[AMBER-Developers] Amber build and environment

From: Jason Swails <jason.swails.gmail.com>
Date: Sat, 22 Mar 2014 01:18:47 -0400

Hi everyone,

We've been trying to come up with the best strategy to limit the complexity
of getting Amber up and running. With recent developments, we've
introduced the use of shared objects in the third-party libraries we
install (NetCDF and FFTW3). With this development comes the requirement
that LD_LIBRARY_PATH be augmented with the location of the shared libraries
that we build.

At the very least, this means that users will have to add $AMBERHOME/lib to
LD_LIBRARY_PATH. An idea that we had which I just implemented was to
generate an Amber resource file (similar to ifortvars.sh or
g09.profile/g03.profile for the Intel compilers and Gaussian, respectively)
that can be sourced in peoples' ~/.bashrc or ~/.cshrc to automatically Do
The Right Thing (tm).

The way this currently works is that configure generates
$AMBERHOME/amber.sh and $AMBERHOME/amber.csh files and prints out a message
telling users to source these files and giving brief instructions for
adding it to the shell login.

The advantage of using a resource script is that we can implement
additional environmental changes (like PYTHONPATH requirements as our
Python tools mature) without adding steps to users' install and use

The updated configure script is present in the master branch now, so please
try it out and see what you think. Comments are welcome and appreciated.


Jason M. Swails
Rutgers University
Postdoctoral Researcher
AMBER-Developers mailing list
Received on Fri Mar 21 2014 - 22:30:02 PDT
Custom Search