RE: Is something wrong with build 2.0?

Did you follow the instructions at
http://www.strongtalk.org/benchmarking.html?  The most common benchmarking
mistake is putting the loop in a doIt, and/or not running it multiple times
so that the compiled version gets used, since the current implementation
doesn't have on-stack replacement (it can't switch an activation from the
interpreted to the compiled form while the activation is running).   You are
probably seeing interpreted performance.
-Dave

> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]]On Behalf Of David Simmons
> Sent: Monday, March 17, 2008 2:02 PM
> To: Strongtalk-general
> Subject: Is something wrong with build 2.0?
>
>
>
> I was just showing it off after talking about its performance. I was
> really surprised to find it was slow...
>
> On a 2.6GHz processer it does 71M loops per second - empty loop.
> It calls an empty function or one which returns a member at about 250M/
> sec.
>
> That is between 1/2 and 1/4 the speed of the dynamic language jits I
> am familiar with.
>
> ?
> David Simmons
> >


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "Strongtalk-general" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [email protected]
For more options, visit this group at http://groups.google.com/group/strongtalk-general?hl=en
-~----------~----~----~----~------~----~------~--~---