Looking for benchmark test tool 2322
On 13 Jul 2006 13:37:14 -0700, dbone staggered into the Black Sun and said:
There are 3 species of mendacity, each worse than the last: Lies, damned lies, and benchmarks.
On Fri, 14 Jul 2006 17:14:12 GMT, Cardiff IT Support Ltd staggered into the Black Sun and said: You can't, at least not easily. If you'd started your initial session with screen...
The best thing to do would be to figure out what these machines are going to be doing, then run sets of processes that are as close to that as possible. Doing anything else will probably give you results that are completely useless, or totally wrong, or both. If you're going to run a bloody huge Java monstrosity on Blackdown 4.5.6, install Blackdown 4.5.6 on all machines. Do something like "export NAME=`hostname`-report.txt ; summary of some important hardware info for the machine. Then run "time javapath-to-stuff dataset1 ; time javapath-to-stuff dataset2 ; ...
Remember that you'll have to make your requirements more specific than what you posted above if you want good data. You *may* be able to get a reasonable disk benchmark out of bonnie++. What exactly did you mean by "thread count"? Something like "How many instances of $CRUD can I run at once before the machine becomes glacially slow"? Ah well, HTH,
Chris F.A. Johnson I think that is unfair to statisticians (other than those kept as pets). A...
-- Matt GThere is no Darkness in Eternity-But only Light too dim for us to see Brainbench MVP for Linux Admin mail: TRAP + SPAN don't belong ----------------------------- penguins, is Tux." --MegaHAL
Looking for benchmark test tool 2323
Linux groups from Newsgroups