| Age | Commit message (Collapse) | Author |
|
- Rename PBO (Profile-Based) to PGO (Performance Guided Optimization)
- Implement true PGO using Go's -pgo compiler flag
- Refactor script into maintainable functions:
- setup_environment(): Initialize paths and variables
- create_test_file(): Generate 100MB test file with 1M lines
- build_baseline(): Build version without PGO optimizations
- collect_training_data(): Generate CPU profiles for training
- build_pgo_optimized(): Build with -pgo flag using training profile
- run_pgo_performance_test(): Profile PGO-optimized version
- run_performance_comparison(): Compare baseline vs PGO performance
- generate_detailed_analysis(): Create comprehensive profile analysis
- cleanup(): Remove temporary files
- show_summary(): Display results and process summary
- Update Makefile target from 'pbo' to 'pgo'
- Update .gitignore patterns for PGO temporary files
- Update CLAUDE.md documentation for new PGO process
- Remove git stash dependencies for simpler automation
- Generate before/after performance comparison reports
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
|
|
improvement
- Add comprehensive PBO script (scripts/pbo.sh) for automated performance analysis
- Implement timer allocation reduction using reusable timers (chunkedreader.go, stats.go, baseclient.go)
- Optimize I/O operations with pre-allocated buffers and bulk writes (chunkedreader.go)
- Enhance memory allocation patterns with improved buffer pooling
- Add CPU and memory profiling support to dgrep command
- Update Makefile with clean PBO target calling scripts/pbo.sh
- Add PBO documentation to CLAUDE.md
Performance improvements:
- 39.9% faster execution time (2.918s → 1.753s average)
- 38% reduction in CPU samples (3.04s → 1.87s)
- Reduced byte-by-byte operations from 21.71% to 8.56% CPU usage
- Eliminated repeated timer allocations across all components
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
|
|
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
|
|
|
|
|
|
|
|
|
|
|
|
this is useful for running unit and integration tests in jenkins
|
|
|
|
|
|
|
|
|
|
|
|
|
|
file unless NONE is specified
|
|
logger package rewrite as dlog
|
|
|
|
|
|
|
|
|
|
less portable
|
|
|
|
|
|
|
|
|
|
|
|
dealing with log streams
|
|
useful for log file monitorig for example. breaking protocol change which allows to mapreduce aggreate messages containing the default field separator |. add of more unit tests. add logformat mapreduce query keyword. add set mapreduce clause support and support to evaluate built-in functions such as md5sum() and maskdigits().
|
|
|
|
|
|
|
|
|