Sure you can get closer to the OP's system performance, but ultimately
there are too many variables for the benchmark to be completely
predictive on the OP's system. Not just data, but the physical ordering
of the data, the distribution over devices and controllers ... just too
much for it to be reliable. Benchmarks run by other people are
sometimes more interesting more for the differences thay show against
your own system than the similarities -- actually, those are the more
interesting discussions ;)
just my 2 cents.
Received on Sat Feb 05 2005 - 11:45:25 CST
