Hi everyone,
which approach should I try to use for official CRAN submissions of R
packages using OpenCL and GPUs, if none of the available solutions would
pass the CRAN checks.
Currently I see two possibilities of building an R package deploying
OpenCL with a GPU, however neither of them would pass the CRAN checks
directly. The first solution would be to use the approach of the
"OpenCL" package from CRAN
(https://cran.r-project.org/web/packages/OpenCL/index.html) relying on
the user to set local variables (OCL = ".../SDK/...", and so forth). The
second approach would be to use something like cmake, to circumvent the
need for the user to take care of certain requirements prior to
installation, as is done in
https://github.com/jmaerte/cmake-rcpp-template. The cmake approach poses
the problem, that the paths generated are too long and thus fails the
CRAN check and a building of the package is not possible.
Best regards
Quirin
______________________________________________
R-package-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-package-devel
