And one more thing, the head node on a cluster is AMD. Everyone uses it to submit jobs, the head node should never be used to run calculations.
But I made a benchmark on head node during some weekend on the night, when nobody wasn't working. This head node with AMD is indeed twice faster than 8 CPU nodes E5440, and four times faster than 16 CPU nodes X7350 (under full load). Four times faster, can you imagine? But of course that head node is only 2 CPUs. So it cannot calculate that many as those 16 CPU nodes. But a single monothread job goes crazy fast. I suspect that when Gael told me that the cluster is twice faster than his old AMD sun workstations in his room (orioles, etc..) is that he was running a simulation on the AMD head node, instead of submitting it to a cluster. But I never had a chance to ask him. best regards -- Janek Kozicki | _______________________________________________ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp