Re: [Resin-interest] My own bemchmark
Scott Ferguson wrote: > > it will be good to know if Quercus is calling the database inefficiently > (very possible) or if > we're stuck with a slow driver. > Hi Scott, How can I answer that point of yours..? What should I look for to answer this question..? Thanks Stephane -- View this message in context: http://www.nabble.com/My-own-bemchmark-tp18797858p18814276.html Sent from the Resin mailing list archive at Nabble.com. ___ resin-interest mailing list resin-interest@caucho.com http://maillist.caucho.com/mailman/listinfo/resin-interest
Re: [Resin-interest] My own bemchmark
On Aug 4, 2008, at 6:02 AM, Stephane Eybert wrote: > > Hi Scott, > > I shall run my code into a profiler with xdebug and pinpoint the > functions > that take most time, and let you know about my results. That would be great. Since it sounds like you're database bound, and we already know the Mysql JDBC driver is pretty inefficient, it will be good to know if Quercus is calling the database inefficiently (very possible) or if we're stuck with a slow driver. -- Scott > > > Cheers > > Stephane > -- > View this message in context: > http://www.nabble.com/My-own-bemchmark-tp18797858p18810855.html > Sent from the Resin mailing list archive at Nabble.com. > > > > ___ > resin-interest mailing list > resin-interest@caucho.com > http://maillist.caucho.com/mailman/listinfo/resin-interest ___ resin-interest mailing list resin-interest@caucho.com http://maillist.caucho.com/mailman/listinfo/resin-interest
Re: [Resin-interest] My own bemchmark
And here is an example using the sql layer. In case it is of any interest. http://www.nabble.com/file/p18799844/example.zip example.zip -- View this message in context: http://www.nabble.com/My-own-bemchmark-tp18797858p18799844.html Sent from the Resin mailing list archive at Nabble.com. ___ resin-interest mailing list resin-interest@caucho.com http://maillist.caucho.com/mailman/listinfo/resin-interest
Re: [Resin-interest] My own bemchmark
Hi Scott, Here is the php source code for the database layer. http://www.nabble.com/file/p18799823/sql.tar.gz sql.tar.gz Stephane -- View this message in context: http://www.nabble.com/My-own-bemchmark-tp18797858p18799823.html Sent from the Resin mailing list archive at Nabble.com. ___ resin-interest mailing list resin-interest@caucho.com http://maillist.caucho.com/mailman/listinfo/resin-interest
Re: [Resin-interest] My own bemchmark
Unfortunately I'm not in a position where I can be able to release the code. -- View this message in context: http://www.nabble.com/My-own-bemchmark-tp18797858p18799409.html Sent from the Resin mailing list archive at Nabble.com. ___ resin-interest mailing list resin-interest@caucho.com http://maillist.caucho.com/mailman/listinfo/resin-interest
Re: [Resin-interest] My own bemchmark
Sorry for that. Probably I should not have called it a benchmark. What I did is call a url served by PHP and by Resin and compare the times for the two responses. This request refreshes an internal cache of a website content, so it's pretty heavy, but it's only lots of database access. -- View this message in context: http://www.nabble.com/My-own-bemchmark-tp18797858p18799403.html Sent from the Resin mailing list archive at Nabble.com. ___ resin-interest mailing list resin-interest@caucho.com http://maillist.caucho.com/mailman/listinfo/resin-interest
Re: [Resin-interest] My own bemchmark
On Aug 3, 2008, at 4:32 AM, Stephane Eybert wrote: > > Dear all, > > I ran a request against a PHP page served by Resin and Quercus on a > Linux > Debian Resin 3.1.6 server. > > The response came after 1mn23s. > > I ran the same request against PHP Version 4.4.4-8 and the response > came > after 53s. > > I was not using any php caching system like eaccelerator. > > The php source code files were actually the same files in both tests. > > When using the php cache system eaccelerator the response came after > 24s. What does your benchmark do? You can't report a benchmark result with mystery code, and a page that takes 1 minute is pretty mysterious. > I wonder how can one achieve the stated benchmark of "4 times > faster" with > Resin. Our benchmarks are against well-known, complete applications: Drupal, Mediawiki, Wordpress. They've been reproduced independently: http://www.workhabit.org/resin-backed-php-drives-4x-performance-improvements-drupal Now, it's entirely possible that your code is running some feature that's slower on Quercus, and that would be useful information for us. But mystery code doesn't help anyone. -- Scott > > > If any one has any tip on this... > > Cheers > > -- > View this message in context: > http://www.nabble.com/My-own-bemchmark-tp18797858p18797858.html > Sent from the Resin mailing list archive at Nabble.com. > > > > ___ > resin-interest mailing list > resin-interest@caucho.com > http://maillist.caucho.com/mailman/listinfo/resin-interest ___ resin-interest mailing list resin-interest@caucho.com http://maillist.caucho.com/mailman/listinfo/resin-interest