search

ingowald / cudaKDTree

206 stars 18 forks source link

cub::DeviceRadixSort or thrust::sort #27

Closed mcmingchang closed 1 month ago

mcmingchang commented 2 months ago

In my previous experience, it has been shown that if sort is called multiple times in a short period of time, the latency of the cube library is much shorter than that of the thrust library. You can choose between cube:: DeviceRadixSort or cube:: DeviceSegmentedRadixSort. I wonder if you are interested in trying them out, and I look forward to your reply.

ingowald commented 2 months ago

Yes, it's usually better; i pretty much changed all my repos over to cub, i just haven't gotten around to doing that for cukd, yet.

ingowald commented 2 months ago

BTW - if you want to submit a PR that changes this library from thrust to cub, i'd be happy to merge it :-)

mcmingchang commented 2 months ago

I am still researching how to call knn and obtain the first N nearest sequences