Hi Stephan, sorry for the late reply.
Thanks for the clarification about the Amber/AmberTools versioning interplay, that helps clarify things. I’ll just trust the download page then and ignore the numbers in the source.
I did not apply any pytraj patches and the .patches folders are all empty in the installation. Is there something I’m missing from my installation procedure?
Yes, I know I could just apply the patch you gave me (thanks for it, it works!), but I like to do “from scratch” installations often, to validate my processes, and it allowed me to identify these issues I stated. I’ll need to change my installation from using miniconda, due to Anaconda ToS. Is what is found on point 3 of section 2.2.5 of the manual the full instructions? Are there any version constraints?
Thanks,
João
From: Stephan Schott <s.schott-verdugo.fz-juelich.de>
Sent: Wednesday, August 9, 2023 1:14 PM
To: Damas Joao CHST <Joao.Damas.syngenta.com>; AMBER Mailing List <amber.ambermd.org>
Subject: Re: [AMBER] questions about AmberTools installation from source
CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.
________________________________
Hi again Joao,
The 22/23 version number has to do with the fact that Amber as a whole has a new version released every 2 years, while AmberTools is released yearly. Because of that, you will find many 22 pointers throughout the code. That said, many codes that were updated for AmberTools23 might have forgotten to change the release date, but overall the last AmberTools is still tailored for Amber22.
I am not sure about the pytraj errors you are seeing, but they might be related to a recent patch that was passed to update pytaj and make it python 3.11 / Cython compatible. Did you apply the last pytraj patch? You should have the update.3 file under ".patches/AmberTools23_Applied_Patches" if that is the case.
P.S., if the trigger was the packmol-memgen numpy floats, the patch is still not there; you don't need to recompile, if you changed the calls as I mentioned before.
Best regards,
Stephan Schott Verdugo
Biochemist
Computational Biophysical Chemistry
Institut für Bio- und Geowissenschaften / Bioinformatik (IBG-4)
Forschungszentrum Jülich GmbH
Wilhelm-Johnen-Straße, 52425 Jülich
Germany
------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------
Forschungszentrum Juelich GmbH
52425 Juelich
Sitz der Gesellschaft: Juelich
Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498
Vorsitzender des Aufsichtsrats: MinDir Volker Rieke
Geschaeftsfuehrung: Prof. Dr.-Ing. Wolfgang Marquardt (Vorsitzender),
Karsten Beneke (stellv. Vorsitzender), Dr. Ir. Pieter Jansens,
Prof. Dr. Astrid Lambrecht, Prof. Dr. Frauke Melchior
------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------
El mié, 9 ago 2023 a las 10:41, Damas Joao CHST via AMBER (<amber.ambermd.org<mailto:amber.ambermd.org>>) escribió:
Dear all,
Triggered by a patching on my installation due to a bug found, I was looking a bit deeper in some details, and I have a few questions.
I am installing AmberTools from source (option 3). This is mainly because I'm doing a cluster-wide installation for multiple users and since we use environment modules, we found that this is the best approach. Up to here, all good.
My first question is: I got the current source through the website, compared the md5sum, all checks out. I was expecting to have version 23, but when I look at the source, several places (e.g., CMakeLists.txt) mention that it is version 22. I wonder if this is just a bug (i.e., it's version 23 but the version strings were not updated) or if it's a source of AmberTools 22.
My second question: when patching, I like to do an installation from scratch, applying the suggested patch after. When I was recompiling, the process errored, while it did not in the past. I paste here a few of the lines of the output, even if there are many more similar errors:
pytraj/topology/topology.cpp: In function '__pyx_CoroutineObject* __Pyx__Coroutine_NewInit(__pyx_CoroutineObject*, __pyx_coroutine_body_t, PyObject*, PyObject*, PyObject*, PyObject*, PyObject*)':
pytraj/topology/topology.cpp:47943:23: error: '_PyErr_StackItem' {aka 'struct _err_stackitem'} has no member named 'exc_type'
47943 | gen->gi_exc_state.exc_type = NULL;
| ^~~~~~~~
pytraj/topology/topology.cpp:47945:23: error: '_PyErr_StackItem' {aka 'struct _err_stackitem'} has no member named 'exc_traceback'
47945 | gen->gi_exc_state.exc_traceback = NULL;
| ^~~~~~~~~~~~~
pytraj/topology/topology.cpp: At global scope:
pytraj/topology/topology.cpp:48152:1: error: too many initializers for 'PyTypeObject' {aka '_typeobject'}
48152 | };
| ^
error: command '/progs/all/opensource/gcc/11.3.0/bin/gcc' failed with exit code 1
make[2]: *** [AmberTools/src/pytraj/CMakeFiles/python-build/pytraj-build.stamp] Error 1
make[1]: *** [AmberTools/src/pytraj/CMakeFiles/pytraj.dir/all] Error 2
make: *** [all] Error 2
My first guess was python version incompatibility, since I think before the miniconda being installed was with python 3.10, and now it was installing 3.11 (since the default in cmake is "latest" and I guess the latest has changed since the last time I installed). I added "-DMINICONDA_VERSION=py310_23.5.2-0" to my cmake flags, and all compiled fine. Do you have cross-compilation matrix for several version of python? It would be good to know which versions AmberTools is compatible with since the only reference in the manual is that it needs to be Python 3.4 or greater.
On the same note, looking at the cmake scripts, I noticed AmberTools is getting the installer from Anaconda/Continuum. Due to Anaconda's ToS, I'll need to change that (follow point 3 of section 2.2.5 of the manual) and it would be good to know which versions of python are possible to use. Additionally, are the python packages references in the manual (numpy, scipy, matplotlib, cython, setuptools, and tkinter) the only ones needed? Any specific versions or the latest should work?
Thank you for your time,
João
This message may contain confidential information. If you are not the designated recipient, please notify the sender immediately, and delete the original and any copies. Any use of the message by you is prohibited.
_______________________________________________
AMBER mailing list
AMBER.ambermd.org<mailto:AMBER.ambermd.org>
http://lists.ambermd.org/mailman/listinfo/amber<
http://lists.ambermd.org/mailman/listinfo/amber>
------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------
Forschungszentrum Juelich GmbH
52425 Juelich
Sitz der Gesellschaft: Juelich
Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498
Vorsitzender des Aufsichtsrats: MinDir Stefan Müller
Geschaeftsfuehrung: Prof. Dr. Astrid Lambrecht (Vorsitzende),
Karsten Beneke (stellv. Vorsitzender), Dr. Ir. Pieter Jansens, Prof. Dr. Frauke Melchior
------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Aug 21 2023 - 02:30:02 PDT
