html5-img
1 / 12

Surplus Fair Scheduling

Ricardo A. Baratto COMS E6118 – Spring 2004. Surplus Fair Scheduling. Proportional-Share CPU Scheduling for Symmetric Multiprocessors. What? Why?. Fair resource allocation Proportional to application's share WFQ [Demers et al.], [Parekh et al.], Stride [Waldspurger et al.]

felton
Download Presentation

Surplus Fair Scheduling

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Ricardo A. Baratto COMS E6118 – Spring 2004 Surplus Fair Scheduling

  2. Proportional-Share CPU Scheduling for Symmetric Multiprocessors

  3. What? Why? • Fair resource allocation • Proportional to application's share • WFQ [Demers et al.], [Parekh et al.], Stride [Waldspurger et al.] • SFQ [Goyal et al.] • SMART [Nieh et al.]

  4. The Problem • Unfeasible Weight Assignment: Starvation • Upper Bound on Assignment • Frequent Arrival – Departures: Unfairness • Burst Behavior

  5. Surplus Fair Scheduling • Weight Readjustment • Generalized Multiprocessor Sharing(Surplus Approximation)

  6. Weight Readjustment • Readjust threads violating feasibility constraint: • Optimal: • Number of threads • Weight change • Efficient • O(p)

  7. GMS and Surplus Approximation • Generalization of GPS [Parekh et al.] • φi: Readjusted weight • Ai(t1, t2): CPU service received • Si: Weighted CPU service = Si + q/φi • αi = Ai(0, t) - AiGMS(0, t)≈ (φi . Si) - (φi . v)

  8. SFS Algorithm • update Fi and compute Si • compute v: min(Si) • compute αi • schedule ti with min(αi) • If {t} changes: readjust weights

  9. Implementation and Complexity • Three ordered queues: • w: Weight readjustment • S: Virtual time • α: Next to schedule • Add thread: Insert + Readjust = O(t+p) • Delete thread: Delete + Readjust = O(p) • Schedule: Update surplus + Sort = O(t log t) [~ every p schedules]

  10. Evaluation: Weight Readjustment

  11. Evaluation: Short Jobs

  12. Evaluation: Application Isolation

More Related