On Sat, Apr 19, 2008 at 1:48 PM, karl <karl.ramberg at comhem.se> wrote:
>
>  Is that 20/30 % overall speedup ? If so, it is great!
>
>  Karl
>

I'm working on one function at time as they step up in the profiling results.

Cannot say what the overall improvements exactly is; in the patched code
execution time is dominated by primitiveWarpBits instead of the
interpret function.
As those two functions  are largely the most cpu intensive ones I think that
the overall speedup is quite sensible.


Anybody already working on performances in squeak ?; I'd really
like to help in this part of the platform !


Riccardo