Re: [R-SIG-Mac] disabling threads when using Accelerate BLAS

2018-09-09 Thread Kasper Daniel Hansen
Thanks for the extremely helpful pointers on this.  Clearly I need to
remove --with-lapack in my configure step.

On Sun, Sep 9, 2018 at 4:12 AM Prof Brian Ripley 
wrote:

> On 08/09/2018 22:01, Kasper Daniel Hansen wrote:
> > For timing purposes, is it possible to control the number of threads /
> > cores when I have compiled R to use the BLAS in the Accelerate framework
> > (by building R using
> >./configure --with-blas="-framework Accelerate" --with-lapack
> > )
> >
> > ? (The alternative is of course to build R using the supplied
> > single-threaded BLAS)
>
> This is a question about Accelerate not R.  But the R-admin manual does say
>
> 'In recent versions of macOS, threading in Accelerate is controlled by
> ‘Grand Central Dispatch’ and is said not to need user control.'
>
> It has quite a few other caveats about Accelerate, not least that it
> uses an LAPACK which is missing many bug fixes, some serious.
>
> --
> Brian D. Ripley,  rip...@stats.ox.ac.uk
> Emeritus Professor of Applied Statistics, University of Oxford
>

[[alternative HTML version deleted]]

___
R-SIG-Mac mailing list
R-SIG-Mac@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-mac


Re: [R-SIG-Mac] disabling threads when using Accelerate BLAS

2018-09-09 Thread Prof Brian Ripley

On 08/09/2018 22:01, Kasper Daniel Hansen wrote:

For timing purposes, is it possible to control the number of threads /
cores when I have compiled R to use the BLAS in the Accelerate framework
(by building R using
   ./configure --with-blas="-framework Accelerate" --with-lapack
)

? (The alternative is of course to build R using the supplied
single-threaded BLAS)


This is a question about Accelerate not R.  But the R-admin manual does say

'In recent versions of macOS, threading in Accelerate is controlled by 
‘Grand Central Dispatch’ and is said not to need user control.'


It has quite a few other caveats about Accelerate, not least that it 
uses an LAPACK which is missing many bug fixes, some serious.


--
Brian D. Ripley,  rip...@stats.ox.ac.uk
Emeritus Professor of Applied Statistics, University of Oxford

___
R-SIG-Mac mailing list
R-SIG-Mac@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-mac