de I have the early stages of a
package up on github: https://github.com/rundel/RcppGP
<https://github.com/rundel/RcppGP>. In particular the gpu_mat class has a
reasonably mature interface for moving data between armadillo and cuBLAS.
-Colin
-----
Colin Rundel
Assistant Professor of th
I do believe that arma uses the correct back solve when everything is called
correctly (I haven’t profiled the code recently to check what BLAS functions
are being called though). I get better performance from Rcpp than R when I
change the back solve line to be
RI = solve(trimatu(R_upptri),I);
I've run into the same issue and have been using the Rcpp's plugin framework to
work around it.
My R file defines a local_include plugin as follows:
registerPlugin("local_include",
function() {
list(env =
list(PKG_CXXFLAGS=paste0("-I",path.expand("~/Desktop/
I think that I've actually found the issue, the liblapack.so and libblas.so
installed in my /usr/lib/R/lib/ were actually from the revolution-mkl package
and not from r-base-core. Once I purged that package everything seems to be
working. Not sure what criteria the g++ linker was using to chose
> For certain operations R does _not_ go to lapack but uses its own. I can
> never remember if chol() was one of them -- but this suggests it. As I
> mentioned in my earlier email you probably really have to follow the chol()
> call all the way down (in the sources).
I've followed both calls to t
> reference blas (packages libblas3 and liblapack)
> atlas (package libatlas3-base or a tuned variant)
> open-blas (package libopenblas-base)
I had been playing around with update-alternatives to swap between the three,
but just in case this was causing issues I've used apt
link to the matrix file as it is somewhat
large.
test.R
Description: Binary data
util.cpp
Description: Binary data
https://www.dropbox.com/s/6i1c8xp2dsthq4y/tmp.dat
-Colin
-----
Colin Rundel
Postdoctoral Associate
Duke University, Department of Statistical Science
colin.run...@stat.duk
I am working on a project using RcppArmadillo and I've run into an issue with
the rgamma function in Rcpp. When calling rgamma the function pegs R's cpu
utilization and the process continues to churn forcing me to kill it. I've let
things run for around 5 mins with no end in sight.
I can replic
