Re:Advanced Data Structures MIT
by gritzko on Thursday February 09, 2012 @09:35AM
(#3571238)
Aha... and for 2^32 I'll have 5-fold speedup. 6-fold for 2^64 entries, obviously.
Frankly, those small-linear-factor improvements tend to be overshadowed by other factors (i.e. whether you hit the cache). I mean, they make sense as an optimization once everything is in place and works correctly.
But, designing something ground up to have a small-constant-factor speedup... that reminds me one physics trick. Namely, that by standing next to an infinite wall and by turning a laser pointer fast you make the spot on the wall move faster than c.
[ Parent
| on HN ]
There are no comments on this comment yet.