[PATCH] D120438: [SLP][NFC] Test for a follow-up fix of the the vector min/max instrinsic cost calculation.

[Vasileios Porpodas via Phabricator via llvm-commits]

vporpo created this revision.
vporpo added reviewers: fhahn, ABataev, RKSimon.
vporpo requested review of this revision.
Herald added a project: LLVM.
Herald added a subscriber: llvm-commits.

The code in this test should not have been vectorized.
It looks worse than the scalar code.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D120438

Files:
  llvm/test/Transforms/SLPVectorizer/X86/max_intrinsic_cost.ll


Index: llvm/test/Transforms/SLPVectorizer/X86/max_intrinsic_cost.ll
===================================================================
--- /dev/null
+++ llvm/test/Transforms/SLPVectorizer/X86/max_intrinsic_cost.ll
@@ -0,0 +1,24 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt < %s -slp-vectorizer -mcpu=corei7-avx -mtriple=x86_64-unknown-linux -S | FileCheck %s
+
+; This test checks whether the cost of the vector max intrinsic is calculated
+; correctly. A max vector intrinsic combines the select and icmp instructions.
+; This maps to a single PMAX instruction in x86.
+define void @max_intrinsic_cost(i64 %arg0, i64 %arg1) {
+; CHECK-LABEL: @max_cost(
+; CHECK-NEXT:    [[TMP1:%.*]] = insertelement <2 x i64> poison, i64 [[ARG0:%.*]], i32 0
+; CHECK-NEXT:    [[TMP2:%.*]] = insertelement <2 x i64> [[TMP1]], i64 [[ARG1:%.*]], i32 1
+; CHECK-NEXT:    [[TMP3:%.*]] = icmp sgt <2 x i64> [[TMP2]], <i64 123, i64 456>
+; CHECK-NEXT:    [[TMP4:%.*]] = select <2 x i1> [[TMP3]], <2 x i64> [[TMP2]], <2 x i64> <i64 123, i64 456>
+; CHECK-NEXT:    [[TMP5:%.*]] = extractelement <2 x i64> [[TMP4]], i32 0
+; CHECK-NEXT:    [[TMP6:%.*]] = extractelement <2 x i64> [[TMP4]], i32 1
+; CHECK-NEXT:    [[ROOT:%.*]] = icmp sle i64 [[TMP5]], [[TMP6]]
+; CHECK-NEXT:    ret void
+;
+  %icmp0 = icmp sgt i64 %arg0, 123
+  %icmp1 = icmp sgt i64 %arg1, 456
+  %select0 = select i1 %icmp0, i64 %arg0, i64 123
+  %select1 = select i1 %icmp1, i64 %arg1, i64 456
+  %root = icmp sle i64 %select0, %select1
+  ret void
+}


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D120438.410956.patch
Type: text/x-patch
Size: 1544 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20220223/dc9f6062/attachment.bin>