[ On Friday, May 17, 2002 at 23:04:30 (-0400), Jason R. Fink wrote: ]
> Subject: Softdeps
>
> I would like to do a *valid* study of softdeps
> impact on system performance. I have enabled it
> on all of my hosts and so far am pleased, but
> I want to find the catch.
> 
> Thusfar I have forumulated the following tools 
> I can use to measure the impact of softdep:
>         
>         bsdbench (paging/mfs usage)  
>         bonnie
>         time ls -1R /usr/pkg
>         time du -s /usr/pkg

Don't forget pkgsrc/benchmarks/postmark....

Also untar pkgsrc.tar.gz and then "rm -rf" it.

> So, my question to the gang is, I want one tool,
> one test, I am thinking bonnie.

no, postmark would be better.  bonnie won't do much different for
softdep -- it's geared for testing maximum I/O rates on a single file,
not general throughput of all operations.

When I was testing external RAID hardware performance vs. cache size and
chunk-size I used the following:

# printf 'set bias read 5\nset bias create 5\nset size 10240 20480\nset transactions 100000\nset subdirectories 10\nset number 10000\nshow\nrun\n' | postmark 


-- 
								Greg A. Woods

+1 416 218-0098;  <gwoods@acm.org>;  <g.a.woods@ieee.org>;  <woods@robohack.ca>
Planix, Inc. <woods@planix.com>; VE3TCP; Secrets of the Weird <woods@weird.com>