Re: [AMBER-Developers] Fwd: Intel MPI in configure?

From: Daniel Roe <daniel.r.roe.gmail.com>
Date: Tue, 1 Apr 2014 15:08:12 -0600

My take on this is that the (now) long-standing behavior of Amber configure
has been to use mpicc/mpif90 as the designated compiler wrappers for MPI
compilation. The intel MPI 'mpiifort' etc. naming represents a deviation
from this, which is why I think it should be handled by a compiler flag,
much like the way '-crayxt5' deals with compiler wrappers on a cray system.
Also, just because a user has Intel MPI installed doesn't necessarily mean
they want to automatically use it for Amber. My vote is to make the Intel
MPI logic separate from the 'intel' compiler flag, and for it to require a
specific flag to configure.

-Dan


On Tue, Apr 1, 2014 at 2:57 PM, Needham, Perri <pneedham.ucsd.edu> wrote:

> Sorry this should have been emailed to the full developers list, not just
> Jason.
>
> What is the final decision on this? I think as few additionally flags as
> possible is always a bonus but if the general consensus is to do away with
> the auto intel mpi check then that's easily changed. Saying that, adding an
> additional error checking line to ensure the user's license is set up
> correctly would also work.
>
> Cheers,
> Perri
>
> Begin forwarded message:
>
> From: Jason Swails <jason.swails.gmail.com<mailto:jason.swails.gmail.com>>
> Subject: Re: [AMBER-Developers] Intel MPI in configure?
> Date: 1 April 2014 13:34:49 GMT-7
> To: "Needham, Perri" <pneedham.ucsd.edu<mailto:pneedham.ucsd.edu>>
>
>
>
> On Apr 1, 2014, at 3:48 PM, "Needham, Perri" <pneedham.ucsd.edu<mailto:
> pneedham.ucsd.edu>> wrote:
>
> Hi,
>
> I'm about to push the modified mpi check but what should I do in the
> scenario that mpiifort is found by 'which' but there is no license file? At
> the moment configure fails with the errors below. Options are to keep it
> that way or to use mpif90 instead?
>
> Thoughts?
>
> This is an even stronger argument for requiring a -intelmpi flag or just
> requiring users to symlink mpiifort to mpif90 and mpiicc to mpicc in my
> opinion...
>
>
> Cheers,
> Perri
>
> Checking for updates...
> Checking for available patches online. This may take a few seconds...
>
> Available AmberTools 14 patches:
>
> No patches available
>
> Available Amber 14 patches:
>
> No patches available
> Searching for python2... Found python2.6: /usr/bin/python2.6
> Detected icc version Error: A license for CCompL is not available
> (-76,61026,2)
> Detected ifort version Error: A license for FCompL is not available
> (-76,61026,2)
> ./configure2: line 940: [: Error: A license for CCompL is not available
> (-76,61026,2): integer expression expected
> ./configure2: line 958: [: Error: A license for CCompL is not available
> (-76,61026,2): integer expression expected
> ./configure2: line 961: [: Error: A license for FCompL is not available
> (-76,61026,2): integer expression expected
> ./configure2: line 993: [: Error: A license for CCompL is not available
> (-76,61026,2): integer expression expected
> ./configure2: line 1001: [: Error: A license for FCompL is not available
> (-76,61026,2): integer expression expected
> Detected mpiifort version 4.1
> Detected mpiicc version 4.1
> ./configure2: line 1032: [: too many arguments
> ./configure2: line 1073: [: Error: A license for CCompL is not available
> (-76,61026,2): integer expression expected
> ./configure2: line 1082: [: Error: A license for FCompL is not available
> (-76,61026,2): integer expression expected
>
> Testing the icc compiler:
> icc -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE -o testp testp.c
>
> Error: A license for CCompL is not available (-76,61026,2).
>
> License file(s) used were (in this order):
> 1. Trusted Storage
> ** 2. /server-home3/netbin/intel2013/composer_xe_2013_sp1.0.080/licenses
> ** 3. /opt/intel/licenses
> ** 4. /server-home1/perri//intel/licenses
> ** 5. /server-home3/netbin/intel2013/composer_xe_2013_sp1.0.080/Licenses
> ** 6. /server-home1/perri/intel/licenses
> ** 7. /Users/Shared/Library/Application Support/Intel/Licenses
> ** 8.
> /server-home3/netbin/intel2013/composer_xe_2013_sp1.0.080/bin/intel64/*.lic
>
> Please visit http://software.intel.com/sites/support/ if you require
> technical assistance.
>
> icc: error #10052: could not checkout FLEXlm license
> ./configure2: line 1651: ./testp: No such file or directory
> Error: Unable to compile a C program using icc -D_FILE_OFFSET_BITS=64
> -D_LARGEFILE_SOURCE
> Please check your compiler settings or configure flags.
> Configure failed due to the errors above!
>
>
>
> On 1 Apr 2014, at 12:12, Jason Swails <jason.swails.gmail.com<mailto:
> jason.swails.gmail.com>> wrote:
>
> On Tue, Apr 1, 2014 at 2:31 PM, Gerald Monard <
> Gerald.Monard.univ-lorraine.fr<mailto:Gerald.Monard.univ-lorraine.fr>>
> wrote:
>
> Hi,
>
> Having mpiifort/mpiicc is very specific I guess.
> On my side, if I use "configure_mpich2/openmpi intel" from
> AmberTools/src to get a working MPI version with the Intel compiler, I
> end up with mpif90/mpicc only. So there are different ways of having
> Intel + MPI...
> As Dave said, it has "always" been required that the user provides a
> working version of mpif90/mpicc. What should this case be so specific?
> Is it now standard to have mpiifort for Intel + MPI? In that case,
> should configure_mpich/openmpi be modified to automagically link the
> correct binaries?
>
>
> Intel MPI is a special MPI developed and released by Intel. Unlike the
> more typical MPI implementations (mpich, OpenMPI, mvapich, etc.), Intel MPI
> names their compiler wrappers "mpiifort" and "mpiicc".
>
> Supposedly Intel MPI scales better, but I've never used it myself...
>
> All the best,
> Jason
>
> --
> Jason M. Swails
> BioMaPS,
> Rutgers University
> Postdoctoral Researcher
> _______________________________________________
> AMBER-Developers mailing list
> AMBER-Developers.ambermd.org<mailto: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
>



-- 
-------------------------
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) 585-6208 (Fax)
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Tue Apr 01 2014 - 14:30:02 PDT
Custom Search