[cairo] New tessellator work and thoughts

[Behdad Esfahbod]

Wow! That was a heck of a lot of information for one email.

Thanks for looking into this Joonas.  Welcome aboard!

On Mon, 2006-11-20 at 21:07 -0500, M Joonas Pihlaja wrote:
> 
> Additionally, on every insert the skip list would call random() 
> twice, which turned out to be taking about 15%.  I suspect it was 
> using the time locking and unlocking the shared mutex used to 
> protect the global random state, or something.  In any case, 
> replacing the random() calls with a shift register based 
> generator seemed to help with that. 

My first thought reading the new tesselator code was that this can be
done using one random() call since you need two 16-bit random numbers,
so a good 32-bit one will do.  May save your approach a multiply too.


Anyway, keep up the awesome work!

-- 
behdad
http://behdad.org/

"Commandment Three says Do Not Kill, Amendment Two says Blood Will Spill"
        -- Dan Bern, "New American Language"