[CP2K:3372] Re: Job stopped after 40 steps

[Jacopo Sgrignani]

Ciao Giulia
I've run a lot of QM/MM calculations on enzimes using QM/MM on SP6,
and it usually works fine.
However I don't use the cp2k version available with the module
command, but I use a version recompiled by myself.
Your system is really very big, but the QM region seems quite small
then, i think, you should run also with 64 procs.

Did you put this line in your submission script?
# @ resources   = ConsumableMemory(1400Mb)

Jacopo






2011/7/13 giulia palermo <giulia.p... at gmail.com>:
> Dear J.Sgrignani,
> Yes, I am running on cineca SP6.
> I will change the number of processors to 128 and 256 and I see what
> will happens.
>
> Thank you,
> Giulia
>
> On Jul 12, 10:50 pm, Jacopo Sgrignani <sgr... at gmail.com> wrote:
>> Hi Giulia
>> Maybe you can solve the memory problem increasing the number of processors.
>>
>> Are you running on the cineca SP6?
>>
>> Jacopo
>>
>> 2011/7/12, giulia palermo <giulia.p... at gmail.com>:
>>
>> > Dear Dr. Laino,
>> > probably CP2K is miscompiled in the machine that I were using. This
>> > conclusion arises from the subsequent arguments:
>>
>> > 1) Coping all the files to another machine the job has started, the
>> > geometry optimization has finished properly again, and we are
>> > currently running the MD simulation without problem for more than 100
>> > steps.
>>
>> > 2) As supposed by the system managers of the IBM sp6 machine, the most
>> > common error that I had is related to memory:
>>
>> > ATTENTION: 0031-408  64 tasks allocated by LoadLeveler, continuing...
>> > ERROR: 0031-161  EOF on socket connection with node sp0025
>>
>> > Even if my calculation are now running on an other machine, the
>> > problem on the IBM sp6 machine persists.
>> > How can be possible that other users of the sp6 are running cp2k,
>> > using the same submission file, without any problem?
>> > I will send an e-mail to the system manager ASAP.
>>
>> > Thank you very much for your quick response.
>> > Giulia Palermo
>>
>> > On Jul 12, 4:07 pm, Teodoro Laino <teodor... at gmail.com> wrote:
>> >> Hi Giulia,
>> >> it's not easy to say with these few info. Don't you get any error message
>> >> from the OS?
>> >> Neither locallog files?
>>
>> >> Teo
>>
>> >> On Jul 12, 2011, at 3:56 PM, giulia palermo wrote:
>>
>> >> > Dear all,
>> >> > I am a new cp2k user and I am working with a big enzyme/inhibitor system
>> >> > (~180000 atoms) in order to study the covalent inhibition of the enzyme.
>> >> > I am running the calculations on an IBM sp6 machine and I asked 64
>> >> > processors for 05.55 hours (in order to quickly enter in the queue).
>>
>> >> > I successfully did the first geometry optimization step for the QM part,
>> >> > maintaining fixed the Classical enzymatic environment.
>>
>> >> > Now I am trying to equilibrate the system in NVT ensemble at 300K,
>> >> > asking 1000 steps (with a timestep of 0.5 fs).
>> >> > Unfortunately, the calculation finishes after 40 steps even if the
>> >> > Energy values seem to be good.
>>
>> >> > I send an e-mail to the IBM-sp6 system managers. They reply that could
>> >> > be a memory problem.
>> >> > So, I tried several times to change the memory in the submission file
>> >> > (run.ll that you will find in attachment),
>> >> > but the problem persists.
>>
>> >> > Have you any suggestion about what could be the problem?
>> >> > You will also find in attachment the input file (md01.inp)
>>
>> >> > Thank you
>> >> > Giulia Palermo
>>
>> >> > --
>> >> > You received this message because you are subscribed to the Google
>> >> > Groups "cp2k" group.
>> >> > To post to this group, send email to cp... at googlegroups.com.
>> >> > To unsubscribe from this group, send email to
>> >> > cp2k+uns... at googlegroups.com.
>> >> > For more options, visit this group
>> >> > athttp://groups.google.com/group/cp2k?hl=en.
>> >> > <md01.inp><run.ll>
>>
>> > --
>> > You received this message because you are subscribed to the Google Groups
>> > "cp2k" group.
>> > To post to this group, send email to cp... at googlegroups.com.
>> > To unsubscribe from this group, send email to
>> > cp2k+uns... at googlegroups.com.
>> > For more options, visit this group at
>> >http://groups.google.com/group/cp2k?hl=en.
>
> --
> You received this message because you are subscribed to the Google Groups "cp2k" group.
> To post to this group, send email to cp... at googlegroups.com.
> To unsubscribe from this group, send email to cp2k+uns... at googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/cp2k?hl=en.
>
>