Can someone tell me if there are CUDA libraries and sources available for things like maps, sets, unordered_maps etc. on FERMI?
I’m interested in a device side equivalent of the C++ TR1 unordered_map (hash table) as well as other STL containers. Has anyone made performance comparisons of HTs operations on FERMI hardware versus on a Quad Core (preferrably the I7) host machine?
Also, has anyone taken the approach of using FERMI device to hash keys transferred from host, and then pass key hashes back to host for use in host side HT operations?
Thanks in advance.
Before FERMI came out someone mentioned Thrust but it seems that it only implements Vector. http://gpgpu.org/2009/05/31/thrust