Adaptive Scheduling with Parallelism Feedback

Abstract

Multiprocessor scheduling in a shared multiprogramming environment is often structured as two-level scheduling, where a kernellevel job scheduler allots processors to jobs and a user-level task scheduler schedules the work of a job on the allotted processors. In this context, the number of processors allotted to a particular job may vary during the job's execution, and the task scheduler must adapt to these changes in processor resources. For overall system efficiency, the task scheduler should also provide parallelism feedback to the job scheduler to avoid the situation where a job is allotted processors that it cannot use productively.We present an adaptive task scheduler for multitasked jobs with dependencies that provides continual parallelism feedback to the job scheduler in the form of requests for processors. Our scheduler guarantees that a job completes near optimally while utilizing at least a constant fraction of the allotted processor cycles. Our scheduler can be applied to schedule data-parallel programs, such as those written in High Performance Fortran (HPF), *Lisp, C*, NESL, and ZPL.Our analysis models the job scheduler as the task scheduler's adversary, challenging the task scheduler to be robust to the system environment and the job scheduler's administrative policies. For example, the job scheduler can make available a huge number of processors exactly when the job has little use for them. To analyze the performance of our adaptive task scheduler under this stringent adversarial assumption, we introduce a new technique called "trim analysis," which allows us to prove that our task scheduler performs poorly on at most a small number of time steps, exhibiting nearoptimal behavior on the vast majority.To be precise, suppose that a job has work T1 and critical-path length T∞ and is running on a machine with P processors. Using trim analysis, we prove that our scheduler completes the job in O(T1/ P + T∞ + L lg P ) time steps, where L is the length of a scheduling quantum and P denotes the O(T∞ + L lg P )-trimmed availability. This quantity is the average of the processor availabilThis research was supported in part by the Singapore-MIT Alliance and NSF Grant ACI-0324974. Yuxiong He is a Visiting Scholar at MIT CSAIL and a Ph.D. candidate at the National University of Singapore. Wen Jing Hsu is a Visiting Scientist at MIT CSAIL and Associate Professor at Nanyang Technological University.Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. ity over all time steps excluding the O(T∞ + L lg P ) time steps with the highest processor availability. When T1/T∞ P (the job's parallelism dominates the O(T∞ + L lg P )-trimmed availability), the job ach...

A Batch System with Efficient Adaptive Scheduling for Malleable and Evolving Applications

“…However, feedback mechanisms introduce other overheads such as too frequent communication, inconsistency (as iteration times are not always constant), and increased complexity for non-iterative malleable applications. Thus, efficient feedback mechanisms for malleable applications have been exclusively studied by many [24], [25]. Scheduling based on feedback from application on its scaling pattern is our interest for investigation in the future and is out of scope of this work.…”

Multi-core real-time scheduling for generalized parallel task models

“…There has also been extensive work on scheduling of one or more parallel jobs on multiprocessors [13]- [22]. However, the work in [13]- [19] does not consider task deadlines, and that in [20]- [22] considers soft real-time scheduling. In contrast to the goal (i.e.…”

Provably-Efficient Online Adaptive Scheduling of Parallel Jobs Based on Simple Greedy Rules

“…Adaptive Greedy Thread Scheduler A-GREEDY [1] is an adaptive greedy thread scheduler with parallelism feedback. Between quanta, it estimates its job's desire, and requests processors from the job scheduler.…”

“…Otherwise, the quantum is deprived. 1 The quantum q is efficient if A-GREEDY utilizes no less than a  fraction of the total allotted processor cycles during the quantum, where  is the utilization parameter. Otherwise, the quantum is inefficient.…”

“…GRAD is a two-level scheduling algorithm that uses simple, greedy-like rules. The threadlevel scheduler called A-GREEDY [1] provides feedback based on two simple indicators acquired from the past quantum, namely, whether its request was satisfied and whether the allotted processors are well utilized. Based on the feedbacks from all jobs, the OS allocator RAD [19] partitions processors as equally as possible.…”