forum@abinit.org
Subject: The ABINIT Users Mailing List ( CLOSED )
List archive
- From: <zhouy6@rpi.edu>
- To: forum@abinit.org
- Subject: Re: Re: [abinit-forum] Abinit 5.4.4 compliation
- Date: Tue, 21 Oct 2008 16:47:57 +0200 (CEST)
Hi
Thank you for you fast reply.
I'm sure there is no other jobs running, so no confliction. I'll try gfortran
now to see what will happen.
btw, what's difference between gfortran and other compilier? and what does "
-static -march=native -O2" mean? I didn't find a comprehensive document in
/abinit/docs.
Thanks again!
Yu Zhou
On Tue, Oct 21, 2008 at 1:12 AM, Anglade Pierre-Matthieu <anglade@gmail.com>
wrote:
> Hi,
>
> That's just a little bit strange : your configure output tells that
> the usual optimization for gfortran have been used. this should
> provide a binary that is about 10% slower than the fastest possible
> one in general.
> 1) Are you sure no other task could have interfered with your benchmarking ?
> 2) Does the other tests show the same discrepancy between your build
> and the provided binary ? (for instance could you run and time "make
> tests_acc") ?
> 3) you could always try to enforce the compilation flags you want. For
> instance you can try something like : ./configure FC="gfotran -static
> -march=native -O2" .
> regards
>
> PMA
>
- [abinit-forum] Abinit 5.4.4 compliation, zhouy6, 10/21/2008
- Re: [abinit-forum] Abinit 5.4.4 compliation, Anglade Pierre-Matthieu, 10/21/2008
- Re: Re: [abinit-forum] Abinit 5.4.4 compliation, zhouy6, 10/21/2008
- Re: Re: [abinit-forum] Abinit 5.4.4 compliation, Anglade Pierre-Matthieu, 10/21/2008
- Re: Re: [abinit-forum] Abinit 5.4.4 compliation, zhouy6, 10/21/2008
- Re: [abinit-forum] Abinit 5.4.4 compliation, Yann Pouillon, 10/21/2008
- Re: [abinit-forum] Abinit 5.4.4 compliation, Anglade Pierre-Matthieu, 10/21/2008
Archive powered by MHonArc 2.6.15.