Implementation of a multithreaded system to efficiently compute trees of boolean formulas, leveraging parallel processing to optimize evaluation performance.
Implementation of an optimized multithreaded application for processing large volumes of computational tasks. The solution employs a batch-checking strategy where threads access the shared task queue only every M iterations and in almost lock-free way, significantly reducing synchronization overhead and improving throughput.
The application is configurable with command-line arguments for thread count and problem parameters.
Performance results are available on Marcin Wrochna's page, showing comparative analysis of throughput and efficiency metrics.
Performance benchmarks demonstrated execution speeds surpassing reference implementations, including those developed by course instructors. That's what I'm proud of.
Explore mutexes and blocking queues. Understand common pitfalls such as deadlocks and thread starvation, and learn strategies to avoid blocking your code.
Learn abstract coding techniques to prevent your applications from becoming unresponsive, focusing on best practices for both synchronous and asynchronous communication between hosts.