> There's always the option of a mixed architecture cluster; buy lots of > cheap 32-bit boxes for the vast majority of compute tasks which can > run in a 4GB address space, and then spend the money saved to buy one > or two large and much more expensive machines to handle the very large > memory tasks. This is exactly the option we have been taking. I think the only difference is that we really cannot afford lots of money on the second part, (i.e. to address more than 4GB of space) and my message was related to cheap options to overcome this issue. > really quite affordable [ as Itanic machines go ], and if you want a > 64GB machine with, say, 4 CPUs, they're only mildly expensive :-) He he, Tim, I am afraid "mildly" is still a scary figure for most labs ;) But they are interesting, I admit my ignorance, I thought the SGIs were not running linux, now I realized they do, that changes things... > The main issue with several types of machine in the same cluster is of > course that the users then have to specify their requirements to the > DRM so that it schedules to the right kind of machine. And I don't > know about you guys, but ours are notoriously bad at estimating what > their jobs' requirements actually are. I know exactly what you mean, luckily we are small enough groups to manage these issues usually, most of the times it really involves talking to the single person to find out what they are planning to do and why they plan to use a certain hardware. Incredibly often things that could be run in small parallel jobs on a cluster are run as one silly job on an SMP... Elia --- Telethon Institute of Genetics and Medicine Via Pietro Castellino, 111 80131 Napoli Tel. +39 081 6132 335 Fax. +39 081 560 98 77 -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: text/enriched Size: 1817 bytes Desc: not available Url : http://bioinformatics.org/pipermail/bioclusters/attachments/20050228/6be1eca1/attachment.bin