Hi Malkeet,
Did you take a look at the mdout file? It should have contained a very clear error message generated by the following code:
Line 2277 of src/pmemd/src/mdin_ctrl_dat.F90 of a fully patched Amber 16 + AmberTools17
#ifdef MPI
if ( imin .ne. 0 ) then
write(mdout, '(a)') 'CUDA (GPU): Minimization is NOT supported in &
¶llel on GPUs. Please use'
write(mdout, '(a)') ' the single GPU code for minimizations.'
inerr = 1
end if
#endif
If it did not then we need to take a closer look at why that error trap is not triggering or is missing from your version of the source code.
All the best
Ross
> On Mar 19, 2018, at 14:14, malkeet singh Bahia <msbahia17.gmail.com> wrote:
>
> one more thing, when I replaced 'min.in' file with 'heat.in' file
> (parameters for NVT) then pmemd.cuda.MPI is working on both of my GPUs!!!
>
>
> On Mon, 19 Mar 2018 at 20:17 malkeet singh Bahia <msbahia17.gmail.com>
> wrote:
>
>> Hello!
>>
>> I am getting the following problems while using
>>
>> mpirun -np 2
>> /private/gnss/AMBER_instal_files_malkeet/amber14/bin/pmemd.cuda.MPI -O -i
>> min.in -o min.out -p pro.top -c pro.crd -r min.rst -inf min.mdinfo >
>> resulted in "application called MPI_Abort(MPI_COMM_WORLD, 1) - process 0"
>> while
>> and /private/gnss/AMBER_instal_files_malkeet/amber14/bin/pmemd.cuda -O -i
>> min.in -o min.out -p pro.top -c pro.crd -r min.rst -inf min.mdinfo >
>> resulted in STOP PMEMD Terminated Abnormally!
>>
>> min.in file contains
>>
>> Minimize
>> & cntrl
>> imin=1,
>> ntx=1,
>> irest=0,
>> maxcyc=2000,
>> ncyc=1000,
>> ntpr=100,
>> ntwx=0,
>> cut=8.0,
>> /
>>
>>
>> Could you please suggest the reasons and solutions?
>>
>> Thanks!
>> MAlkeet
>>
>>
>>
>>
>> On Wed, 14 Mar 2018 at 22:07 David A Case <david.case.rutgers.edu> wrote:
>>
>>> On Wed, Mar 14, 2018, malkeet singh Bahia wrote:
>>>>
>>>> I didn't find any *.dif file. Below are make test results
>>>>
>>>> 62 file comparisons passed
>>>> 35 file comparisons failed
>>>> 0 tests experienced errors
>>>
>>> Right below this output, there should be pointers to where all the logs
>>> are: they will be under $AMBERHOME/logs, and labelled by the type of
>>> test and the date.
>>>
>>> Note that just saying you typed "make test" is not very helpful, since
>>> we would need to know what the most recent flags to configure were. The
>>> expected Amber workflow is this:
>>>
>>> ./configure gnu (say)
>>> make install #installs serial programs
>>> make test # tests serial programs
>>>
>>> ./configure -cuda gnu
>>> make install #installs cuda programs
>>> make test # tests cuda programs
>>>
>>> Be *sure* that serial cuda programs are working for you before
>>> proceeding to parallel cuda programs. And try to ensure that you really
>>> need parallel cuda capability before trying to install it. Unless you
>>> have good hardware, and a need to run a single simulation at the fastest
>>> possible speed, you will probably find that installing parallel cuda
>>> programs are not very helpful.
>>>
>>> ....dac
>>>
>>>
>>> _______________________________________________
>>> AMBER mailing list
>>> AMBER.ambermd.org
>>> http://lists.ambermd.org/mailman/listinfo/amber
>>>
>> --
>> Thanks & Warm Regards
>>
>>
>> Malkeet S. Bahia, PhD
>> Post Doctoral Fellow at the lab of Prof. Hanoch Senderowitz
>> PBC Indo-Israel Post-doc fellowship
>> Dept. of Chemistry, Building 211, Room no. 211
>> Bar Ilan University, Ramat Gan, 5290002, ISRAEL
>>
>> Mobile: +91-9815988501 <098159%2088501> (India, watts app only)
>> +972-584553434 <+972%2058-455-3434> (Israel number)
>>
> --
> Thanks & Warm Regards
>
>
> Malkeet S. Bahia, PhD
> Post Doctoral Fellow at the lab of Prof. Hanoch Senderowitz
> PBC Indo-Israel Post-doc fellowship
> Dept. of Chemistry, Building 211, Room no. 211
> Bar Ilan University, Ramat Gan, 5290002, ISRAEL
>
> Mobile: +91-9815988501 (India, watts app only)
> +972-584553434 (Israel number)
> _______________________________________________
> 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 20 2018 - 08:00:02 PDT
