Re: [AMBER] Compiling AMBER23

From: accuratefreeenergy--- via AMBER <amber.ambermd.org>
Date: Tue, 29 Aug 2023 06:50:12 -0400

Hi Javier,

        It is likely due to a bug in Miniconda-Python3.11. (I posted a similar message on 7/30/23):

        You can try to modify the file $AMBERHOME/cmake/PythonInterpreterConfig.cmake, around line 60. Comment out the default "latest" version and change to 3.10.

=================================
# set(MINICONDA_VERSION latest)
        set(MINICONDA_VERSION py310_23.5.2-0)
=================================

(or go to https://repo.anaconda.com/miniconda/ and pick up the one you need)

Taisung




-----Original Message-----
From: Javier O Sanlley-hernandez via AMBER <amber.ambermd.org>
Sent: Monday, August 28, 2023 7:22 PM
To: amber.ambermd.org
Subject: [AMBER] Compiling AMBER23

Hello,

I am writing because I am trying to compile pmemd.cuda to run molecular dynamics simulations through the cmake approach as described in the manual. However, when I run the make install command after building I get the following error message.

This is the error message I am getting:
src/mpi4py.MPI.c:198:12: fatal error: longintrepr.h: No such file or directory
 198 | #include "longintrepr.h"
   | ^~~~~~~~~~~~~~~
compilation terminated.
error: command '/sw/spack/delta-2022-03/apps/openmpi/4.1.2-gcc-11.2.0-37px7gc/bin/mpicc' failed with exit code 1
make[2]: *** [AmberTools/src/mpi4py-3.1.3/CMakeFiles/mpi4py.dir/build.make:203: AmberTools/src/mpi4py-3.1.3/CMakeFiles/mpi4py-build/mpi4py-build.stamp] Error 1
make[1]: *** [CMakeFiles/Makefile2:4628: AmberTools/src/mpi4py-3.1.3/CMakeFiles/mpi4py.dir/all] Error 2
make: *** [Makefile:156: all] Error 2
exit

These are the options in my cmake_file to compile with gnu and use mpi:

cmake $AMBER_PREFIX/amber22_src \

    -DCMAKE_INSTALL_PREFIX=$AMBER_PREFIX/amber22 \

    -DCOMPILER=GNU \

    -DMPI=TRUE -DCUDA=TRUE -DINSTALL_TESTS=TRUE \

    -DDOWNLOAD_MINICONDA=TRUE \

    2>&1 | tee cmake.log

These are the loaded modules I am working with:

  1) cue-login-env/1.0 3) ucx/1.11.2 5) cuda/11.6.1 7) default

  2) gcc/11.2.0 4) openmpi/4.1.2 6) modtree/gpu

I also tried the options listed on an earlier post<http://archive.ambermd.org/202307/0083.html> but that didn’t work either, still got the same error.

Let me know how to best proceed or if you need anything else.

Best,
_______________________________________________
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 Aug 29 2023 - 04:00:02 PDT
Custom Search