On Fri, 2005-06-24 at 12:51 -0400, William Van Etten wrote: > James, > > I wanted to mention IBT (informatics benchmarking toolkit) as a > potential extensible informatics cluster benchmarking tool. Cool, that was actually one of the first things pointed out in my first message. Thanks for the clarification though Bill. IBT could be a way to get the portability issues nailed, I've never got the patience to write that kinda portable code, this combined with Joe's stuff will no doubt help out a lot. Just for giggles, I've knocked up a dev/test/play mysql/php thingy at http://topbiocluster.org/testdat/ to start to have a think about what things could look like if we used the simple minded protein recipe I was blethering about in the last message. The data's nonsense, but as a start it gives a feel for how we could think about presenting the data. Right now I really quite like the concept of wallclock/qsize/dbsize vector as a potential measure of 'bestness'. Ok, enough for today, have a good weekend folks, and have a bit of a think about what we should do next. Best, j.