6502/llvm-6502
1
0
Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2025-01-15 07:34:33 +00:00
Code Issues Projects Releases Wiki Activity

Allow loop-idiom to run on multiple BB loops, but still only scan the loop

Browse Source 
header for now for memset/memcpy opportunities.  It turns out that loop-rotate
is successfully rotating loops, but *DOESN'T MERGE THE BLOCKS*, turning "for 
loops" into 2 basic block loops that loop-idiom was ignoring.

With this fix, we form many *many* more memcpy and memsets than before, including
on the "history" loops in the viterbi benchmark, which look like this:

        for (j=0; j<MAX_history; ++j) {
          history_new[i][j+1] = history[2*i][j];
        }

Transforming these loops into memcpy's speeds up the viterbi benchmark from
11.98s to 3.55s on my machine.  Woo.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@122685 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Chris Lattner 2011-01-02 07:58:36 +00:00
parent c9e152b778
commit cf078f2b20
3 changed files with 29 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
lib/Target/README.txt
View File

@ -414,14 +414,6 @@ this construct.
//===---------------------------------------------------------------------===//
[LOOP RECOGNITION]
viterbi speeds up *significantly* if the various "history" related copy loops
are turned into memcpy calls at the source level. We need a "loops to memcpy"
pass.
//===---------------------------------------------------------------------===//
[LOOP OPTIMIZATION]
SingleSource/Benchmarks/Misc/dt.c shows several interesting optimization

10
lib/Transforms/Scalar/LoopIdiomRecognize.cpp
View File

@ -126,11 +126,6 @@ static void DeleteDeadInstruction(Instruction *I, ScalarEvolution &SE) {
bool LoopIdiomRecognize::runOnLoop(Loop *L, LPPassManager &LPM) {
CurLoop = L;
// We only look at trivial single basic block loops.
// TODO: eventually support more complex loops, scanning the header.
if (L->getBlocks().size() != 1)
return false;
// The trip count of the loop must be analyzable.
SE = &getAnalysis<ScalarEvolution>();
if (!SE->hasLoopInvariantBackedgeTakenCount(L))
@ -142,6 +137,11 @@ bool LoopIdiomRecognize::runOnLoop(Loop *L, LPPassManager &LPM) {
TD = getAnalysisIfAvailable<TargetData>();
if (TD == 0) return false;
// TODO: We currently only scan the header of the loop, because it is the only
// part that is known to execute and we don't want to make a conditional store
// into an unconditional one in the preheader. However, there can be diamonds
// and other things in the loop that would make other blocks "always executed"
// we should get the full set and scan each block.
BasicBlock *BB = L->getHeader();
DEBUG(dbgs() << "loop-idiom Scanning: F[" << BB->getParent()->getName()
<< "] Loop %" << BB->getName() << "\n");

24
test/Transforms/LoopIdiom/basic.ll
View File

@ -21,6 +21,30 @@ for.end: ; preds = %for.body, %entry
; CHECK-NOT: store
}
; This is a loop that was rotated but where the blocks weren't merged. This
; shouldn't perturb us.
define void @test1a(i8* %Base, i64 %Size) nounwind ssp {
bb.nph: ; preds = %entry
br label %for.body
for.body: ; preds = %bb.nph, %for.body
%indvar = phi i64 [ 0, %bb.nph ], [ %indvar.next, %for.body.cont ]
%I.0.014 = getelementptr i8* %Base, i64 %indvar
store i8 0, i8* %I.0.014, align 1
%indvar.next = add i64 %indvar, 1
br label %for.body.cont
for.body.cont:
%exitcond = icmp eq i64 %indvar.next, %Size
br i1 %exitcond, label %for.end, label %for.body
for.end: ; preds = %for.body, %entry
ret void
; CHECK: @test1a
; CHECK: call void @llvm.memset.p0i8.i64(i8* %Base, i8 0, i64 %Size, i32 1, i1 false)
; CHECK-NOT: store
}
define void @test2(i32* %Base, i64 %Size) nounwind ssp {
entry:
%cmp10 = icmp eq i64 %Size, 0