[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: OpenBLAS and performance
From: |
Pjotr Prins |
Subject: |
Re: OpenBLAS and performance |
Date: |
Wed, 20 Dec 2017 20:28:02 +0100 |
User-agent: |
Mutt/1.5.21 (2010-09-15) |
On Wed, Dec 20, 2017 at 07:15:16PM +0100, Ricardo Wurmus wrote:
> Is it really single-threaded? I remember having a couple of problems
> with OpenBLAS on our cluster when it is used with Numpy as both would
> spawn lots of threads. The solution was to limit OpenBLAS to at most
> two threads.
Looks like 1 on my system.
> > If I compile for a target it
> > makes a large difference.
>
> The FAQ document[1] says this:
>
> The environment variable which control the kernel selection is
> OPENBLAS_CORETYPE (see driver/others/dynamic.c) e.g. export
> OPENBLAS_CORETYPE=Haswell. And the function char*
> openblas_get_corename() returns the used target.
>
> [1]: https://github.com/xianyi/OpenBLAS/wiki/Faq
>
> Have you tried this and compared the performance?
About 10x difference on 24+ cores for matrix multiplication (my
version vs what comes with Guix).
I do think we need to default to a conservative openblas for general
use. Question is how we make it fly on dedicated hardware.
package python-numpy:openblas-haswellp
for the parallel version?
also for R and others. Problem is that we blow up the types of
packages.
Pj.
- OpenBLAS and performance, Pjotr Prins, 2017/12/19
- Re: OpenBLAS and performance, Ludovic Courtès, 2017/12/19
- Re: OpenBLAS and performance, Dave Love, 2017/12/20
- Re: OpenBLAS and performance, Dave Love, 2017/12/20
- Re: OpenBLAS and performance, Pjotr Prins, 2017/12/20
- Re: OpenBLAS and performance, Ricardo Wurmus, 2017/12/20
- Re: OpenBLAS and performance,
Pjotr Prins <=
- Re: OpenBLAS and performance, Ludovic Courtès, 2017/12/21
- Re: OpenBLAS and performance, Dave Love, 2017/12/22
- Re: OpenBLAS and performance, Ricardo Wurmus, 2017/12/22
- Re: OpenBLAS and performance, Pjotr Prins, 2017/12/20
- Re: OpenBLAS and performance, Ricardo Wurmus, 2017/12/20
- Re: OpenBLAS and performance, Pjotr Prins, 2017/12/21
- Re: OpenBLAS and performance, Eric Bavier, 2017/12/20
- Re: OpenBLAS and performance, Dave Love, 2017/12/21
- Re: OpenBLAS and performance, Dave Love, 2017/12/21
- Re: OpenBLAS and performance, Ricardo Wurmus, 2017/12/21