107 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*%v0 = load i16, ptr %in0"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=AVX512DQ6; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx512vl,+avx512bw --debug-only=loop-vectorize --disable-output < %s 2>&1 | FileCheck %s --check-prefix=AVX512BW7; REQUIRES: asserts8 9target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"10target triple = "x86_64-unknown-linux-gnu"11 12@A = global [1024 x i16] zeroinitializer, align 12813@B = global [1024 x i8] zeroinitializer, align 12814 15define void @test() {16; SSE2-LABEL: 'test'17; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i16, ptr %in0, align 218; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i16, ptr %in0, align 219; SSE2: LV: Found an estimated cost of 26 for VF 2 For instruction: %v0 = load i16, ptr %in0, align 220; SSE2: LV: Found an estimated cost of 51 for VF 4 For instruction: %v0 = load i16, ptr %in0, align 221; SSE2: LV: Found an estimated cost of 102 for VF 8 For instruction: %v0 = load i16, ptr %in0, align 222; SSE2: LV: Found an estimated cost of 204 for VF 16 For instruction: %v0 = load i16, ptr %in0, align 223;24; AVX1-LABEL: 'test'25; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i16, ptr %in0, align 226; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i16, ptr %in0, align 227; AVX1: LV: Found an estimated cost of 28 for VF 2 For instruction: %v0 = load i16, ptr %in0, align 228; AVX1: LV: Found an estimated cost of 51 for VF 4 For instruction: %v0 = load i16, ptr %in0, align 229; AVX1: LV: Found an estimated cost of 102 for VF 8 For instruction: %v0 = load i16, ptr %in0, align 230; AVX1: LV: Found an estimated cost of 210 for VF 16 For instruction: %v0 = load i16, ptr %in0, align 231; AVX1: LV: Found an estimated cost of 420 for VF 32 For instruction: %v0 = load i16, ptr %in0, align 232;33; AVX2-LABEL: 'test'34; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i16, ptr %in0, align 235; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i16, ptr %in0, align 236; AVX2: LV: Found an estimated cost of 16 for VF 2 For instruction: %v0 = load i16, ptr %in0, align 237; AVX2: LV: Found an estimated cost of 11 for VF 4 For instruction: %v0 = load i16, ptr %in0, align 238; AVX2: LV: Found an estimated cost of 42 for VF 8 For instruction: %v0 = load i16, ptr %in0, align 239; AVX2: LV: Found an estimated cost of 112 for VF 16 For instruction: %v0 = load i16, ptr %in0, align 240; AVX2: LV: Found an estimated cost of 224 for VF 32 For instruction: %v0 = load i16, ptr %in0, align 241;42; AVX512DQ-LABEL: 'test'43; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i16, ptr %in0, align 244; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i16, ptr %in0, align 245; AVX512DQ: LV: Found an estimated cost of 16 for VF 2 For instruction: %v0 = load i16, ptr %in0, align 246; AVX512DQ: LV: Found an estimated cost of 12 for VF 4 For instruction: %v0 = load i16, ptr %in0, align 247; AVX512DQ: LV: Found an estimated cost of 41 for VF 8 For instruction: %v0 = load i16, ptr %in0, align 248; AVX512DQ: LV: Found an estimated cost of 109 for VF 16 For instruction: %v0 = load i16, ptr %in0, align 249; AVX512DQ: LV: Found an estimated cost of 218 for VF 32 For instruction: %v0 = load i16, ptr %in0, align 250; AVX512DQ: LV: Found an estimated cost of 852 for VF 64 For instruction: %v0 = load i16, ptr %in0, align 251;52; AVX512BW-LABEL: 'test'53; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i16, ptr %in0, align 254; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i16, ptr %in0, align 255; AVX512BW: LV: Found an estimated cost of 13 for VF 2 For instruction: %v0 = load i16, ptr %in0, align 256; AVX512BW: LV: Found an estimated cost of 13 for VF 4 For instruction: %v0 = load i16, ptr %in0, align 257; AVX512BW: LV: Found an estimated cost of 17 for VF 8 For instruction: %v0 = load i16, ptr %in0, align 258; AVX512BW: LV: Found an estimated cost of 33 for VF 16 For instruction: %v0 = load i16, ptr %in0, align 259; AVX512BW: LV: Found an estimated cost of 81 for VF 32 For instruction: %v0 = load i16, ptr %in0, align 260; AVX512BW: LV: Found an estimated cost of 342 for VF 64 For instruction: %v0 = load i16, ptr %in0, align 261;62entry:63 br label %for.body64 65for.body:66 %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ]67 68 %iv.0 = add nuw nsw i64 %iv, 069 %iv.1 = add nuw nsw i64 %iv, 170 %iv.2 = add nuw nsw i64 %iv, 271 %iv.3 = add nuw nsw i64 %iv, 372 %iv.4 = add nuw nsw i64 %iv, 473 %iv.5 = add nuw nsw i64 %iv, 574 75 %in0 = getelementptr inbounds [1024 x i16], ptr @A, i64 0, i64 %iv.076 %in1 = getelementptr inbounds [1024 x i16], ptr @A, i64 0, i64 %iv.177 %in2 = getelementptr inbounds [1024 x i16], ptr @A, i64 0, i64 %iv.278 %in3 = getelementptr inbounds [1024 x i16], ptr @A, i64 0, i64 %iv.379 %in4 = getelementptr inbounds [1024 x i16], ptr @A, i64 0, i64 %iv.480 %in5 = getelementptr inbounds [1024 x i16], ptr @A, i64 0, i64 %iv.581 82 %v0 = load i16, ptr %in083 %v1 = load i16, ptr %in184 %v2 = load i16, ptr %in285 %v3 = load i16, ptr %in386 %v4 = load i16, ptr %in487 %v5 = load i16, ptr %in588 89 %reduce.add.0 = add i16 %v0, %v190 %reduce.add.1 = add i16 %reduce.add.0, %v291 %reduce.add.2 = add i16 %reduce.add.1, %v392 %reduce.add.3 = add i16 %reduce.add.2, %v493 %reduce.add.4 = add i16 %reduce.add.3, %v594 95 %reduce.add.4.narrow = trunc i16 %reduce.add.4 to i896 97 %out = getelementptr inbounds [1024 x i8], ptr @B, i64 0, i64 %iv.098 store i8 %reduce.add.4.narrow, ptr %out99 100 %iv.next = add nuw nsw i64 %iv.0, 6101 %cmp = icmp ult i64 %iv.next, 1024102 br i1 %cmp, label %for.body, label %for.cond.cleanup103 104for.cond.cleanup:105 ret void106}107