A Proposal of Job-Worker Assignment Algorithm Considering CPU Core Utilization for User-PC Computing System

Kamoyedji, Ariel; Funabiki, Nobuo; Htet, Hein; Kuribayashi, Minoru

doi:10.18178/ijfcc.2022.11.2.586

Cited by 1 publication

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Previously, we proposed the algorithm of assigning non-uniform jobs to workers in the UPC system [2]. In non-uniform jobs, the programs are much different from each other, including the developed programming languages, the number of threads, and the requiring data.…”

Section: Introductionmentioning

confidence: 99%

A Static Assignment Algorithm of Uniform Jobs to Workers in a User-PC Computing System Using Simultaneous Linear Equations

et al. 2022

Self Cite

View full text Add to dashboard Cite

Currently, the User-PC computingsystem (UPC) has been studied as a low-cost and high-performance distributed computing platform. It uses idling resources of personal computers (PCs) in a group.The job-worker assignment for minimizing makespan is critical to determine the performance of the UPC system.Some applications need to execute a lot of uniform jobs that use the identical program but with slightly different data, where they take the similar CPU time on a PC. Then, the total CPU time of a worker is almost linear to the number of assigned jobs. In this paper, we propose a static assignment algorithm of uniform jobs to workers in the UPC system, using simultaneous linear equations to find the lower bound on makespan, where every worker requires the same CPU time to complete the assigned jobs. For the evaluations of the proposal, we consider the uniform jobs in three applications. In OpenPose, the CNN-based keypoint estimation program runs with various images of human bodies. In OpenFOAM, the physics simulation program runs with various parameter sets. In code testing, two open-source programs run with various source codes from students for the Android programming learning assistance system (APLAS). Using the proposal, we assigned the jobs to six workers in the testbed UPC system and measured the CPU time. The results show that makespan was reduced by 10% on average, which confirms the effectiveness of the proposal.

show abstract

Section: Introductionmentioning

confidence: 99%