skip to main content

Explore Research Products in the PAR
It may take a few hours for recently added research products to appear in PAR search results.


Title: Sojourn Time Minimization of Successful Jobs
Due to a growing interest in deep learning applications [5], compute-intensive and long-running (hours to days) training jobs have become a significant component of datacenter workloads. A large fraction of these jobs is often exploratory, with the goal of determining the best model structure (e.g., the number of layers and channels in a convolutional neural network), hyperparameters (e.g., the learning rate), and data augmentation strategies for the target application. Notably, training jobs are often terminated early if their learning metrics (e.g., training and validation accuracy) are not converging, with only a few completing successfully. For this motivating application, we consider the problem of scheduling a set of jobs that can be terminated at predetermined checkpoints with known probabilities estimated from historical data. We prove that, in order to minimize the time to complete the first K successful jobs on a single server, optimal scheduling does not require preemption (even when preemption overhead is negligible) and provide an optimal policy; advantages of this policy are quantified through simulation. Related Work. While job scheduling has been investigated extensively in many scenarios (see [6] and [2] for a survey of recent result), most policies require that the cost of waiting times of each job be known at scheduling time; in contrast, in our setting the scheduler does not know which job will be the K-th successful job, and sojourn times of subsequent jobs do not contribute to the target metric. For example, [4, 3] minimize makespan (i.e., the time to complete all jobs) for known execution times and waiting time costs; similarly, Gittins index [1] and SR rank [7] minimize expected sojourn time of all jobs, i.e., both successfully completed jobs and jobs terminated early. Unfortunately, scheduling policies not distinguishing between these two types of jobs may favor jobs where the next stage is short and leads to early termination with high probability, which is an undesirable outcome in our applications of interest.  more » « less
Award ID(s):
1816887
PAR ID:
10376773
Author(s) / Creator(s):
; ;
Date Published:
Journal Name:
ACM SIGMETRICS Performance Evaluation Review
Volume:
50
Issue:
2
ISSN:
0163-5999
Page Range / eLocation ID:
24 to 26
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; ( , IEEE Transactions on Parallel and Distributed Systems)
    null (Ed.)
    While cloud platforms enable users to rent computing resources on demand to execute their jobs, buying fixed resources is still much cheaper than renting if their utilization is high. Thus, optimizing cloud costs requires users to determine how many fixed resources to buy versus rent based on their workload. In this paper, we introduce the concept of a waiting policy for cloud-enabled schedulers, which is the dual of a scheduling policy, and show that the optimal cost depends on it. We define multiple waiting policies and develop simple analytical models to reveal their tradeoff between fixed resource provisioning, cost, and job waiting time. We evaluate the impact of these waiting policies on a year-long production batch workload consisting of 14M jobs run on a 14.3k-core cluster, and show that a compound waiting policy decreases the cost (by 5%) and mean job waiting time (by 7x) compared to a fixed cluster of the current size. 
    more » « less
  2. ; ; ; ( , Proceedings of the ACM on Measurement and Analysis of Computing Systems)
    Multiserver-job systems, where jobs require concurrent service at many servers, occur widely in practice. Essentially all of the theoretical work on multiserver-job systems focuses on maximizing utilization, with almost nothing known about mean response time. In simpler settings, such as various known-size single-server-job settings, minimizing mean response time is merely a matter of prioritizing small jobs. However, for the multiserver-job system, prioritizing small jobs is not enough, because we must also ensure servers are not unnecessarily left idle. Thus, minimizing mean response time requires prioritizing small jobs while simultaneously maximizing throughput. Our question is how to achieve these joint objectives. We devise the ServerFilling-SRPT scheduling policy, which is the first policy to minimize mean response time in the multiserver-job model in the heavy traffic limit. In addition to proving this heavy-traffic result, we present empirical evidence that ServerFilling-SRPT outperforms all existing scheduling policies for all loads, with improvements by orders of magnitude at higher loads. Because ServerFilling-SRPT requires knowing job sizes, we also define the ServerFilling-Gittins policy, which is optimal when sizes are unknown or partially known. 
    more » « less
  3. ; ; ; ( , SC20: International Conference for High Performance Computing, Networking, Storage and Analysis)
    null (Ed.)
