Ok - this matches behavior I have seen on other kepler based systems (notably K80s) when running multiple GPUs. My current theory is that one of the more recent updates must have broken things. I'll investigate some more when I get time. In the meantime I suggest just running single GPU runs.
All the best
Ross
> On Mar 31, 2015, at 8:32 PM, Himanshu Joshi <himanshuphy87.gmail.com> wrote:
>
> Hello Ross,
> I ran the with pmemd.cuda and found it printing correctly.
> So the problem lies somwhere in pmemd.cuda.MPI.
>
> Thanks for the help.
>
> On Tue, Mar 31, 2015 at 11:36 PM, Ross Walker <ross.rosswalker.co.uk> wrote:
>
>> Hi Joshi,
>>
>> If you run with just pmemd.cuda instead of cuda.MPI do you see different
>> behavior?
>>
>> All the best
>> Ross
>>
>>> On Mar 31, 2015, at 10:58 AM, Himanshu Joshi <himanshuphy87.gmail.com>
>> wrote:
>>>
>>> Dear Friends,
>>> I am running pmemd.cuda.MPI (AMBER14) in Tesla K40s gpu cards for an
>>> explicit water DNA system with SPFP precision model. After running few
>>> steps sucessfully, I am getting **** for temperature and energy values as
>>> typed below followed by the md input file.
>>>
>>>
>>> NSTEP = 45000 TIME(PS) = 54230.000 TEMP(K) = 300.15 PRESS =
>>> 0.0
>>> Etot = -1102980.4479 EKtot = 216143.9219 EPtot =
>>> -1319124.3698
>>> BOND = 8297.0959 ANGLE = 17306.9385 DIHED =
>>> 22929.3885
>>> 1-4 NB = 8627.8031 1-4 EEL = -71177.7576 VDWAALS =
>>> 134764.0741
>>> EELEC = -1439871.9123 EHBOND = 0.0000 RESTRAINT =
>>> 0.0000
>>>
>> ------------------------------------------------------------------------------
>>>
>>> check COM velocity, temp: NaN NaN(Removed)
>>>
>>> NSTEP = 50000 TIME(PS) = 54240.000 TEMP(K) = NaN PRESS =
>>> 0.0
>>> Etot = NaN EKtot = NaN EPtot =
>>> **************
>>> BOND = 0.0000 ANGLE = 4770613.6638 DIHED =
>>> 0.0000
>>> 1-4 NB = ************** 1-4 EEL = ************** VDWAALS =
>>> **************
>>> EELEC = ************** EHBOND = 0.0000 RESTRAINT =
>>> 0.0000
>>>
>> ------------------------------------------------------------------------------
>>> Here is my md input file
>>>
>>> Production run with constant volume
>>> &cntrl
>>> ntx=6, irest=1,
>>> nmropt=0,
>>> ntrx=1, ntxo = 2,
>>> ntpr=5000, ntwx=5000, ntwr=10000,
>>> ntwv=0,
>>> ntf=2, ntb=1,
>>> cut=9.0, nsnb=10,
>>> nstlim=1000000, nscm=10000,
>>> t=0.0, dt=0.002,
>>> ntt=1, tautp=0.5
>>> ntp=0, pres0=1.0, taup=0.5, comp=44.6
>>> tempi=300.0, temp0=300.0,
>>> ipol=0, ntc=2, ioutfm = 1
>>> /
>>> &ewald
>>> /
>>>
>>> Though, I am running another NPT run using Langevin dynamics using
>>> anisotropic pressure coupling and MC barostate with same pmemd.cuda.MPI
>> and
>>> it is running fine.
>>>
>>> Is it something known problem or I am missing something, please comment.
>>> Thanks in advance for your valuable time.
>>>
>>>
>>>
>>>
>>>
>>> *With Regards,HIMANSHU JOSHI Graduate Scholar, Center for Condense Matter
>>> TheoryDepartment of Physics IISc.,Bangalore India 560012*
>>> _______________________________________________
>>> 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
>>
>
>
>
> --
>
>
>
> *With Regards,HIMANSHU JOSHI Graduate Scholar, Center for Condense Matter
> TheoryDepartment of Physics IISc.,Bangalore India 560012*
> _______________________________________________
> 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 Tue Mar 31 2015 - 21:00:08 PDT
