On 26 May 2024 at 13:31, Kurt Hornik wrote:
| >>>>> Dirk Eddelbuettel writes:
|
| > Kurt,
|
| > Could you do me a favour and run on that clang18-using machine in question
| > the following one-liner (provided your session has access to a .libPaths()
| > including Rcpp) and, in the case of success, the resulting function?
|
| I can:
|
| R> Rcpp::cppFunction("int ompconfigtest() { return omp_get_num_threads(); }",
includes="#include <omp.h>", plugin="openmp")
| Warning in Rcpp::cppFunction("int ompconfigtest() { return
omp_get_num_threads(); }", :
| partial argument match of 'plugin' to 'plugins'
| R> ompconfigtest()
| [1] 1
Hmpf.
| The system has OpenMP, but R was configured not to use it.
Again, it would be lovely if we could _query_ that.
| In general, packages should leave the decision to use OpenMP to the
| *user*, who can use their own Makevars files to override the R system
| Makevars SHLIB_OPENMP_* settings as desired.
Didn't help here, did it? Or maybe Rcpp was too eager with
.plugins[["openmp"]] <- function() {
list(env = list(PKG_CXXFLAGS="-fopenmp",
PKG_LIBS="-fopenmp"))
}
and we'd need something like this (untested)
.plugins[["openmp"]] <- function() {
list(env = list(PKG_CXXFLAGS=Sys.getenv("SHLIB_OPENMP_CXXFLAGS, ""),
PKG_LIBS=Sys.getenv("SHLIB_OPENMP_CXXFLAGS, "")))
}
Actually, does not work it does come into the R session. Only goes to `make`.
So I got nothing. Sorry.
Dirk
--
dirk.eddelbuettel.com | @eddelbuettel | e...@debian.org
______________________________________________
R-package-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-package-devel
