Christian Seiler wrote:
>
> yes, uname -a: Linux cs-laptop 2.6.24-19-generic
> java 1.6.0_06-b02
>
> Best out of 10.times {puts Benchmark.measure {10000.times {Profile.find
> :first}}.total}
>
> Profile is a simple AR model (a few properties).
>
> I just noticed that 10.times {puts Benchmark.measure {10000.times
> {Profile.new }}.total} is slower than with MRI (something like 0.79 vs
> 0.89), too. I can imagine it's hard for JRuby/AR-JDBC to compete with
> MRI/MySQL/Unix sockets, but the seond benchmark is just about
> instantiating objects.
>
I just found out that jirb turns on ObjectSpace, so script/console it not
suited for benchmarking. Running without ObjectSpace the "Profile.new" test
performs blazingly fast (twice as fast as MRI!). "Profile.find" is about
20-30% slower, which I think is still pretty good. Using the pure Ruby
driver (MySQL) would be interesting too.
http://jira.codehaus.org/browse/JRUBY-2184
--
View this message in context:
http://www.nabble.com/JRuby-1.1.3-by-end-of-week...Nominate-problems-here...-tp18448971p18502776.html
Sent from the JRuby - Dev mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe from this list, please visit:
http://xircles.codehaus.org/manage_email
