[CombToSynth] Compute Kogge-Stone prefix tree lazily in unsigned comparison lowering #9050
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
uenoku
commented
Oct 4, 2025
uenoku
force-pushed
the
dev/hidetou/lazy-Kogge-Stone
branch
from
6c9ce16 to
ec10b65
Compare
…arison lowering Add LazyKoggeStonePrefixTree class for on-demand computation of prefix values, use lazy evaluation in ICmpOp conversion to avoid computing all intermediate prefix values.
uenoku
force-pushed
the
dev/hidetou/lazy-Kogge-Stone
branch
from
ec10b65 to
a7acda3
Compare
cowardsa
approved these changes
Oct 7, 2025
| if (it != prefixCache.end()) | ||
| return it->second; | ||
|
|
||
| assert(level > 0 && "level must be positive"); |
There was a problem hiding this comment.
Nit: perhaps a comment to justify that we're now moving to compute the values and we can only do that after level 0 would help - took me a minute to get that
There was a problem hiding this comment.
Absolutely, sorry for the confusion.
Yeah Brent-Kung would be really non-trivial and not sure if there is a clean way :) I'll probably prioritize Sklanskey as it's used in 32-bit comparison (which creates ~500 dead operations right now). |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Add LazyKoggeStonePrefixTree class for on-demand computation of prefix values, use lazy evaluation in ICmpOp conversion to avoid computing all intermediate prefix values. The class could be generalized to handle other prefix tree as well. KoggeStone is prioritized for now since it's currently used for wide comparisons (actually for smaller width it's not too bad to just rely on DCE). Locally checked LEC for larger width like i128 etc.