Salut Laurent,

I totally agree, recursive fibonacci is not the most realwold routine, but still I think it yields some significant data.

The benchmark will probably get some sequel, I will try then with an AOT compilation, also maybe 0.7 will be out by that time, looking forward to it!

thanks for the tips,

L-P

2010/9/25 Laurent Sansonetti <lsansonetti@apple.com>
Bonjour Louis-Philippe :)

That's a cool micro-benchmark. I'm not sure if recursive arithmetic algorithms really represent the real world, but it's cool anyways.

If you remove the startup times from the final number you should see better results for MacRuby. But I agree that the current MacRuby startup time, which is about 0.1s, is bad, and should be fixed (if you file a ticket that would be great).

I recommend to give it a try with MacRuby trunk (you can grab a nightly build) and set up the VM_OPT_LEVEL environment variable to 3.

$ VM_OPT_LEVEL=3 macruby test.rb

You should get good results this way. This will enable optimized codegen and primitives inlining.

Laurent

On Sep 24, 2010, at 5:06 AM, Louis-Philippe wrote:

Hi All,

I had fun building this benchmark which involves MacRuby:

I think there might be some points of interest to MacRuby developers in there.

L-P
_______________________________________________
MacRuby-devel mailing list
MacRuby-devel@lists.macosforge.org
http://lists.macosforge.org/mailman/listinfo.cgi/macruby-devel


_______________________________________________
MacRuby-devel mailing list
MacRuby-devel@lists.macosforge.org
http://lists.macosforge.org/mailman/listinfo.cgi/macruby-devel