search

jermp / pthash

Fast and compact minimal perfect hash functions in C++.
MIT License
179 stars 25 forks source link

Parallel partitioning using cpp threads #47

Closed stefanfred closed 3 months ago

jermp commented 3 months ago

Hi @stefanfred, this one looks a lot cleaner! Thanks. How is the performance on many cores?

-Giulio

stefanfred commented 3 months ago

Hi @jermp,

I measured a speedup of 20 for the partitioning step using 64 hardware threads. This is also what I measured for ipsra which ist used by several competitors. But I expect that my code will not scale well for even larger machines.

jermp commented 3 months ago

That's fantastic. Better to reduce dependencies :)