On Wednesday 31 March 2010 at 12:53, Andrew Whitworth wrote:

> Do you have any benchmark numbers available? Any idea how this branch
> compares performance-wise with trunk? The goal here was to improve
> performance, correct?

It still needs work to improve performance; it's about 10% slower.  In a quick 
benchmark, we spend too much time doing memcmp(), and it doesn't detect 
duplicate STRING constants with complete accuracy yet.

I have some ideas, but I don't have time to type them up for a bit.

-- c
_______________________________________________
http://lists.parrot.org/mailman/listinfo/parrot-dev

Reply via email to