Thanks, Jeff! You are absolutely correct. We have OPENBLAS installed in our
environment that causes this.
One way to "fix" it is:
library(RhpcBLASctl)
blas_get_num_procs()
blas_set_num_threads(1)
stopifnot(blas_get_num_procs()==1)
On Sat, Apr 20, 2019 at 9:06 AM Jeff Newmiller
wrote:
> My
My guess would be that you are running with a non-CRAN distribution of R like
Anaconda or MRAN that has MKL enabled?
On April 19, 2019 10:25:57 AM PDT, Youyi Fong wrote:
>Hi, I am wondering why it is the case that in R 3.3.3, calling
>chngpt:chngptm uses only 1 core, but in later releases, e.g.
Hi, I am wondering why it is the case that in R 3.3.3, calling
chngpt:chngptm uses only 1 core, but in later releases, e.g. R 3.4.3, it
uses multiple cores on linux. The function chngpt:chngptm has a .Call to
invoke a C/C++ function that performs bootstrapping. No explicit parallel
computing
3 matches
Mail list logo