    While cloud platforms enable users to rent computing resources on demand to execute their jobs, buying fixed resources is still much cheaper than renting if their utilization is high. Thus, optimizing cloud costs requires users to determine how many fixed resources to buy versus rent based on their workload. In this paper, we introduce the concept of a waiting policy for cloud-enabled schedulers, which is the dual of a scheduling policy, and show that the optimal cost depends on it. We define multiple waiting policies and develop simple analytical models to reveal their tradeoff between fixed resource provisioning, cost, and job waiting time. We evaluate the impact of these waiting policies on a year-long production batch workload consisting of 14Mjobs run on a 14.3k-core cluster, and show that a compound waiting policy decreases the cost (by 5%) and mean job waiting time (by 7×) compared to a fixed cluster of the current size. 
    more » « less
  4. ; ( , Proceedings of the ACM on Measurement and Analysis of Computing Systems)
    We study the problem of scheduling jobs in a queueing system, specifically an M/G/1 with light-tailed job sizes, to asymptotically optimize the response time tail. This means scheduling to make P[T > t], the chance a job's response time exceeds t, decay as quickly as possible in the t \to \infty limit. For some time, the best known policy was First-Come First-Served (FCFS), which has an asymptotically exponential tail: P[T > t] ~ C e^-γ t . FCFS achieves the optimal decay rate γ, but its tail constant C is suboptimal. Only recently have policies that improve upon FCFS's tail constant been discovered. But it is unknown what the optimal tail constant is, let alone what policy might achieve it. In this paper, we derive a closed-form expression for the optimal tail constant C, and we introduce γ-Boost, a new policy that achieves this optimal tail constant. Roughly speaking, γ-Boost operates similarly to FCFS, but it pretends that small jobs arrive earlier than their true arrival times. This significantly reduces the response time of small jobs without unduly delaying large jobs, improving upon FCFS's tail constant by up to 50% with only moderate job size variability, with even larger improvements for higher variability. While these results are for systems with full job size information, we also introduce and analyze a version of γ-Boost that works in settings with partial job size information, showing it too achieves significant gains over FCFS. Finally, we show via simulation that γ-Boost has excellent practical performance. 
    more » « less
  5. ; ( , Leibniz international proceedings in informatics)
    The BUNDLE and BUNDLEP scheduling algorithms are cache-cognizant thread-level scheduling algorithms and associated worst case execution time and cache overhead (WCETO) techniques for hard real-time multi-threaded tasks. The BUNDLE-based approaches utilize the inter-thread cache benefit to reduce WCETO values for jobs. Currently, the BUNDLE-based approaches are limited to scheduling a single task. This work aims to expand the applicability of BUNDLE-based scheduling to multiple task multi-threaded task sets. BUNDLE-based scheduling leverages knowledge of potential cache conflicts to selectively preempt one thread in favor of another from the same job. This thread-level preemption is a requirement for the run-time behavior and WCETO calculation to receive the benefit of BUNDLE-based approaches. This work proposes scheduling BUNDLE-based jobs non-preemptively according to the earliest deadline first (EDF) policy. Jobs are forbidden from preempting one another, while threads within a job are allowed to preempt other threads. An accompanying schedulability test is provided, named Threads Per Job (TPJ). TPJ is a novel schedulability test, input is a task set specification which may be transformed (under certain restrictions); dividing threads among tasks in an effort to find a feasible task set. Enhanced by the flexibility to transform task sets and taking advantage of the inter-thread cache benefit, the evaluation shows TPJ scheduling task sets fully preemptive EDF cannot. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email