Abstract
Commodity architectures are now parallel by default, yet apart from exceptional cases, the notorious challenges of parallel programming endure. On one hand, a few application domains have sustained a performance boom since the shift to multicores, in part due to their abundant obvious sources of parallelism. On the other hand, parallel algorithm and implementation experts have uncovered surprising opportunities for task-level parallelism in conventionally challenging domains [4, 12, 13, 17, 18, 30, 33]. Algorithms in the former domains typically have abundant regular parallelism, where data and control dependences among tasks are statically identifiable, while algorithms in the latter are challenged with irregular parallelism, with dynamically manifesting data and control dependences [30]. Scheduling tasks and synchronizing irregular data accesses continues to challenge programmers with pitfalls such as non-determinism [26], deadlock, data races, and other concurrency bugs [15, 27, 36]. While a plethora of work in programming languages [5, 10], language extensions [6-8, 32], and type systems [16, 28] has sought to curtail concurrency bugs, few have reached mainstream adoption.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
