Optimize StrictMetricsEvaluator reference binding and add benchmarks#823
Open
SURYAS1306 wants to merge 2 commits into
Open
Optimize StrictMetricsEvaluator reference binding and add benchmarks#823SURYAS1306 wants to merge 2 commits into
SURYAS1306 wants to merge 2 commits into
Conversation
Reuse dynamic_cast results in predicate handlers to avoid redundant shared_ptr refcount bumps, and add an optional Google Benchmark target behind ICEBERG_BUILD_BENCHMARKS for the filtering hot path (issue apache#690). Co-authored-by: Cursor <cursoragent@cursor.com>
Apply clang-format and cmake-format, move benchmarks under src/iceberg/benchmark, and enable ICEBERG_BUILD_BENCHMARKS in the cpp-linter workflow so Google Benchmark headers are available. Co-authored-by: Cursor <cursoragent@cursor.com>
There was a problem hiding this comment.
Pull request overview
This PR optimizes StrictMetricsEvaluator’s predicate reference binding by reusing the dynamic_cast result (avoiding repeated shared_from_this() refcount bumps), and adds an off-by-default benchmark build option plus an initial Google Benchmark target to measure the filtering hot path.
Changes:
- Replace the
RETURN_IF_NOT_REFERENCEpattern withBIND_REFERENCE_OR_RETURNso each predicate handler reuses the already-castedBoundReference*. - Add
ICEBERG_BUILD_BENCHMARKSCMake option and asrc/iceberg/benchmark/subdirectory that fetches Google Benchmark and buildsstrict_metrics_evaluator_bench. - Enable
ICEBERG_BUILD_BENCHMARKS=ONin thecpp-linterworkflow build step to ensure the benchmark target stays buildable in CI.
Reviewed changes
Copilot reviewed 6 out of 6 changed files in this pull request and generated 2 comments.
Show a summary per file
| File | Description |
|---|---|
src/iceberg/expression/strict_metrics_evaluator.cc |
Hot-path optimization: reuse dynamic_cast result instead of calling reference() repeatedly. |
CMakeLists.txt |
Add ICEBERG_BUILD_BENCHMARKS option (default OFF). |
src/iceberg/CMakeLists.txt |
Wire benchmark subdir behind ICEBERG_BUILD_BENCHMARKS. |
src/iceberg/benchmark/CMakeLists.txt |
Fetch Google Benchmark and define the benchmark executable target. |
src/iceberg/benchmark/strict_metrics_evaluator_bench.cc |
Add initial micro-benchmark for StrictMetricsEvaluator::Evaluate. |
.github/workflows/cpp-linter.yml |
Turn on benchmarks during CI build to catch build regressions. |
Comment on lines
+119
to
+120
| BENCHMARK(iceberg::BM_StrictMetricsEvaluate)->Arg(1)->Arg(14)->Arg(20); | ||
| BENCHMARK_MAIN(); |
Comment on lines
+65
to
+66
| std::shared_ptr<Expression> MakeFilter(int predicate_count) { | ||
| std::vector<std::shared_ptr<Expression>> predicates; |
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
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.
Summary
This PR implements the filtering hot-path optimization and benchmark infrastructure proposed in #690.
RETURN_IF_NOT_REFERENCEwithBIND_REFERENCE_OR_RETURNinStrictMetricsEvaluatorso predicate handlers reuse thedynamic_castresult instead of callingexpr->reference()(which bumpsshared_ptrrefcounts viashared_from_this()).ICEBERG_BUILD_BENCHMARKSCMake option that fetches Google Benchmark (v1.9.1) and buildsstrict_metrics_evaluator_bench.Motivation
StrictMetricsEvaluatorruns once per data file during scan planning. The previous macro discarded a successfuldynamic_castand then re-fetched the same reference through a virtual call + atomic refcount operation on every predicate handler (14+ predicate types).Benchmark results (Release, Apple Clang 17, macOS)
Micro-benchmark model of the predicate binding hot path:
End-to-end
StrictMetricsEvaluator::Evaluatebenchmark (with binding + evaluation):Test plan
expression_test --gtest_filter='StrictMetrics*'(42 tests pass)-DICEBERG_BUILD_BENCHMARKS=ON./benchmark/strict_metrics_evaluator_benchCloses #690