maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity

1. Fix a bug in getTypeConversion. When a *simple* type is split, we need to return the type of the split result. 

2. Change the maximum vectorization width from 4 to 8.
3. A test for both.

llvm-svn: 166864
This commit is contained in:
Nadav Rotem 2012-10-27 04:11:32 +00:00
parent 8c0739d332
commit 859366f93f
3 changed files with 66 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
llvm/include/llvm/Target/TargetLowering.h
View File

@ -1980,6 +1980,9 @@ public:
ValueTypeActions.getTypeAction(NVT.getSimpleVT()) != TypePromoteInteger)
&& "Promote may not follow Expand or Promote");
if (LA == TypeSplitVector)
NVT = EVT::getVectorVT(Context, VT.getVectorElementType(),
VT.getVectorNumElements() / 2);
return LegalizeKind(LA, NVT);
}

2
llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
View File

@ -317,7 +317,7 @@ public:
/// Returns the most profitable vectorization factor for the loop that is
/// smaller or equal to the VF argument. This method checks every power
/// of two up to VF.
unsigned findBestVectorizationFactor(unsigned VF = 4);
unsigned findBestVectorizationFactor(unsigned VF = 8);
private:
/// Returns the expected execution cost. The unit of the cost does

62
llvm/test/Transforms/LoopVectorize/X86/gcc-examples.ll Normal file
View File

@ -0,0 +1,62 @@
; RUN: opt < %s -loop-vectorize -mtriple=x86_64-apple-macosx10.8.0 -mcpu=corei7 -dce -instcombine -licm -S | FileCheck %s
target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"
target triple = "x86_64-apple-macosx10.8.0"
@b = common global [2048 x i32] zeroinitializer, align 16
@c = common global [2048 x i32] zeroinitializer, align 16
@a = common global [2048 x i32] zeroinitializer, align 16
; Select VF = 8;
;CHECK: @example1
;CHECK: load <8 x i32>
;CHECK: add <8 x i32>
;CHECK: store <8 x i32>
;CHECK: ret void
define void @example1() nounwind uwtable ssp {
br label %1
; <label>:1 ; preds = %1, %0
%indvars.iv = phi i64 [ 0, %0 ], [ %indvars.iv.next, %1 ]
%2 = getelementptr inbounds [2048 x i32]* @b, i64 0, i64 %indvars.iv
%3 = load i32* %2, align 4
%4 = getelementptr inbounds [2048 x i32]* @c, i64 0, i64 %indvars.iv
%5 = load i32* %4, align 4
%6 = add nsw i32 %5, %3
%7 = getelementptr inbounds [2048 x i32]* @a, i64 0, i64 %indvars.iv
store i32 %6, i32* %7, align 4
%indvars.iv.next = add i64 %indvars.iv, 1
%lftr.wideiv = trunc i64 %indvars.iv.next to i32
%exitcond = icmp eq i32 %lftr.wideiv, 256
br i1 %exitcond, label %8, label %1
; <label>:8 ; preds = %1
ret void
}
; Select VF=4 because sext <8 x i1> to <8 x i32> is expensive.
;CHECK: @example10b
;CHECK: load <4 x i16>
;CHECK: sext <4 x i16>
;CHECK: store <4 x i32>
;CHECK: ret void
define void @example10b(i16* noalias nocapture %sa, i16* noalias nocapture %sb, i16* noalias nocapture %sc, i32* noalias nocapture %ia, i32* noalias nocapture %ib, i32* noalias nocapture %ic) nounwind uwtable ssp {
br label %1
; <label>:1 ; preds = %1, %0
%indvars.iv = phi i64 [ 0, %0 ], [ %indvars.iv.next, %1 ]
%2 = getelementptr inbounds i16* %sb, i64 %indvars.iv
%3 = load i16* %2, align 2
%4 = sext i16 %3 to i32
%5 = getelementptr inbounds i32* %ia, i64 %indvars.iv
store i32 %4, i32* %5, align 4
%indvars.iv.next = add i64 %indvars.iv, 1
%lftr.wideiv = trunc i64 %indvars.iv.next to i32
%exitcond = icmp eq i32 %lftr.wideiv, 1024
br i1 %exitcond, label %6, label %1
; <label>:6 ; preds = %1
ret void
}