6502/llvm-6502
1
0
Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2025-01-13 09:33:50 +00:00
Code Issues Projects Releases Wiki Activity

LoopVectorizer: Preserve fast-math flags

Browse Source 
Fixes PR19045.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@203008 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Arnold Schwaighofer 2014-03-05 21:10:47 +00:00
parent fc5436c951
commit 9d84b4d70c
3 changed files with 48 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

26
lib/Transforms/Vectorize/LoopVectorize.cpp
View File

@ -2489,6 +2489,16 @@ static void cse(SmallVector<BasicBlock *, 4> &BBs) {
}
}
/// \brief Adds a 'fast' flag to floating point operations.
static Value *addFastMathFlag(Value *V) {
if (isa<FPMathOperator>(V)){
FastMathFlags Flags;
Flags.setUnsafeAlgebra();
cast<Instruction>(V)->setFastMathFlags(Flags);
}
return V;
}
void InnerLoopVectorizer::vectorizeLoop() {
//===------------------------------------------------===//
//
@ -2632,9 +2642,10 @@ void InnerLoopVectorizer::vectorizeLoop() {
setDebugLocFromInst(Builder, ReducedPartRdx);
for (unsigned part = 1; part < UF; ++part) {
if (Op != Instruction::ICmp && Op != Instruction::FCmp)
ReducedPartRdx = Builder.CreateBinOp((Instruction::BinaryOps)Op,
RdxParts[part], ReducedPartRdx,
"bin.rdx");
// Floating point operations had to be 'fast' to enable the reduction.
ReducedPartRdx = addFastMathFlag(
Builder.CreateBinOp((Instruction::BinaryOps)Op, RdxParts[part],
ReducedPartRdx, "bin.rdx"));
else
ReducedPartRdx = createMinMaxOp(Builder, RdxDesc.MinMaxKind,
ReducedPartRdx, RdxParts[part]);
@ -2664,8 +2675,9 @@ void InnerLoopVectorizer::vectorizeLoop() {
"rdx.shuf");
if (Op != Instruction::ICmp && Op != Instruction::FCmp)
TmpVec = Builder.CreateBinOp((Instruction::BinaryOps)Op, TmpVec, Shuf,
"bin.rdx");
// Floating point operations had to be 'fast' to enable the reduction.
TmpVec = addFastMathFlag(Builder.CreateBinOp(
(Instruction::BinaryOps)Op, TmpVec, Shuf, "bin.rdx"));
else
TmpVec = createMinMaxOp(Builder, RdxDesc.MinMaxKind, TmpVec, Shuf);
}
@ -2999,6 +3011,10 @@ void InnerLoopVectorizer::vectorizeBlockInLoop(BasicBlock *BB, PhiVector *PV) {
if (VecOp && isa<PossiblyExactOperator>(VecOp))
VecOp->setIsExact(BinOp->isExact());
// Copy the fast-math flags.
if (VecOp && isa<FPMathOperator>(V))
VecOp->setFastMathFlags(it->getFastMathFlags());
Entry[Part] = V;
}
break;

26
test/Transforms/LoopVectorize/flags.ll
View File

@ -51,3 +51,29 @@ define i32 @flags2(i32 %n, i32* nocapture %A) nounwind uwtable ssp {
._crit_edge: ; preds = %.lr.ph, %0
ret i32 undef
}
; Make sure we copy fast math flags and use them for the final reduction.
; CHECK-LABEL: fast_math
; CHECK: load <4 x float>
; CHECK: fadd fast <4 x float>
; CHECK: br
; CHECK: fadd fast <4 x float>
; CHECK: fadd fast <4 x float>
define float @fast_math(float* noalias %s) {
entry:
br label %for.body
for.body:
%indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
%q.04 = phi float [ 0.000000e+00, %entry ], [ %add, %for.body ]
%arrayidx = getelementptr inbounds float* %s, i64 %indvars.iv
%0 = load float* %arrayidx, align 4
%add = fadd fast float %q.04, %0
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%exitcond = icmp eq i64 %indvars.iv.next, 256
br i1 %exitcond, label %for.end, label %for.body
for.end:
%add.lcssa = phi float [ %add, %for.body ]
ret float %add.lcssa
}

2
test/Transforms/LoopVectorize/float-reduction.ll
View File

@ -3,7 +3,7 @@
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"
;CHECK-LABEL: @foo(
;CHECK: fadd <4 x float>
;CHECK: fadd fast <4 x float>
;CHECK: ret
define float @foo(float* nocapture %A, i32* nocapture %n) nounwind uwtable readonly ssp {
entry: