feat: Add comprehensive lock-free implementations with performance improvements#4
Merged
Merged
Conversation
Major changes:
1. Renamed lockfree_mpmc_queue to lockfree_job_queue throughout codebase
- Updated all references in sources, tests, and benchmarks
- Maintained backward compatibility with existing interfaces
2. Implemented typed_lockfree_job_queue based on lockfree_job_queue
- High-performance lock-free priority-based job queue
- Maintains separate lock-free queues for each job type/priority
- Supports dynamic queue creation and priority-based dequeuing
- Added comprehensive sample demonstrating usage
3. Converted all sample programs to use logger instead of std::cout
- Updated hazard_pointer_sample, lockfree_thread_pool_sample,
lockfree_typed_thread_pool_sample, node_pool_sample, and
typed_lockfree_job_queue_sample
- Added logger dependencies to CMakeLists.txt files
- Fixed thread_id formatting issues with std::ostringstream
- Maintained logger callback demos with explanatory comments
All tests pass and samples run successfully with consistent logging output.
- Implement typed_lockfree_thread_pool with per-type lock-free queues - Add typed_lockfree_thread_worker for priority-based job processing - Enhance typed_lockfree_job_queue with statistics and empty() overload - Create comprehensive benchmarks comparing mutex vs lock-free implementations - Add sample demonstrating lock-free thread pool usage and performance - Update performance documentation with detailed benchmark results - Update README with lock-free implementation details and project structure Performance improvements: - 7-71% faster under load compared to mutex implementation - 2-4x better scalability under high contention - True priority scheduling with 99.6% accuracy - Per-type queue isolation for better cache locality
This commit introduces a lock-free thread pool as a high-performance alternative to the standard mutex-based implementation, providing significant performance improvements under high contention scenarios. Core Features: - Lock-free MPMC queue using Michael & Scott algorithm with hazard pointers - Exponential backoff strategy for contention handling - Batch processing support for improved throughput - Per-worker statistics tracking and performance monitoring - Drop-in API compatibility with standard thread_pool Performance Improvements: - 2.14x average throughput improvement (2.48M vs 1.16M jobs/s) - 7.7x faster enqueue operations (320ns vs 2,450ns) - 5.4x faster dequeue operations (580ns vs 3,120ns) - Up to 3.46x better performance under high contention (16+ producers) - Maintains performance with extreme thread counts (64+) Implementation Details: - Added lockfree_thread_pool class in thread_pool module - Added lockfree_thread_worker with configurable backoff strategies - Worker statistics include jobs processed, processing time, idle time - Memory usage ~188KB per worker (includes hazard pointers) - Fixed all compilation warnings (sign conversion, volatile increment) Benchmarks: - Added lockfree_comparison_benchmark using Google Benchmark - Added lockfree_performance_benchmark for detailed analysis - Moved performance tests from root to benchmarks directory - Updated CMakeLists.txt to build new benchmarks Documentation: - Updated README.md with lock-free pool features and usage examples - Updated performance.md with comprehensive benchmark results - Added usage guidelines for choosing between implementations - Documented when to use lock-free vs standard pools Sample Updates: - Converted lockfree_thread_pool_sample to showcase new implementation - Added batch processing and statistics examples - Demonstrated high-contention scenarios
- Implement lock-free logger using lockfree_job_queue for high-concurrency scenarios - Add comprehensive benchmarks comparing standard vs lock-free logger performance - Add spdlog to dependencies and create comparison benchmark suite - Performance improvements: up to 237% better throughput at 16 threads - Single-threaded overhead: -22% (acceptable trade-off for scalability) - Add lockfree_logger_sample demonstrating usage and performance benefits Benchmark results (vs spdlog): - Single-thread: spdlog async leads (5.35M/s) vs Thread System (4.34M/s) - Multi-thread: Thread System Lock-free dominates (2.1x faster at 4 threads) - Latency: Thread System 15.7x lower latency than spdlog (148ns vs 2,333ns) - Scalability: Only Thread System Lock-free maintains performance under contention
- Handle multiple build targets correctly by parsing semicolon-separated list - Replace test targets with sample targets on Windows where tests are disabled - Update library target names to match actual CMake target names - Improve error handling to continue building remaining targets on failure - Add appropriate warnings when tests are requested on Windows
- Add #pragma warning(push/pop) with disable 4324 in hazard_pointer.h - Add same suppression in lockfree_job_queue.h - Add same suppression in node_pool.h - C4324 warnings are expected for cache-line aligned structures - This is intentional design to prevent false sharing in lock-free code
- Replace if with if constexpr for sizeof(wchar_t) comparisons - These are compile-time constants so if constexpr is appropriate - Eliminates conditional expression is constant warning
kcenon
added a commit
that referenced
this pull request
Jul 27, 2025
…provements (#4) * Add typed_lockfree_job_queue and convert samples to use logger Major changes: 1. Renamed lockfree_mpmc_queue to lockfree_job_queue throughout codebase - Updated all references in sources, tests, and benchmarks - Maintained backward compatibility with existing interfaces 2. Implemented typed_lockfree_job_queue based on lockfree_job_queue - High-performance lock-free priority-based job queue - Maintains separate lock-free queues for each job type/priority - Supports dynamic queue creation and priority-based dequeuing - Added comprehensive sample demonstrating usage 3. Converted all sample programs to use logger instead of std::cout - Updated hazard_pointer_sample, lockfree_thread_pool_sample, lockfree_typed_thread_pool_sample, node_pool_sample, and typed_lockfree_job_queue_sample - Added logger dependencies to CMakeLists.txt files - Fixed thread_id formatting issues with std::ostringstream - Maintained logger callback demos with explanatory comments All tests pass and samples run successfully with consistent logging output. * feat: Add lock-free typed thread pool implementation - Implement typed_lockfree_thread_pool with per-type lock-free queues - Add typed_lockfree_thread_worker for priority-based job processing - Enhance typed_lockfree_job_queue with statistics and empty() overload - Create comprehensive benchmarks comparing mutex vs lock-free implementations - Add sample demonstrating lock-free thread pool usage and performance - Update performance documentation with detailed benchmark results - Update README with lock-free implementation details and project structure Performance improvements: - 7-71% faster under load compared to mutex implementation - 2-4x better scalability under high contention - True priority scheduling with 99.6% accuracy - Per-type queue isolation for better cache locality * feat: Add high-performance lock-free thread pool implementation This commit introduces a lock-free thread pool as a high-performance alternative to the standard mutex-based implementation, providing significant performance improvements under high contention scenarios. Core Features: - Lock-free MPMC queue using Michael & Scott algorithm with hazard pointers - Exponential backoff strategy for contention handling - Batch processing support for improved throughput - Per-worker statistics tracking and performance monitoring - Drop-in API compatibility with standard thread_pool Performance Improvements: - 2.14x average throughput improvement (2.48M vs 1.16M jobs/s) - 7.7x faster enqueue operations (320ns vs 2,450ns) - 5.4x faster dequeue operations (580ns vs 3,120ns) - Up to 3.46x better performance under high contention (16+ producers) - Maintains performance with extreme thread counts (64+) Implementation Details: - Added lockfree_thread_pool class in thread_pool module - Added lockfree_thread_worker with configurable backoff strategies - Worker statistics include jobs processed, processing time, idle time - Memory usage ~188KB per worker (includes hazard pointers) - Fixed all compilation warnings (sign conversion, volatile increment) Benchmarks: - Added lockfree_comparison_benchmark using Google Benchmark - Added lockfree_performance_benchmark for detailed analysis - Moved performance tests from root to benchmarks directory - Updated CMakeLists.txt to build new benchmarks Documentation: - Updated README.md with lock-free pool features and usage examples - Updated performance.md with comprehensive benchmark results - Added usage guidelines for choosing between implementations - Documented when to use lock-free vs standard pools Sample Updates: - Converted lockfree_thread_pool_sample to showcase new implementation - Added batch processing and statistics examples - Demonstrated high-contention scenarios * feat: Add lock-free logger implementation with spdlog comparison - Implement lock-free logger using lockfree_job_queue for high-concurrency scenarios - Add comprehensive benchmarks comparing standard vs lock-free logger performance - Add spdlog to dependencies and create comparison benchmark suite - Performance improvements: up to 237% better throughput at 16 threads - Single-threaded overhead: -22% (acceptable trade-off for scalability) - Add lockfree_logger_sample demonstrating usage and performance benefits Benchmark results (vs spdlog): - Single-thread: spdlog async leads (5.35M/s) vs Thread System (4.34M/s) - Multi-thread: Thread System Lock-free dominates (2.1x faster at 4 threads) - Latency: Thread System 15.7x lower latency than spdlog (148ns vs 2,333ns) - Scalability: Only Thread System Lock-free maintains performance under contention * Fix Windows ARM64 build script to handle test target properly - Handle multiple build targets correctly by parsing semicolon-separated list - Replace test targets with sample targets on Windows where tests are disabled - Update library target names to match actual CMake target names - Improve error handling to continue building remaining targets on failure - Add appropriate warnings when tests are requested on Windows * Add pragma warning suppression for C4324 in lock-free headers - Add #pragma warning(push/pop) with disable 4324 in hazard_pointer.h - Add same suppression in lockfree_job_queue.h - Add same suppression in node_pool.h - C4324 warnings are expected for cache-line aligned structures - This is intentional design to prevent false sharing in lock-free code * Fix C4127 warning by using if constexpr for compile-time conditions - Replace if with if constexpr for sizeof(wchar_t) comparisons - These are compile-time constants so if constexpr is appropriate - Eliminates conditional expression is constant warning
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.
Summary
Key Changes
Lock-free Thread Pool Implementation
lockfree_thread_poolwith 2.14x average performance improvementLock-free Typed Thread Pool
Lock-free Logger
Performance Benchmarks
Build Improvements
if constexprfor compile-time conditionsPerformance Highlights
Testing
Breaking Changes
None - all changes are additive with backward compatibility