Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

A bloom filter needs to multiple loads from different memory locations for each single lookup. (7, in the example 1.2GB filter.) But unlike, say, with a tree, it knows all the addresses after computing the hashes, without having to wait for results from the previous loads. So it can start all of them in parallel.


Yeah, sorry. I was thinking of a single hash function bloom filter, which is of course the exception. My bad.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: