On 9 Feb 2006, at 4:08 pm, Bonnie Hurwitz wrote: > Hello, > > We are in the process of deciding on a distributive management > software for a 250 node cluster and are concerned about performance > and issues that come up with a cluster this size. > > Are there are web sites were we can get some nice comparisons and > solid examples of issues that people run into using different DSM > software? We would like to compare and contrast SGE, LSF and PBS. I can't speak for PBS, but both SGE and LSF can happily manage clusters that size. Have a look through the archives of this list - a few months ago Chris Dagdigian posted a very good message about SGE and LSF, and which is "best". As with most things, it depends on what you want to do. SGE wins on price of course, but Platform offer some layered products over the top of LSF which can give it the edge, but only if you specifically need that functionality. For simple batch scheduling tasks both systems are more than adequate. I tend to favour LSF myself, but that's purely because I am more experienced with it, and because I have found Platform's support to be generally excellent. But we have in the past run the Ensembl pipeline on our cluster (over 1000 CPUs) using SGE instead of LSF, just to see whether it would work, and it was fine. Tim -- Dr Tim Cutts Informatics Systems Group, Wellcome Trust Sanger Institute GPG: 1024D/E3134233 FE3D 6C73 BBD6 726A A3F5 860B 3CDD 3F56 E313 4233