On Wed, May 29, 2002 at 10:58:41PM +1200, James Kahn wrote: > On Wed, 2002-05-29 at 22:16, Andrew Suffield wrote: > > On Wed, May 29, 2002 at 09:48:33AM +0200, Ulrich Eckhardt wrote: > > > Hmmm, everyone can do so on their own machines easily .... how are the > > > results on yours ? :) > > > > > > I'll give it a more systematic try with a linux-kernel sourcetree instead of > > > my cvsroot, but that's later today. > > > > That is not a deterministic test. It doesn't provide a particularly > > useful benchmark; it involves a semi-random pattern of disk access > > which is hopelessly skewed by environmental factors. > > Far from being a "deterministic" test, it's not a test of CPU at all. > The CPU will be mostly idle waiting for data from the hard disk and bus. > > > If you are trying to do a benchmark on processor-related performance > > changes (which it seems like you are), you will have to construct a > > benchmark which actually measures that (no, I don't have any > > suggestions offhand). > > Ray tracing tends to be pretty CPU intensive, how about using povray? > Of course, to get "real" results, you'll need it to trace a rather large > scene a few times with each buildarch, throw away any outliers and draw > the mean of the rest. If bzip2 should be used as a benchmark, write a small program that generates random data and pipe this to bzip2: genran | bzip2 -c > /dev/null This should not generate any disk IO. Of course, the data generating program should not be CPU intensive itself. Greetings, Oliver >
Attachment:
pgpXdPYUPj_k_.pgp
Description: PGP signature