Parallel Programming and Concurrent Data Structures

The advent of multi-cores and other multi-processor architectures has led to an increasing interest in various parallel programming paradigms. Our research here focuses on parallelizing Graph and Linear Algebra algorithms targeting multi-cores and GPGPUs. Concurrent data-structures is another focus area of our research group. Programmers have been using increasingly complex data structures rather than simple arrays like a decade ago. These pointer intensive data structures could be sources of bottlenecks in parallelizing programs. Here we are studying the benefits of transactional memories and lock-free data structures.

Faculty

1. R Govindarajulu (Head)