[Beowulf] Re: Beowulf Digest, Vol 22, Issue 12
Many of your questions may have already been answered in earlier discussions or in the FAQ. The search results page will indicate current discussions as well as past list serves, articles, and papers.
rod mach rwm at absoft.comTue Dec 13 12:30:33 PST 2005
- Previous message: [Beowulf] [owner-chemistry@ccl.net: CCL:G: dual core opteron]
- Next message: [Beowulf] Software Raid and Raid in general
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
> > On Mon, Dec 05, 2005 at 12:28:01PM +0000, enver ever wrote: > >> Can we assume that the job arrivals to a typical Beowulf system >> can be >> exponential? > Research done at the University of Michigan by Tom Hacker (now at Indiana University) indicated that for two sites he analyzed (Both U of M and UT Austin TACC) the job arrival rate can be modeled as a Poisson distribution. That might not hold true for all sites, but the paper is a good read. The paper analyzed 7,294 jobs over a period of 44 weeks. The University of Texas logs contain 3,160 jobs over a period of 27 weeks. The paper discusses formulas and methods to predicate peak jobs, unique number of users, and other interesting items in relation to using template accounts for grid computing. See: http://www.mgrid.umich.edu/publications/grid_computing_2001.pdf -Rod
- Previous message: [Beowulf] [owner-chemistry@ccl.net: CCL:G: dual core opteron]
- Next message: [Beowulf] Software Raid and Raid in general
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Beowulf mailing list
