6502/llvm-6502
1
0
Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2024-10-02 17:55:18 +00:00
Code Issues Projects Releases Wiki Activity

LoopVectorizer: Only allow vectorization of intrinsics. We can't know for sure that the functions 'abs' or 'round' are the functions from libm.

Browse Source 
rdar://15012650



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@191122 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Nadav Rotem 2013-09-21 00:27:05 +00:00
parent 74d3482f76
commit 2ecd8c90b0
2 changed files with 39 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
lib/Transforms/Vectorize/LoopVectorize.cpp
View File

@ -2925,9 +2925,18 @@ bool LoopVectorizationLegality::canVectorizeInstrs() {
// We still don't handle functions. However, we can ignore dbg intrinsic
// calls and we do handle certain intrinsic and libm functions.
CallInst *CI = dyn_cast<CallInst>(it);
if (CI && !getIntrinsicIDForCall(CI, TLI) && !isa<DbgInfoIntrinsic>(CI)) {
if (CI) {
DEBUG(dbgs() << "LV: Found a call site.\n");
return false;
if (!isa<IntrinsicInst>(it)) {
DEBUG(dbgs() << "LV: We only vectorize intrinsics.\n");
return false;
}
if (!getIntrinsicIDForCall(CI, TLI) && !isa<DbgInfoIntrinsic>(CI)) {
DEBUG(dbgs() << "LV: Found an unknown intrinsic.\n");
return false;
}
}
// Check that the instruction return type is vectorizable.

31
test/Transforms/LoopVectorize/intrinsic.ll
View File

@ -1018,7 +1018,7 @@ for.body: ; preds = %entry, %for.body
%indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
%arrayidx = getelementptr inbounds float* %x, i64 %indvars.iv
%0 = load float* %arrayidx, align 4
%call = tail call float @fabsf(float %0) nounwind readnone
%call = tail call float @llvm.fabs.f32(float %0) nounwind readnone
store float %call, float* %arrayidx, align 4
%indvars.iv.next = add i64 %indvars.iv, 1
%lftr.wideiv = trunc i64 %indvars.iv.next to i32
@ -1029,6 +1029,31 @@ for.end: ; preds = %for.body
ret void
}
declare float @fabsf(float) nounwind readnone
declare double @llvm.pow.f64(double, double) nounwind readnone
;CHECK: @not_intrin
;CHECK: @round
;CHECK-NOT: @round
;CHECK: ret
define void @not_intrin(i32* nocapture %A) nounwind ssp uwtable {
br label %1
; <label>:1 ; preds = %1, %0
%indvars.iv = phi i64 [ 0, %0 ], [ %indvars.iv.next, %1 ]
%2 = getelementptr inbounds i32* %A, i64 %indvars.iv
%3 = load i32* %2, align 4
%4 = add nsw i32 %3, 3
store i32 %4, i32* %2, align 4
%5 = trunc i64 %indvars.iv to i32
tail call void @round(i32 %5) nounwind
%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 %6, label %1
; <label>:6 ; preds = %1
ret void
}
declare void @round(i32)