On Fri, 30 Dec 2005 06:43:35 -0700 (MST), Jim Prewett download@hpc.unm.edu wrote:
I'd like to compare cl-ppcre's performance on different systems. It seemed to me that the benchmarks (under #performance on the webpage) are more focused on comparing Perl to CL-PPCRE. Is that true?
Yes.
Anyone have thoughts on how to benchmark different systems in this way?
As long as you want to compare different Lisp implementations on the same OS you should be fine. If you want to compare across different operating systems note that the tests also output the number of repetitions and the time for each test so it should be relatively easy to tweak this a bit and get what you want.
Don't forget that those benchmark tests use GET-INTERNAL-REAL-TIME which isn't perfectly accurate. You should probably use TIME and measure user and system time.
If you come across interesting results please post them here.
Cheers, Edi.