2
$\begingroup$

Suggestions of widely used and fast GPU libraries that calculate the Euclidean distance of a vector from a matrix of vectors (of the same dimension) ?

Preferably one that works on NVIDIA GPUs.

Need it as a part of comparing solutions that evaluate Radial Basis Function Neural Networks.

$\endgroup$
  • $\begingroup$ What language you prefer? I estimate writing this function should not take much effort (in c language). $\endgroup$ – William Jul 1 '16 at 20:33
  • $\begingroup$ @William I'll be running these libraries on a small cluster / HPC workstation, so I require libraries with high performance. $\endgroup$ – kesari Jul 2 '16 at 10:09
1
$\begingroup$

If you are using python, use CUDA tool-kit with cudamat library. You will be able to perform matrix operations on GPU using cudamat.

$\endgroup$
  • $\begingroup$ Can I time these functions to the precision that cudaEvents can ? $\endgroup$ – kesari Jul 2 '16 at 13:14
  • $\begingroup$ I am not sure about the precision bt in python you can definitely time these functions. $\endgroup$ – Riken Mehta Jul 2 '16 at 13:24

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.