Hi,
On 1/9/06, Rony Bill <[EMAIL PROTECTED]> wrote: > > JTD wrote: > > If you or anyone else has a cluster running somewhere for graphics, > could I get a short writeup on the hardware used, how it is joined > together (pc to pc) and just a wee bit on its software and setup. It is > just for virtual practicals. :) Rony there are only 2 large installations that I have seen. A render farm typically consist about a few tens to few hundreds of dual processor HT / xeon (em64t) class systems with some expensive display cards and atleast a few gigs of high speed ram . needless to say about raid sata / scsi disk . say if there are 11 boxes the first box serves as the grid master with 10 boxes running the grid engine client. the master distributes the task to the clients . the task can be copied over a nfs share on a network . this is for a company that some real anim work for movies etc. u can certainly setup a lowerend system for small testing etc ... sun grid engine project might be a nice place to start. maybe sejal can do a write up once they are up
Check out this month's PCQuest for instructions on setting up your own clustered super-computer. Uses the openmosix kernel patch for process migration. Other options are also discussed.
-RR -- http://mm.ilug-bom.org.in/mailman/listinfo/linuxers
