80 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 i16 %valB, ptr %out, align 2"2; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+sse2 --debug-only=loop-vectorize --disable-output < %s 2>&1 | FileCheck %s --check-prefixes=SSE3; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+sse4.2 --debug-only=loop-vectorize --disable-output < %s 2>&1 | FileCheck %s --check-prefixes=SSE4; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx --debug-only=loop-vectorize --disable-output < %s 2>&1 | FileCheck %s --check-prefixes=AVX15; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx2,-fast-gather --debug-only=loop-vectorize --disable-output < %s 2>&1 | FileCheck %s --check-prefixes=AVX26; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx2,+fast-gather --debug-only=loop-vectorize --disable-output < %s 2>&1 | FileCheck %s --check-prefixes=AVX27; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx512bw --debug-only=loop-vectorize --disable-output < %s 2>&1 | FileCheck %s --check-prefixes=AVX5128 9; REQUIRES: asserts10 11target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"12target triple = "x86_64-unknown-linux-gnu"13 14@A = global [1024 x i8] zeroinitializer, align 12815@B = global [1024 x i16] zeroinitializer, align 12816 17define void @test(ptr %C) {18; SSE-LABEL: 'test'19; SSE: LV: Found an estimated cost of 1 for VF 1 For instruction: store i16 %valB, ptr %out, align 220; SSE: LV: Found an estimated cost of 2 for VF 2 For instruction: store i16 %valB, ptr %out, align 221; SSE: LV: Found an estimated cost of 4 for VF 4 For instruction: store i16 %valB, ptr %out, align 222; SSE: LV: Found an estimated cost of 8 for VF 8 For instruction: store i16 %valB, ptr %out, align 223; SSE: LV: Found an estimated cost of 16 for VF 16 For instruction: store i16 %valB, ptr %out, align 224;25; AVX1-LABEL: 'test'26; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: store i16 %valB, ptr %out, align 227; AVX1: LV: Found an estimated cost of 2 for VF 2 For instruction: store i16 %valB, ptr %out, align 228; AVX1: LV: Found an estimated cost of 4 for VF 4 For instruction: store i16 %valB, ptr %out, align 229; AVX1: LV: Found an estimated cost of 8 for VF 8 For instruction: store i16 %valB, ptr %out, align 230; AVX1: LV: Found an estimated cost of 16 for VF 16 For instruction: store i16 %valB, ptr %out, align 231; AVX1: LV: Found an estimated cost of 33 for VF 32 For instruction: store i16 %valB, ptr %out, align 232;33; AVX2-LABEL: 'test'34; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: store i16 %valB, ptr %out, align 235; AVX2: LV: Found an estimated cost of 2 for VF 2 For instruction: store i16 %valB, ptr %out, align 236; AVX2: LV: Found an estimated cost of 4 for VF 4 For instruction: store i16 %valB, ptr %out, align 237; AVX2: LV: Found an estimated cost of 8 for VF 8 For instruction: store i16 %valB, ptr %out, align 238; AVX2: LV: Found an estimated cost of 16 for VF 16 For instruction: store i16 %valB, ptr %out, align 239; AVX2: LV: Found an estimated cost of 33 for VF 32 For instruction: store i16 %valB, ptr %out, align 240;41; AVX512-LABEL: 'test'42; AVX512: LV: Found an estimated cost of 1 for VF 1 For instruction: store i16 %valB, ptr %out, align 243; AVX512: LV: Found an estimated cost of 2 for VF 2 For instruction: store i16 %valB, ptr %out, align 244; AVX512: LV: Found an estimated cost of 2 for VF 4 For instruction: store i16 %valB, ptr %out, align 245; AVX512: LV: Found an estimated cost of 1 for VF 8 For instruction: store i16 %valB, ptr %out, align 246; AVX512: LV: Found an estimated cost of 1 for VF 16 For instruction: store i16 %valB, ptr %out, align 247; AVX512: LV: Found an estimated cost of 1 for VF 32 For instruction: store i16 %valB, ptr %out, align 248; AVX512: LV: Found an estimated cost of 2 for VF 64 For instruction: store i16 %valB, ptr %out, align 249;50entry:51 br label %for.body52 53for.body:54 %iv = phi i64 [ 0, %entry ], [ %iv.next, %end ]55 56 %inB = getelementptr inbounds [1024 x i16], ptr @B, i64 0, i64 %iv57 %valB = load i16, ptr %inB58 59 %inA = getelementptr inbounds [1024 x i8], ptr @A, i64 0, i64 %iv60 %valA = load i8, ptr %inA61 %canStore = icmp ne i8 %valA, 062 br i1 %canStore, label %store, label %mask63 64store:65 %out = getelementptr inbounds [1024 x i16], ptr %C, i64 0, i64 %iv66 store i16 %valB, ptr %out67 br label %end68 69mask:70 br label %end71 72end:73 %iv.next = add nuw nsw i64 %iv, 174 %cmp = icmp ult i64 %iv.next, 102475 br i1 %cmp, label %for.body, label %for.cond.cleanup76 77for.cond.cleanup:78 ret void79}80