.Jason:
I already turned mkl off a while ago (for exact reason about downloading
size)
Hai
On Tue, Mar 22, 2016 at 9:36 PM, Jason Swails <jason.swails.gmail.com>
wrote:
> On Tue, Mar 22, 2016 at 9:30 PM, David A Case <david.case.rutgers.edu>
> wrote:
>
> > On Tue, Mar 22, 2016, Ross Walker wrote:
> >
> > > [rcw.comet-ln3 amber]$ cat /etc/redhat-release
> > > CentOS release 6.7 (Final)
> > >
> > > ./configure intel
> > > (I said yes to download the miniconda stuff)
> > > make
> > >
> > > (cd parmed && /home/rcw/amber/bin/amber.python setup.py install
> > --no-setuptools -f --prefix=/home/rcw/amber
> > --install-scripts=/home/rcw/amber/bin)
> > > OpenBLAS: pthread_creat error in blas_thread_init function. Error
> code:11
> >
> > Just curious: does the --skip-python flag at least allow one to get
> around
> > this problem?
> >
> > Second question: Does it make a difference if you load the MKL module?
> > (MKL provides its own version of BLAS, and it may be most people (like
> me)
> > that test intel builds also use MKL.)
> >
>
> This may have something to do with the fact that recently Anaconda (and
> Miniconda) have started providing a version of numpy and scipy linked to
> MKL. I wonder if this is causing problems. I'm trying to reproduce the
> error on my machine. I know how to disable MKL (and this should probably
> be done, anyway, since it increases the amount that needs to be downloaded
> by quite a bit, I think).
>
> If I can't repro it here, I'll see if I can on a CentOS 6.7 vagrant box.
> As an aside, what version of the Intel compilers are you using?
>
> --
> Jason M. Swails
> _______________________________________________
> AMBER-Developers mailing list
> AMBER-Developers.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber-developers
>
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Tue Mar 22 2016 - 19:00:08 PDT
