Some test results, thanks to Ian H.: http://webperf.org/a2/v30/ This compares today's httpd-2.0 CVS head with my "recycle pools" patch for the worker MPM and Sander's new implementation of pools.
Sander's pool code yielded a substantial drop in usr CPU consumption (see the "cpuu" links on the test result pages) and a corresponding increase in throughput. My worker MPM patch was faster than the current CVS head, but not as fast as Sander's pool code. (I'd expected the two changes to be complementary, but the combination of my worker patch with Sander's pool changes wasn't as fast as the pool changes alone. Maybe because recycling the pools, rather than destroying them, reduced the cache locality on the multiprocessor test server?)
--Brian