Re: Optimize numeric.c mul_var() using the Karatsuba algorithm

Hi Joel, thanks for posting this.  Although I have only a cursory 
familiarity with fast multiplication algorithms, I'd like to try and 
give it a review.  To start with, can you help me understand the choice 
of this algorithm versus others like Toom?  If this looks correct on a 
closer view I'll propose it for inclusion. Along the way though I'd like 
to have it explicitly called out whether this is superior in general to 
other choices, better for more realistic use cases, simpler, clearer to 
license or something similar.  It would be nice for future dicussions to 
have some context around whether it would make sense to have conditions 
to choose other algorithms as well, or if this one is generally the best 
for what Postgres users are usually doing.

Continuing with code review in any case.  Interested to hear more.

Regards,

Aaron Altman