Coscheduling in Clusters [electronic resource] : Is It a Viable Alternative?

As clusters are widely accepted as cost-effective infrastructures for many scientific and commercial applications, improving the deliverable performance and reducing the energy consumption of such systems has become a pressing issue. In this paper, we exploit the feasibility of achieving these objec...

Full description

Saved in:
Bibliographic Details
Online Access: Online Access (via OSTI)
Corporate Author: Lawrence Berkeley National Laboratory (Researcher)
Format: Government Document Electronic eBook
Language:English
Published: Washington, D.C. : Oak Ridge, Tenn. : United States. Department of Energy. ; distributed by the Office of Scientific and Technical Information, U.S. Department of Energy, 2003.
Subjects:

MARC

LEADER 00000nam a22000003u 4500
001 b7012893
003 CoU
005 20180821000443.1
006 m o d f
007 cr |||||||||||
008 190514e20031110||| o| f1|||||eng|d
035 |a (TOE)ost15013790 
035 |a (TOE)15013790 
040 |a TOE  |c TOE 
049 |a GDWR 
072 7 |a 99  |2 edbsc 
086 0 |a E 1.99:ucrl-conf-200869 
086 0 |a E 1.99:ucrl-conf-200869 
088 |a ucrl-conf-200869 
245 0 0 |a Coscheduling in Clusters  |h [electronic resource] :  |b Is It a Viable Alternative? 
260 |a Washington, D.C. :  |b United States. Department of Energy. ;  |a Oak Ridge, Tenn. :  |b distributed by the Office of Scientific and Technical Information, U.S. Department of Energy,  |c 2003. 
300 |a PDF-file: 14 pages; size: 0.8 Mbytes. 
336 |a text  |b txt  |2 rdacontent. 
337 |a computer  |b c  |2 rdamedia. 
338 |a online resource  |b cr  |2 rdacarrier. 
500 |a Published through SciTech Connect. 
500 |a 11/10/2003. 
500 |a "ucrl-conf-200869" 
500 |a Presented at: ACM SIGMETRICS - Performance 2004, New York, NY, United States, Jun 12 - Jun 16, 2004. 
500 |a Choi, G S; Kim, J H; Ersoz, D; Yoo, A B; Das, C R. 
520 3 |a As clusters are widely accepted as cost-effective infrastructures for many scientific and commercial applications, improving the deliverable performance and reducing the energy consumption of such systems has become a pressing issue. In this paper, we exploit the feasibility of achieving these objectives through efficiently scheduling the communicating processes of parallel applications. In this context, we conduct an in-depth evaluation of a broad spectrum of scheduling alternatives for clusters. These include the widely used batch scheduling, local scheduling, gang scheduling, all prior communication-driven coscheduling algorithms, and a newly proposed HYBRID coscheduling algorithm. In order to provide ease of implementation and portability across many cluster platforms, we propose a generic framework for deploying any coscheduling algorithm. We have implemented four prior coscheduling algorithms (Dynamic Coscheduling (DCS), Spin Block (SB), Periodic Boost (PB), and Co-ordinated Coscheduling (CC)) and the HYBRID coscheduling using this framework on a 16-node, Myrinet connected Linux cluster that uses GM as the communication layer. In addition, we use PBS as the batch scheduler and a previously proposed gang scheduler (SCore) to analyze all classes of scheduling techniques. Performance and energy measurements using several NAS and LLNL benchmarks on the Linux cluster provide several interesting conclusions. First, although batch scheduling is currently used in most clusters, all blocking-based coscheduling techniques such as SB, CC and HYBRID and the gang scheduling can provide much better performance even in a dedicated cluster platform. Under high system load, these coscheduling schemes can provide orders of magnitude reduction in average response time and much better performance-energy behavior compared to the PBS scheme. Second, in contrast to some of the prior studies, we observe that blocking-based schemes like SB and HYBRID can provide better performance than spin-based techniques like PB on a Linux platform. Third, the proposed HYBRID scheduling provides the best performance-energy behavior and can be implemented on any cluster with little effort. All these results suggest that blocking-based coscheduling techniques are viable candidates to be used instead of batching scheme for significant performance-energy benefits. 
536 |b W-7405-ENG-48. 
650 7 |a Energy Consumption.  |2 local. 
650 7 |a Lawrence Livermore National Laboratory.  |2 local. 
650 7 |a Computer Architecture.  |2 local. 
650 7 |a Performance.  |2 local. 
650 7 |a Computers.  |2 local. 
650 7 |a General And Miscellaneous//Mathematics, Computing, And Information Science.  |2 edbsc. 
710 2 |a Lawrence Berkeley National Laboratory.  |4 res. 
710 1 |a United States.  |b Department of Energy.  |4 spn. 
710 1 |a United States.  |b Department of Energy.  |b Office of Scientific and Technical Information.  |4 dst. 
856 4 0 |u http://www.osti.gov/scitech/biblio/15013790  |z Online Access (via OSTI) 
907 |a .b70128935  |b 03-09-23  |c 03-31-12 
998 |a web  |b 05-20-19  |c f  |d m   |e p  |f eng  |g    |h 0  |i 2 
956 |a Information bridge 
999 f f |i 21a78914-562d-59d4-811b-0951eee21027  |s 07bd1678-6d88-5282-8515-2f7718cfa6ce 
952 f f |p Can circulate  |a University of Colorado Boulder  |b Online  |c Online  |d Online  |e E 1.99:ucrl-conf-200869  |h Superintendent of Documents classification  |i web  |n 1