Advanced search
1 file | 629.73 KB Add to list
Author
Organization
Abstract
Symbiotic job scheduling exploits the fact that in a system with shared resources, the performance of jobs is impacted by the behavior of other co-running jobs. By coscheduling combinations of jobs that have low interference, the performance of a system can be increased. In this paper, we investigate the impact of using symbiotic job scheduling for increasing throughput. We find that even for a theoretically optimal scheduler, this impact is very low, despite the substantial sensitivity of per job performance to which other jobs are coscheduled: for example, our experiments on a 4-thread SMT processor show that, on average, the job IPC varies by 37% depending on coscheduled jobs, the per-coschedule throughput varies by 69%, and yet the average throughput gain brought by optimal symbiotic scheduling is only 3%. This small margin of improvement can be explained by the observation that all the jobs need to be eventually executed, restricting the job combinations a symbiotic job scheduler can select to optimize throughput. We explain why previous work reported a substantial gain from symbiotic job scheduling, and we find that (only) reporting turnaround time can lead to misleading conclusions. Furthermore, we show how the impact of scheduling can be evaluated in microarchitectural studies, without having to implement a scheduler.
Keywords
PROCESSORS

Downloads

  • (...).pdf
    • full text
    • |
    • UGent only
    • |
    • PDF
    • |
    • 629.73 KB

Citation

Please use this url to cite or link to this publication:

MLA
Eyerman, Stijn, Pierre Michaud, and Wouter Rogiest. “Revisiting Symbiotic Job Scheduling.” IEEE International Symposium on Performance Analysis of Systems and Software-ISPASS. 2015. 124–134. Print.
APA
Eyerman, S., Michaud, P., & Rogiest, W. (2015). Revisiting symbiotic job scheduling. IEEE International Symposium on Performance Analysis of Systems and Software-ISPASS (pp. 124–134). Presented at the IEEE INTERNATIONAL SYMPOSIUM ON PERFORMANCE ANALYSIS OF SYSTEMS AND SOFTWARE (ISPASS).
Chicago author-date
Eyerman, Stijn, Pierre Michaud, and Wouter Rogiest. 2015. “Revisiting Symbiotic Job Scheduling.” In IEEE International Symposium on Performance Analysis of Systems and Software-ISPASS, 124–134.
Chicago author-date (all authors)
Eyerman, Stijn, Pierre Michaud, and Wouter Rogiest. 2015. “Revisiting Symbiotic Job Scheduling.” In IEEE International Symposium on Performance Analysis of Systems and Software-ISPASS, 124–134.
Vancouver
1.
Eyerman S, Michaud P, Rogiest W. Revisiting symbiotic job scheduling. IEEE International Symposium on Performance Analysis of Systems and Software-ISPASS. 2015. p. 124–34.
IEEE
[1]
S. Eyerman, P. Michaud, and W. Rogiest, “Revisiting symbiotic job scheduling,” in IEEE International Symposium on Performance Analysis of Systems and Software-ISPASS, Philidelphia, PA, 2015, pp. 124–134.
@inproceedings{6855659,
  abstract     = {Symbiotic job scheduling exploits the fact that in a system with shared resources, the performance of jobs is impacted by the behavior of other co-running jobs. By coscheduling combinations of jobs that have low interference, the performance of a system can be increased. In this paper, we investigate the impact of using symbiotic job scheduling for increasing throughput. We find that even for a theoretically optimal scheduler, this impact is very low, despite the substantial sensitivity of per job performance to which other jobs are coscheduled: for example, our experiments on a 4-thread SMT processor show that, on average, the job IPC varies by 37% depending on coscheduled jobs, the per-coschedule throughput varies by 69%, and yet the average throughput gain brought by optimal symbiotic scheduling is only 3%. This small margin of improvement can be explained by the observation that all the jobs need to be eventually executed, restricting the job combinations a symbiotic job scheduler can select to optimize throughput.

We explain why previous work reported a substantial gain from symbiotic job scheduling, and we find that (only) reporting turnaround time can lead to misleading conclusions. Furthermore, we show how the impact of scheduling can be evaluated in microarchitectural studies, without having to implement a scheduler.},
  author       = {Eyerman, Stijn and Michaud, Pierre and Rogiest, Wouter},
  booktitle    = {IEEE International Symposium on Performance Analysis of Systems and Software-ISPASS},
  isbn         = {978-1-4799-1957-4},
  keywords     = {PROCESSORS},
  language     = {eng},
  location     = {Philidelphia, PA},
  pages        = {124--134},
  title        = {Revisiting symbiotic job scheduling},
  year         = {2015},
}

Web of Science
Times cited: