Responsive Parallelism
Overview
Multicore hardware is no longer present only in large data centers handling specially-written batch computations: almost all consumer hardware today is multicore, and so the landscape of parallel computing has changed. Parallel software must now be able to handle features of consumer software like user interaction and responsiveness requirements. This is particularly a challenge for the frameworks and models of implicit parallelism, which are elegant and have been well-designed and tested over decades, but are not prepared for this changing landscape. This project seeks to extend the models and frameworks with facilities for interaction, responsiveness, fairness, and other hallmarks of consumer software.
People
Collaborators
- Umut A. Acar (Carnegie Mellon)
- Angelina Lee (Washington U. in St. Louis)
- Kunal Agrawal (Washington U. in St. Louis)
Former Collaborators
- Jatin Arora (Carnegie Mellon)
- Sam Westrick (Carnegie Mellon)
- Kyle Singer (Washington U. in St. Louis)
- Noah Goldstein (Washington U. in St. Louis)
Funding
NSF CCF-2107289: Responsive Parallelism for Interactive Applications: Theory and Practice