[Beowulf] Why one might want a bunch o' processors under your desk.
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.
Jim Lux james.p.lux at jpl.nasa.govFri May 6 21:22:14 PDT 2005
- Previous message: [Beowulf] Why one might want a bunch o' processors under your desk.
- Next message: [Beowulf] CCL:Opteron or Nocona ? (fwd from cavallo@chemistry.unina.it)
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
From: "Greg Lindahl" <lindahl at pathscale.com> > On Fri, May 06, 2005 at 05:49:25PM -0700, Jim Lux wrote: > > > I suspect that I am not alone. There are probably hundreds of people who > > have similar kinds of problems, and would be well served by a desktop or > > personal supercomputer. > > Embarrassingly parallel computations should be embarrassingly easy to > run on a cluster. > > Few clusters meet that standard. > > -- greg Fascinatingly, the problem is not in starting all the jobs in parallel (that was a script taking <30minutes to write). It's gathering and looking at the data afterwards. I suppose the real problem is in the user interface. The code I'm running has a long punched card heritage (yesirree, the input file is called a "NEC deck", and the output file has carriage control characters in column 1), so there's a wide variety of "front ends" to it to make the deck prep and output viewing easier. It's these prep and output viewer programs that need to be "cluster aware".
- Previous message: [Beowulf] Why one might want a bunch o' processors under your desk.
- Next message: [Beowulf] CCL:Opteron or Nocona ? (fwd from cavallo@chemistry.unina.it)
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Beowulf mailing list
