Hi
I am still working on this with hpc. I am 100% sure it has to do with
their mpi installation. They are checking things now.
adrian
On 9/19/22 1:50 PM, amirhossein taghavi via AMBER wrote:
> [External Email]
>
> Hello,
>
> I got the following responses from HPC admin and it seems a separate python
> installation is not an option on HPC. That would be great if any of the
> Amber team could help with HPC's questions.
>
> Thanks,
> Amir
>
> 1) they should correct the cmake to work with python + miniconda, no
> separate python installs which mess up our python for general users.
>
> 2) Which versions of compiler+openmpi and python are officially supported
> by the Amber team? The help says the latest is intel17 which is quite old,
> is that true?
>
>
> On Sun, Sep 18, 2022 at 7:43 PM Hai Nguyen <nhai.qn.gmail.com> wrote:
>
>> Are you able to install mpi4py via
>>
>> conda install mpi4py -c conda-forge
>>
>> Hai
>>
>> On Sun, Sep 18, 2022 at 2:44 PM amirhossein taghavi via AMBER <
>> amber.ambermd.org> wrote:
>>
>>> Hello Amber Developers,
>>>
>>> We have a problem with building the MMPBSA.py.MPI on HPC following the
>>> recommended installation by Amber manual. We have also tested the
>>> suggestion by Prof. Case on copying the .py to .py.MPI and it didn't work.
>>> The summary of the problem is provided as follows and any suggestions to
>>> solve this problem is very much appreciated.
>>>
>>> Best,
>>> Amir
>>>
>>> Executive summary:
>>> ==================
>>>
>>> In a "standard" cmake MPI install
>>>
>>> ll /apps/intel/2019.1.144/openmpi/4.0.1/amber/20/bin/MM*
>>> ls: cannot access /apps/intel/2019.1.144/openmpi/4.0.1/amber/20/bin/MM*:
>>> No
>>> such file or directory
>>>
>>> using miniconda, which is supposed to do the python part, the resulting
>>> /bin
>>> has no
>>>
>>> MMPBSA.py and MMPBSA.py.MPI
>>>
>>> and is non-python-fuctional.
>>>
>>> In a cmake non-MPI install, e.g.
>>>
>>> ll /apps/gcc/7.3.0/amber/20/bin/MM*
>>> -rwxrwxr-x 1 dimi amber 4129 Jun 4 2020
>>> /apps/gcc/7.3.0/amber/20/bin/MMPBSA.py
>>>
>>> python is OK, MMPBSA.py is there and works, no hacking involved.
>>>
>>> The idea to just cp MMPBSA.py to MMPBSA.py.MPI fails, not surprisingly
>>> for a
>>> serial install, on the lines
>>>
>>> if sys.argv[0].endswith('MPI'):
>>> try:
>>> from mpi4py import MPI
>>>
>>> MMPBSA.py.MPI
>>> Traceback (most recent call last):
>>> File "./MMPBSA.py.MPI", line 53, in <module>
>>> from mpi4py import MPI
>>> ModuleNotFoundError: No module named 'mpi4py'
>>>
>>> During handling of the above exception, another exception occurred:
>>>
>>> Traceback (most recent call last):
>>> File "./MMPBSA.py.MPI", line 55, in <module>
>>> raise MMPBSA_Error('Could not import mpi4py package! Use serial
>>> version
>>> '
>>> MMPBSA_mods.exceptions.MMPBSA_Error: Could not import mpi4py package! Use
>>> serial version or install mpi4py.
>>>
>>>
>>> Summary: cmake serial and OMP installs work for python, but not MPI
>>> installs.
>>> _______________________________________________
>>> AMBER mailing list
>>> AMBER.ambermd.org
>>> https://nam10.safelinks.protection.outlook.com/?url=http%3A%2F%2Flists.ambermd.org%2Fmailman%2Flistinfo%2Famber&data=05%7C01%7Croitberg%40ufl.edu%7C43cffb18d1154a3a7ba308da9a67a5d9%7C0d4da0f84a314d76ace60a62331e1b84%7C0%7C0%7C637992067197691750%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=fAxc76PX8zHXBZXP%2FJBgiCoTn4SSWDQjniKoarq6Iqs%3D&reserved=0
>>>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> https://nam10.safelinks.protection.outlook.com/?url=http%3A%2F%2Flists.ambermd.org%2Fmailman%2Flistinfo%2Famber&data=05%7C01%7Croitberg%40ufl.edu%7C43cffb18d1154a3a7ba308da9a67a5d9%7C0d4da0f84a314d76ace60a62331e1b84%7C0%7C0%7C637992067197691750%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=fAxc76PX8zHXBZXP%2FJBgiCoTn4SSWDQjniKoarq6Iqs%3D&reserved=0
--
Dr. Adrian E. Roitberg (he/him/el)
V.T. and Louise Jackson Professor in Chemistry
Department of Chemistry
University of Florida
roitberg.ufl.edu
352-392-6972
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Sep 19 2022 - 11:30:03 PDT
