Member of the Scalable Systems and Software Research Group at Lehigh University
Advised by Professor Roberto Palmieri
Research Interests
Concurrent data structures
Multi-threaded and parallel programming
Algorithm Design
NUMA Aware Designs
Publications
Olivia Grimes, Jacob Nelson-Slivon, Ahmed Hassan, Roberto Palmieri. Opportunities and Limitations of Hardware Timestamps in Concurrent Data Structures. 37th IEEE International Parallel & Distributed Processing Symposium, IPDPS, May 15-19, 2023, St. Petersburg, Florida USA.
Ongoing Work
Insert Optimized Priority Queue (Submitted to SPAA ‘25)
Designed and implemented a strict (i.e., not relaxed), linearizable, concurrent priority queue optimized for the insert operation that outperforms other strict competitors in a variety of cases, especially for insert heavy workloads. The design employs a hierarchical technique which follows the non-uniform memory access pattern of the underlying machine to optimize insertions by minimizing sharing and synchronization, and leveraging thread locality, without losing performance of the delete-min operation as compared to competitors.
Relaxed Priority Queue (Current Work)
Leveraging lessons learned from development of Insert Optimized Priority Queue to implement a relaxed priority queue whose main innovation is minimizing the overhead of delete-min via processing thread-local elements most of the time, and performing batch-removal on the main structure when local elements become exhausted. The approach maintains a tighter bound on the priority of the removed element compared to recent, state-of-the-art relaxed designs.