[gmx-users] Re: intel or gcc for opteron

[David van der Spoel]

Myunggi Yi wrote:
> I have tested with 8, 16, 24 and 32 cpus.
> Only parallel with 8 and 16 cpus works, and the other cases (24 and
> 32) are even much slower than 8 cpus' case.
> 
> Any idea about this?
larger systems
check mailing list
use 16 cpus until gmx 4

> 
> 
> On Nov 24, 2007 11:42 AM, David van der Spoel <spoel at xray.bmc.uu.se> wrote:
>> Myunggi Yi wrote:
>>> Thank you for the answer.
>>>
>>> I compiled gmx with gcc and gnu onpempi.
>>>
>>> I have seen very good performance with 16 cpus (about 12,000 atoms).
>>>
>>> +++++++++++++++++++++++++++++++++++
>>>                (Mnbf/s)   (GFlops)   (ns/day)  (hour/ns)
>>> Performance:     50.828     19.920     14.648      1.638
>>> Total Scaling: 95% of max performance
>>> +++++++++++++++++++++++++++++++++++
>>>
>>> Now I just submit a job with 32 cpus (about 48,000 atoms).
>>> It looks very very slow.
>>>
>>>
>>> Would you recommend the optimal numer of cpus?
>> No. You have to test for yourself (as you are doing right now).
>> It is very dependent on simulation size, network etc.
>>
>>>
>>>
>>>
>>> On Nov 24, 2007 10:53 AM, David van der Spoel <spoel at xray.bmc.uu.se> wrote:
>>>> Myunggi Yi wrote:
>>>>> Sorry. Here is more information.
>>>>>
>>>>>
>>>>> GNU C/C++     3.4.6
>>>>> Intel C/C++   10.0.026
>>>>> and
>>>>> OpenMPI GNU   1.2.4
>>>>> OpenMPI Intel         1.2.4
>>>> Everything will probably be easier with the gcc compiler, and the
>>>> performance difference is negligible for gromacs.
>>>>
>>>>>
>>>>> On Nov 24, 2007 10:31 AM, Myunggi Yi <myunggi at gmail.com> wrote:
>>>>>> Dear gromacs users,
>>>>>>
>>>>>> I'm a beginner.
>>>>>> I'm trying to compile gromacs 3.3.1 or 3.3.2 on the following linux
>>>>>> cluster (centOS).
>>>>>> Dual, Dual-Core 2220 2.8 GHz Opterons, 2 GB RAM per core, 64-bit.
>>>>>> I will use 16 to 32 (64) cpus for the parallel simulations.
>>>>>>
>>>>>> I tested mpich2, but there is a problem with infiniband communications.
>>>>>>
>>>>>> I have intel and GNU compilers and openMPI.
>>>>>>
>>>>>> Which option will give me the best performance?
>>>>>>
>>>>>> Have a nice weekend.
>>>>>>
>>>>>>
>>>>>> --
>>>>>> Best wishes,
>>>>>>
>>>>>> MYUNGGI YI
>>>>>> ==================================
>>>>>> KLB 419
>>>>>> Institute of Molecular Biophysics
>>>>>> Florida State University
>>>>>> Tallahassee, FL 32306
>>>>>>
>>>>>> Office: (850) 645-1334
>>>>>> http://www.scs.fsu.edu/~myunggi
>>>>>>
>>>>>
>>>> --
>>>> David.
>>>> ________________________________________________________________________
>>>> David van der Spoel, PhD, Assoc. Prof., Molecular Biophysics group,
>>>> Dept. of Cell and Molecular Biology, Uppsala University.
>>>> Husargatan 3, Box 596,          75124 Uppsala, Sweden
>>>> phone:  46 18 471 4205          fax: 46 18 511 755
>>>> spoel at xray.bmc.uu.se    spoel at gromacs.org   http://folding.bmc.uu.se
>>>> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>>>> _______________________________________________
>>>> gmx-users mailing list    gmx-users at gromacs.org
>>>> http://www.gromacs.org/mailman/listinfo/gmx-users
>>>> Please search the archive at http://www.gromacs.org/search before posting!
>>>> Please don't post (un)subscribe requests to the list. Use the
>>>> www interface or send it to gmx-users-request at gromacs.org.
>>>> Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>>>>
>>>
>>>
>>
>> --
>> David.
>> ________________________________________________________________________
>> David van der Spoel, PhD, Assoc. Prof., Molecular Biophysics group,
>> Dept. of Cell and Molecular Biology, Uppsala University.
>> Husargatan 3, Box 596,          75124 Uppsala, Sweden
>> phone:  46 18 471 4205          fax: 46 18 511 755
>> spoel at xray.bmc.uu.se    spoel at gromacs.org   http://folding.bmc.uu.se
>> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>> _______________________________________________
>> gmx-users mailing list    gmx-users at gromacs.org
>> http://www.gromacs.org/mailman/listinfo/gmx-users
>> Please search the archive at http://www.gromacs.org/search before posting!
>> Please don't post (un)subscribe requests to the list. Use the
>> www interface or send it to gmx-users-request at gromacs.org.
>> Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>>
> 
> 
> 


-- 
David.
________________________________________________________________________
David van der Spoel, PhD, Assoc. Prof., Molecular Biophysics group,
Dept. of Cell and Molecular Biology, Uppsala University.
Husargatan 3, Box 596,  	75124 Uppsala, Sweden
phone:	46 18 471 4205		fax: 46 18 511 755
spoel at xray.bmc.uu.se	spoel at gromacs.org   http://folding.bmc.uu.se
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++