Xor Filters: Faster and Smaller Than Bloom Filtershttps://lemire.me/blog/2019/12/19/xor-filters-faster-and-smaller-than-bloom-filters/ …
CF nerd here, thanks for the paper I'll make sure to read it carefully. In the meantime, two nitpicks: 1. Cuckoo's aren't that hard, I have a <250 LOC implementation that doesn't cut corners.https://github.com/kristoff-it/zig-cuckoofilter …
-
-
Your implementation is great and quite simple. I don't think I ever wrote that this wasn't possible. I will write however that your implementation of an xor filter would be simpler. At least the query part...
-
Might do that in the near future, if
@ArielShtul doesn't beat me on that :)
Kraj razgovora
Novi razgovor -
-
-
2. "Wasting so much space seems problematic, especially since it does not improve the false-positive probability." The pow2 thing is annoying indeed, but this statement seems wrong. A filter filled up to 50% will waste space, but the err% will be lower than filling the filter.
-
If you want to have tight memory usage in cuckoo filters, you need to pack your arrays, do away with the power-of-two restriction. It is easy enough to fix... and there are follow-up papers like the recent one on vaccuum filters that are related. cc
@tlipcon - Još 1 odgovor
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.