Loading...
Please wait, while we are loading the content...
Similar Documents
Best of Both Worlds: High Performance Interactive and Batch Launching
| Content Provider | arXiv |
|---|---|
| Author | Byun, Chansup Kepner, Jeremy Arcand, William Bestor, David Bergeron, Bill Gadepally, Vijay Houle, Michael Hubbell, Matthew Jones, Michael Kirby, Andrew Klein, Anna Michaleas, Peter Milechin, Lauren Mullen, Julie Prout, Andrew Rosa, Antonio Samsi, Siddharth Yee, Charles Reuther, Albert |
| Date of Submission | 2020-08-05 |
| Abstract | Rapid launch of thousands of jobs is essential for effective interactive supercomputing, big data analysis, and AI algorithm development. Achieving thousands of launches per second has required hardware to be available to receive these jobs. This paper presents a novel preemptive approach to implement spot jobs on MIT SuperCloud systems allowing the resources to be fully utilized for both long running batch jobs while still providing fast launch for interactive jobs. The new approach separates the job preemption and scheduling operations and can achieve 100 times faster performance in the scheduling of a job with preemption when compared to using the standard scheduler-provided automatic preemption-based capability. The results demonstrate that the new approach can schedule interactive jobs preemptively at a performance comparable to when the required computing resources are idle and available. The spot job capability can be deployed without disrupting the interactive user experience while increasing the overall system utilization. |
| Related Links | https://arxiv.org/pdf/2008.02223.pdf |
| arXiv | 2008.02223 |
| Language | English |
| Access Restriction | Open |
| Subject Keyword | Computer Science - Distributed, Parallel, and Cluster Computing Computer Science |
| Content Type | Text |
| Resource Type | Article |
| Subject | Computer Science |