fix(compression): bring in Felipe's fixes and improvements to compression exploration - #1726
Open
joosthooz wants to merge 1 commit into
Open
fix(compression): bring in Felipe's fixes and improvements to compression exploration#1726joosthooz wants to merge 1 commit into
joosthooz wants to merge 1 commit into
Conversation
…or Pareto pick The rerank pass's round-trip timer let an NVRTC cold compile land inside the timed bracket on a JIT-cache miss, and a bare one-shot wall clock carried launch/sync + first-touch overhead — reported rates were 6.7-12x off. Replace it with a warmup phase (absorbs the compile, default 2 passes) followed by `rerank_iters` timed, event-bracketed, median-of-N measurements, configurable via new `rerank_warmup`/`rerank_iters` exploration_config fields and `--rerank-warmup`/`--rerank-iters` CLI flags. The old one-shot metric survives as `old_wall_compress_gbps`/`old_wall_decompress_gbps` so existing plan files stay comparable. Also add a Pareto decompress-throughput floor (`pareto_decomp_floor_gbps` / `--decomp-floor`): pick max-ratio among frontier points at or above the floor, falling back to the fastest-decompress point if none qualify. The full measured frontier is now exposed on `exploration_result` and can be dumped via `--frontier-out` as TSV, so a floor re-pick doesn't require re-measuring. Extracted from Felipe's sirius-db#1391 (exp/fused-scan-filter), where these explorer fixes were bundled with unrelated dict-encode and pin-logging changes that belong in their own PR. Co-Authored-By: Felipe Aramburu <faramburu@nvidia.com> Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
joosthooz
force-pushed
the
fix/simpatico-explorer-rerank-timing
branch
from
September 7, 2026 12:13
27a218f to
27e3867
Compare
joosthooz
marked this pull request as ready for review
September 7, 2026 19:51
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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.
Description
This PR brings in the changes from Felipe's late-mat branch #1391 to fix and improve simpatico's explore function.
It includes excluding JIT costs from the evaluation, and outputting all pareto-optimal points so we can change the throughput floor without re-measuring.
Checklist
References