That's true. The gnu compiler is smooth here too. Maybe I should not worry about it. I am looking to get the Intel version working and do benchmark myself.
*******************************
Mohd Farid Ismail
Graduate Student
Dept. of Chemistry/Biochemistry
University of Oklahoma
Norman 73019
________________________________________
From: ET [sketchfoot.gmail.com]
Sent: Monday, June 24, 2013 10:18 PM
To: AMBER Mailing List
Subject: Re: [AMBER] Actual Intel Compiler Version Used
Hi,
I tried compiling AMBER on a dual opteron machine
with Intel_CplusplusCompilers_2013.2.146
and Intel_FortranCompilers_2013.2.146
Had no end of problems. Possibly because you need to set a different combo
of switches for AMD? Started looking into -xhost, etc, but never got to the
bottom of it, so went with GNU compilers in the end, which made the
installation trivial. As the machine was going to mainly run cuda, I
understood from the list that the intel code optimisation impact was
minimal to none. So haven't looked back.. Not sure if there is such a
dramatic code improvement in the CPU code either, but I'm sure someone who
knows better could inform you on the latter point. :)
br,
g
On 25 June 2013 01:28, Ismail, Mohd F. <farid.ou.edu> wrote:
> Dr. Roe,
>
> Here is my system specs:
> AMD Opteron Magny-cours 8-core
> CentOS 6.4
> Intel Compiler 11.1 073
>
> What I did:
> unpack AT1.5
> set AMBERHOME
> cd $AMBERHOME
> export PATH="AMBERHOME/bin:$PATH"
> apply patch (patch -p0 -N < bugfix.all)
> cd to $AMBERHOME/AmberTools/src
> ./configure intel
> make serial
>
> The compile will stop at compiling MTK+. I have included the error log
> and the config.h generated. The error log is truncated to reduce the size.
>
>
> *******************************
> Mohd Farid Ismail
> Graduate Student
> Dept. of Chemistry/Biochemistry
> University of Oklahoma
> Norman 73019
>
> ________________________________________
> From: Daniel Roe [daniel.r.roe.gmail.com]
> Sent: Sunday, June 23, 2013 10:10 PM
> To: AMBER Mailing List
> Subject: Re: [AMBER] Actual Intel Compiler Version Used
>
> Hi,
>
> On Sun, Jun 23, 2013 at 7:47 PM, Ismail, Mohd F. <farid.ou.edu> wrote:
> > Hi, can anyone let me know what was the actual version of intel compiler
> used for Amber11 and Amber12? And AmberTools 1.5 and AT12/13?
>
> Amber is tested with a variety of compiler suites and versions. I was
> just able to compile AmberTools 1.5 with Intel compilers version 11.1
> (20090511). Could you post your exact configuration, the flags you
> provided to configure, and the error you received?
>
> -Dan
>
> --
> -------------------------
> Daniel R. Roe, PhD
> Department of Medicinal Chemistry
> University of Utah
> 30 South 2000 East, Room 201
> Salt Lake City, UT 84112-5820
> http://home.chpc.utah.edu/~cheatham/
> (801) 587-9652
> (801) 585-9119 (Fax)
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Jun 24 2013 - 21:00:03 PDT
