVX2: LV: Found an estimated cost of 20 for VF 8 For instruction: store i32 %v3, ptr %out3, align 438; AVX2: LV: Found an estimated cost of 40 for VF 16 For instruction: store i32 %v3, ptr %out3, align 439; AVX2: LV: Found an estimated cost of 80 for VF 32 For instruction: store i32 %v3, ptr %out3, align 440;41; AVX512-LABEL: 'test'42; AVX512: LV: Found an estimated cost of 1 for VF 1 For instruction: store i32 %v3, ptr %out3, align 443; AVX512: LV: Found an estimated cost of 1 for VF 1 For instruction: store i32 %v3, ptr %out3, align 444; AVX512: LV: Found an estimated cost of 5 for VF 2 For instruction: store i32 %v3, ptr %out3, align 445; AVX512: LV: Found an estimated cost of 5 for VF 4 For instruction: store i32 %v3, ptr %out3, align 446; AVX512: LV: Found an estimated cost of 11 for VF 8 For instruction: store i32 %v3, ptr %out3, align 447; AVX512: LV: Found an estimated cost of 22 for VF 16 For instruction: store i32 %v3, ptr %out3, align 448; AVX512: LV: Found an estimated cost of 44 for VF 32 For instruction: store i32 %v3, ptr %out3, align 449; AVX512: LV: Found an estimated cost of 88 for VF 64 For instruction: store i32 %v3, ptr %out3, align 450;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 %iv.2 = add nuw nsw i64 %iv, 260 %iv.3 = add nuw nsw i64 %iv, 361 62 %in = getelementptr inbounds [1024 x i8], ptr @A, i64 0, i64 %iv.063 %v.narrow = load i8, ptr %in64 65 %v = zext i8 %v.narrow to i3266 67 %v0 = add i32 %v, 068 %v1 = add i32 %v, 169 %v2 = add i32 %v, 270 %v3 = add i32 %v, 371 72 %out0 = getelementptr inbounds [1024 x i32], ptr @B, i64 0, i64 %iv.073 %out1 = getelementptr inbounds [1024 x i32], ptr @B, i64 0, i64 %iv.174 %out2 = getelementptr inbounds [1024 x i32], ptr @B, i64 0, i64 %iv.275 %out3 = getelementptr inbounds [1024 x i32], ptr @B, i64 0, i64 %iv.376 77 store i32 %v0, ptr %out078 store i32 %v1, ptr %out179 store i32 %v2, ptr %out280 store i32 %v3, ptr %out381 82 %iv.next = add nuw nsw i64 %iv.0, 483 %cmp = icmp ult i64 %iv.next, 102484 br i1 %cmp, label %for.body, label %for.cond.cleanup85 86for.cond.cleanup:87 ret void88}89