Skip to content

[MachineOutliner] Efficient Implementation of MachineOutliner::findCandidates() #90260

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 5 commits into from
Jun 3, 2024
Merged

[MachineOutliner] Efficient Implementation of MachineOutliner::findCandidates() #90260

Show file tree
Hide file tree
Changes from 4 commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
046ddd4
efficient implementation of MachineOutliner::findCandidates()
xuanzhang816 Apr 26, 2024
de64e6f
update check on empty vector
xuanzhang816 May 3, 2024
358c954
retrigger checks
xuanzhang816 May 6, 2024
632f482
remove const and sort in place
xuanzhang816 May 28, 2024
95ddb95
modify comments
xuanzhang816 Jun 3, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
21 changes: 11 additions & 10 deletions llvm/lib/CodeGen/MachineOutliner.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -584,7 +584,7 @@ void MachineOutliner::findCandidates(
LLVM_DEBUG(dbgs() << "*** Discarding overlapping candidates *** \n");
LLVM_DEBUG(
dbgs() << "Searching for overlaps in all repeated sequences...\n");
for (const SuffixTree::RepeatedSubstring &RS : ST) {
for (SuffixTree::RepeatedSubstring &RS : ST) {
CandidatesForRepeatedSeq.clear();
unsigned StringLen = RS.Length;
LLVM_DEBUG(dbgs() << " Sequence length: " << StringLen << "\n");
Expand All @@ -593,6 +593,9 @@ void MachineOutliner::findCandidates(
unsigned NumDiscarded = 0;
unsigned NumKept = 0;
#endif
// Sort the start indices so that we can efficiently check if candidates
// overlap with each other in MachineOutliner::findCandidates().
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm a bit confused why mention this function's name here (MachineOutliner::findCandidates()). From the wording it would appear that it would referring to another function. Perhaps "overlap with each other further down" will be clearer.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for the comments. Modified!

llvm::sort(RS.StartIndices);
for (const unsigned &StartIdx : RS.StartIndices) {
// Trick: Discard some candidates that would be incompatible with the
// ones we've already found for this sequence. This will save us some
Expand All @@ -616,17 +619,15 @@ void MachineOutliner::findCandidates(
// * End before the other starts
// * Start after the other ends
unsigned EndIdx = StartIdx + StringLen - 1;
auto FirstOverlap = find_if(
CandidatesForRepeatedSeq, [StartIdx, EndIdx](const Candidate &C) {
return EndIdx >= C.getStartIdx() && StartIdx <= C.getEndIdx();
});
if (FirstOverlap != CandidatesForRepeatedSeq.end()) {
if (!CandidatesForRepeatedSeq.empty() &&
StartIdx <= CandidatesForRepeatedSeq.back().getEndIdx()) {
#ifndef NDEBUG
++NumDiscarded;
LLVM_DEBUG(dbgs() << " .. DISCARD candidate @ [" << StartIdx
<< ", " << EndIdx << "]; overlaps with candidate @ ["
<< FirstOverlap->getStartIdx() << ", "
<< FirstOverlap->getEndIdx() << "]\n");
LLVM_DEBUG(dbgs() << " .. DISCARD candidate @ [" << StartIdx << ", "
<< EndIdx << "]; overlaps with candidate @ ["
<< CandidatesForRepeatedSeq.back().getStartIdx()
<< ", " << CandidatesForRepeatedSeq.back().getEndIdx()
<< "]\n");
#endif
continue;
}
Expand Down
12 changes: 6 additions & 6 deletions llvm/test/CodeGen/AArch64/machine-outliner-overlap.mir
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,27 +8,27 @@
# CHECK-NEXT: Candidates discarded: 0
# CHECK-NEXT: Candidates kept: 2
# CHECK-DAG: Sequence length: 8
# CHECK-NEXT: .. DISCARD candidate @ [5, 12]; overlaps with candidate @ [12, 19]
# CHECK-NEXT: .. DISCARD candidate @ [12, 19]; overlaps with candidate @ [5, 12]
# CHECK-NEXT: Candidates discarded: 1
# CHECK-NEXT: Candidates kept: 1
# CHECK-DAG: Sequence length: 9
# CHECK-NEXT: .. DISCARD candidate @ [4, 12]; overlaps with candidate @ [11, 19]
# CHECK-NEXT: .. DISCARD candidate @ [11, 19]; overlaps with candidate @ [4, 12]
# CHECK-NEXT: Candidates discarded: 1
# CHECK-NEXT: Candidates kept: 1
# CHECK-DAG: Sequence length: 10
# CHECK-NEXT: .. DISCARD candidate @ [3, 12]; overlaps with candidate @ [10, 19]
# CHECK-NEXT: .. DISCARD candidate @ [10, 19]; overlaps with candidate @ [3, 12]
# CHECK-NEXT: Candidates discarded: 1
# CHECK-NEXT: Candidates kept: 1
# CHECK-DAG: Sequence length: 11
# CHECK-NEXT: .. DISCARD candidate @ [2, 12]; overlaps with candidate @ [9, 19]
# CHECK-NEXT: .. DISCARD candidate @ [9, 19]; overlaps with candidate @ [2, 12]
# CHECK-NEXT: Candidates discarded: 1
# CHECK-NEXT: Candidates kept: 1
# CHECK-DAG: Sequence length: 12
# CHECK-NEXT: .. DISCARD candidate @ [1, 12]; overlaps with candidate @ [8, 19]
# CHECK-NEXT: .. DISCARD candidate @ [8, 19]; overlaps with candidate @ [1, 12]
# CHECK-NEXT: Candidates discarded: 1
# CHECK-NEXT: Candidates kept: 1
# CHECK-DAG: Sequence length: 13
# CHECK-NEXT: .. DISCARD candidate @ [0, 12]; overlaps with candidate @ [7, 19]
# CHECK-NEXT: .. DISCARD candidate @ [7, 19]; overlaps with candidate @ [0, 12]
# CHECK-NEXT: Candidates discarded: 1
# CHECK-NEXT: Candidates kept: 1

Expand Down
Loading