[R-pkgs] New package: gpuR

Charles Determan cdetermanjr at gmail.com
Mon Nov 23 14:10:09 CET 2015


R Users,

I am happy to inform you that my 'gpuR' package has just been accepted to
CRAN.

https://cran.r-project.org/web/packages/gpuR/index.html

The gpuR package is designed to provide simple to use functions for
leveraging GPU's for computing.  Although there are a couple existing
packages for GPU's in R most are specific to NVIDIA GPU's and have very
limited and specific functions.  The packaged is based on an OpenCL backend
in conjunction with the ViennaCL library (which is packaged within the
RViennaCL package).  This allows the user to use almost any GPU (Intel,
AMD, or NVIDIA).  It is my hope that these functions can be used to more
rapidly develop algorithms within R that can leverage GPUs.

The package is structured to use a few new S4 classes that retain the
object either on the host CPU or in GPU memory (thereby avoiding transfer
time).  I have included a minimal introductory vignette describing the
package further, providing a simple use case, and listing currently
available functions.

https://cran.r-project.org/web/packages/gpuR/vignettes/gpuR.pdf

You can view the github page here:

https://github.com/cdeterman/gpuR

which also contains a wiki to help with installation.  Although it must be
compiled, it is able to be installed on Linux, Mac OSX, and Windows
platforms.

I welcome any comments, issues (please submit on the github), and of course
additional contributions.

Regards,
Charles Determan

	[[alternative HTML version deleted]]



More information about the R-packages mailing list