You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Code that runs on CUDA devices needs to be compiled specifically for the user's device. Compilation is done by evoking the function 'install.m' from Matlab. 'install.m' will compile two mex files
- fnearneigh_gpu.mexa64
- rangesearch_all_gpu.mexa64
from the corresponding *.cpp files using the library libgpuKnnLibrary.a. libgpuKnnLibrary.a contains uncompiled CUDA code for all devices to date (as of May 2013).
(Make sure, both mex files are within the TRENTOOL folder before starting any calculations.)