Dear Prof. Case,
Thanks a lot for your kindly reply. Firstly, the underscore of "$AMBERHOME"
is just a typo, please ignore it. Secondly, when I built Amber20 with
run_cmake file, the only thing I change is to set CUDA flag to TURE. Please
see it as following.
-------------------------------------------------------------------------
# Assume this is Linux:
cmake $AMBER_PREFIX/amber20_src \
-DCMAKE_INSTALL_PREFIX=$AMBER_PREFIX/amber20 \
-DCOMPILER=GNU \
-DMPI=FALSE -DCUDA=TRUE -DINSTALL_TESTS=TRUE \
-DDOWNLOAD_MINICONDA=TRUE -DMINICONDA_USE_PY3=TRUE \
2>&1 | tee cmake.log
----------------------------------------------------------------------------
I have tried other application related with minoconda python, such as
parmed, and they worked fine. So I guess there might be some configuration
problems with pdb4amber?
Many thanks,
Tao
David A Case <david.case.rutgers.edu> 于2020年7月22日周三 下午7:55写道：
> On Wed, Jul 22, 2020, Huang Tao wrote:
> >
> >When I tried to use pdb4amber after the installation of AMBER20, one error
> >happened.
> >
> >"importlib_metadata.PackageNotFoundError: No package metadata was found
> for
> >pdb4amber"
> >
> >I have checked the first line of pdb4amber file under $AMBER_HOME/bin. The
> >first line is as following.
> >
> >#!$AMBER_HOME//miniconda/bin/python
>
> 1. Can you say when you built Amber20? There has been a recent change in
> what miniconda does that seems to affect pytraj and pdb4amber. Did you
> make any changes to the run_cmake script?
>
> I think we may know what to do here, and (if so), we will publish a update
> soon. (cc-ing to Jason)
>
> Possible aside:
>
> The reference to $AMBER_HOME is very odd: the "amber.sh" script (in your
> installation folder), and everything else we use, uses $AMBERHOME (no
> underscore). We are trying to make things independent of this variable,
> but
> there are still places that refer to $AMBERHOME. Do you have that also
> set on
> your machine? Do you know if you did something non-standard to get
> $AMBER_HOME? This may be unrelated to your pdb4amber problem, but using a
> non-standard name here is likely to lead to trouble.
>
> ...regards...dac
>
>
> _______________________________________________
> 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 Wed Jul 22 2020 - 05:30:02 PDT
