[llvm] 0e0295f - [LV] Pass explicit vector width to not require a X86 target.

[Florian Hahn via llvm-commits]

Author: Florian Hahn
Date: 2020-12-15T12:52:22Z
New Revision: 0e0295fd6124163b478603a6cc357867aadf5d70

URL: https://github.com/llvm/llvm-project/commit/0e0295fd6124163b478603a6cc357867aadf5d70
DIFF: https://github.com/llvm/llvm-project/commit/0e0295fd6124163b478603a6cc357867aadf5d70.diff

LOG: [LV] Pass explicit vector width to not require a X86 target.

Added: 
    

Modified: 
    llvm/test/Transforms/LoopVectorize/select-reduction.ll

Removed: 
    


################################################################################
diff  --git a/llvm/test/Transforms/LoopVectorize/select-reduction.ll b/llvm/test/Transforms/LoopVectorize/select-reduction.ll
index 579cf5df2860..4c3ad8ee8aa9 100644
--- a/llvm/test/Transforms/LoopVectorize/select-reduction.ll
+++ b/llvm/test/Transforms/LoopVectorize/select-reduction.ll
@@ -1,8 +1,7 @@
 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
-; RUN: opt -loop-vectorize -S %s | FileCheck %s
+; RUN: opt -force-vector-width=4 -force-vector-interleave=1 -loop-vectorize -S %s | FileCheck %s
 
 target datalayout = "e-m:o-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
-target triple = "x86_64-apple-macosx10.15.0"
 
 define i32 @test(i64 %N, i32 %x) {
 ; CHECK-LABEL: @test(
@@ -42,7 +41,13 @@ define i32 @test(i64 %N, i32 %x) {
 ; CHECK-NEXT:    [[TMP5:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
 ; CHECK-NEXT:    br i1 [[TMP5]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP0:!llvm.loop !.*]]
 ; CHECK:       middle.block:
-; CHECK-NEXT:    [[TMP6:%.*]] = call i32 @llvm.vector.reduce.smax.v4i32(<4 x i32> [[TMP4]])
+; CHECK-NEXT:    [[RDX_SHUF:%.*]] = shufflevector <4 x i32> [[TMP4]], <4 x i32> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
+; CHECK-NEXT:    [[RDX_MINMAX_CMP:%.*]] = icmp sgt <4 x i32> [[TMP4]], [[RDX_SHUF]]
+; CHECK-NEXT:    [[RDX_MINMAX_SELECT:%.*]] = select <4 x i1> [[RDX_MINMAX_CMP]], <4 x i32> [[TMP4]], <4 x i32> [[RDX_SHUF]]
+; CHECK-NEXT:    [[RDX_SHUF5:%.*]] = shufflevector <4 x i32> [[RDX_MINMAX_SELECT]], <4 x i32> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT:    [[RDX_MINMAX_CMP6:%.*]] = icmp sgt <4 x i32> [[RDX_MINMAX_SELECT]], [[RDX_SHUF5]]
+; CHECK-NEXT:    [[RDX_MINMAX_SELECT7:%.*]] = select <4 x i1> [[RDX_MINMAX_CMP6]], <4 x i32> [[RDX_MINMAX_SELECT]], <4 x i32> [[RDX_SHUF5]]
+; CHECK-NEXT:    [[TMP6:%.*]] = extractelement <4 x i32> [[RDX_MINMAX_SELECT7]], i32 0
 ; CHECK-NEXT:    br i1 true, label [[EXIT_LOOPEXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[IND_END]], [[MIDDLE_BLOCK]] ], [ [[EXTRA_ITER]], [[LOOP_PREHEADER]] ]