Hello Everybody,
I have a few bugs that are giving me a problem, the general theme is
'mmpbsa' calculations.
My approaches are currently to either use pytraj or the dedicated python
script (MMPBSA.py) to do these calculations in parallel (i.e. using MPI).
Here are the failures:
1. MMPSA.py.MPI
Using this raises two errors for the import of "_distutils_hack" and also
"mpi4py". The former error is raised on the single threaded version and
works fine, but the latter one only appears in the MPI version. Checking
the mailing list archives, I also tested if mpi4py is installed on both my
current environment and on the amber environment using conda. Running conda
install mpi4py or amber.conda mpi4py gives identical results, the module is
installed. Trying to import it in python (i.e., running "python -c 'import
mpi4py') works fine in my conda environment but fails when run using
amber.python (i.e., using the amber python environment). This suggests to
me that there is some unforeseen disconnect between the two.
For completeness, here is the specific error message:
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
> File "/scistor/gt/dpo206/software/amber20/bin/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.
> Traceback (most recent call last):
> File "/scistor/gt/dpo206/software/amber20/bin/MMPBSA.py.MPI", line 53,
> in <module>
> from mpi4py import MPI
> ModuleNotFoundError: No module named 'mpi4py'
2. esander
Running these calculations in pytraj with esander is possible, and ideally
preferable to using the old mmpbsa.py script since then I can generate nice
visualizations or use the data generated downstream as part of a larger
workflow. While this is fine for GB and PME calculations, it fails for PB
(in line with the theme of MMPBSA). Specifically, pb init fails on the
second time the method is called in a python script, which is unfortunate
and non-functional for the intended application. Checking the mailing list
archives, I see a similar error for LIE but I cannot tell if it is from a
similar cause as it originates from the sander API. Here is the error
message:
PB Bomb in pb_init(): Allocation aborted 5014
> � 5014
> � 5014
> � 5014
> � 5014
> � 5014
> � 5014
> � 5014
> � 5014
> � 5014
> [some symbols did not survive the copy]
3. AmberTools (conda installed)
Lastly, MMPBSA is also included in the conda install of ambertools, when I
load them in a conda environment it becomes the default (determined using
'which MMPBSA.py'). This version gives a new error in regards to the
version number, I will include it here:
File "/scistor/gt/dpo206/.conda/envs/amber/bin/MMPBSA.py", line 4, in
> <module>
> __import__('pkg_resources').run_script('MMPBSA.py==16.0', 'MMPBSA.py')
> File
> "/scistor/gt/dpo206/.conda/envs/amber/lib/python3.9/site-packages/pkg_resources/__init__.py",
> line 662, in run_script
> self.require(requires)[0].run_script(script_name, ns)
> File
> "/scistor/gt/dpo206/.conda/envs/amber/lib/python3.9/site-packages/pkg_resources/__init__.py",
> line 1447, in run_script
> raise ResolutionError(
> pkg_resources.ResolutionError: Script 'scripts/MMPBSA.py' not found in
> metadata at None
The same error is raised for both the MPI and the serial version of MMPBSA
installed by conda.
So that is what I've got, three unfortunate stories for likely three
different people to respond to :^D I would really appreciate any help on
the matter.
Kindly,
David Poole
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Mar 22 2022 - 08:30:02 PDT
