I am working on an R package that contains some CUDA code. As such, it
needs to use the 'nvcc' compiler. I then need to use some of the R header
files such as R.h. On a linux machine, I can handle this with my configure
script using autotools that will replace @R_INCL@ with AC_SUBST(R_INCL).
However, AFAIK, I cannot do this on a Windows machine. I am trying to
write the Makevars.win file to create the appropriate variables that I can
pass to the nvcc compiler. Here is an excerpt:
R_INCL=`$(shell ${R_HOME}/bin/R CMD config --cppflags)`
CU_INCL = -I../inst/include $(R_INCL)
%.o: %.cu $(cu_sources)
$(NVCC) $(CU_ARCH) $(CU_ARGS) $(CU_INCL) $< -c
However, I keep getting a truncated output from the --cppflags output. If
I run from the commandline R CMD config --cppflags I get:
-IC:/Users/cdeterman/Documents/R/R-3.3.2/include
-IC:/Users/cdeterman/Documents/R/R-3.3.2/include/x64
but when run in the build/install I get
-IC:/Users/CDETER~1/DOCUME~1/R/R-33~1.2/include
-IC:/Users/CDETER~1/DOCUME~1/R/R-33~1.2/include/x64
and the compliation fails saying that the
-IC:/Users/CDETER~1/DOCUME~1/R/R-33~1.2/include is not found whereas the
full path is found.
Any assistance is appreciated.
Regards,
Charles
[[alternative HTML version deleted]]
______________________________________________
R-package-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-package-devel
