On Tue, Jul 7, 2009 at 8:32 AM, Rocco Carbone <ro...@ntop.org> wrote:
> Hi list, > while reading niels's blog about performances increment in > libevent 2.x series, I was really surprised about lack of > "usable" benchmark tools in the current libevent distribution. > > So I cloned some functionalities of the so popular Apache 'ab' > benchmark tool and I started coding my own libevent-based > version from the ground. > > The resulting program 'lb', which is an acronym for > "libevent-based benchmark" tool, has been released > under the BSD license and it is available for download at: > > http://lb.tecsiel.it > > This is only a preliminary version which lacks from 'ab' > a lot of functionalities but adds: > > o benchmarking multi-servers at the same time > o reading HTTP servers parameters from a file (not full tested) > > > The package comes with: > 1. a simple patch file to add the directory contrib/lb to the > libevent SUBDIRS. > It only changes the {configure.in,Makefile.am} libevent files > in order to allow the top-level Makefile to compile > under the contrib/lb directory too. > > 2. The 'lb' sources in a tarball format > > Included in the package there is also the program 'wsl' which is > a micro libevent-based web server useful for testing lb itself. > > I really hope in the near future the tool be included in the > main libevent distribution, so the choice of the contrib/lb directory. > > have fun > > /rocco > -- > Rocco Carbone mail-to: <rocco /at/ ntop /dot/ org> > Pisa Italy > > Life is in the details > > In other words, you are welcome to use, share and improve me. > You are forbidden to forbid anyone else to use, share and > improve what I gave you. > _______________________________________________ > Libevent-users mailing list > Libevent-users@monkey.org > http://monkeymail.org/mailman/listinfo/libevent-users > Your work is much appreciated. I haven't looked at it yet, but I am very interested in trying it. Hopefully some time this week I can spare some time to give it a go. I have some development servers that are fairly old, and surprisingly the slower machine seems to have a quicker responce time than the faster one, although it's peak load is a bit lower. I was planning on writing a little benchmark to test libevent on them, but I never got around to it. I can determine how many 'transactions' my apps can handle, but it would be nice to know the maximum number of event cycles the machine can handle too. molto buono, grazie mille! -- "Be excellent to each other"
_______________________________________________ Libevent-users mailing list Libevent-users@monkey.org http://monkeymail.org/mailman/listinfo/libevent-users
