Hi Avraham, IIUC, when you install / compile RcppArmadillo, it sets the BLAS, LAPACK linker arguments based on the results of
R CMD config BLAS_LIBS R CMD config LAPACK_LIBS So if you've compiled R using OpenBLAS, RcppArmadillo should also be using OpenBLAS. You can check R CMD config BLAS_LIBS, as well as the actual compilation steps invoked when installing RcppArmadillo from source, to see how exactly it is linking to BLAS / LAPACK. But (I think) Rblas.dll is still just the packaged version of BLAS distributed with R; I don't think it even gets generated if you specify a system BLAS library explicitly. But (assuming you're on Windows) I'm not sure if R does something Windows-specific here. Cheers, Kevin On Thu, Feb 6, 2014 at 3:08 PM, Avraham Adler <avraham.ad...@gmail.com> wrote: > If I have compiled my installation of R linking to OpenBLAS so that > Rblas.dll is fast and not reference, if I install RcppArmadillo from source, > will it take advantage of that Fast Blas? > > Thank you, > > Avraham > > _______________________________________________ > Rcpp-devel mailing list > Rcpp-devel@lists.r-forge.r-project.org > https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/rcpp-devel _______________________________________________ Rcpp-devel mailing list Rcpp-devel@lists.r-forge.r-project.org https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/rcpp-devel