Hi,
I have now installed a fresh amber on gpu cluster as follows:
Serial version:
tar -jxf AmberTools14.tar.bz2
cd amber14
export AMBERHOME=`pwd`
./configure gnu
make $AMBERHOME/bin/install
source amber.sh
make test
Further, for parallel version, the mpich is old in our cluster, I had downloaded new version
mpich-3.1.4.tar.gz and then started the parallel installation:
tar xvf mpich-3.1.4.tar.gz
building directory to the path of user i.e. my user and connected to the path of amber and adding the path to the bash profile.
than
cd amber14
./configure -mpi gnu
make install
source amber.sh
DO_PARALLEL="mpirun -np 4"
make test ?
This all what I had done.
Now how to understand that the installation is successful, and the files what else missed here please let me know.
thanks
Abhi
?
>>>>>>>>>>>>
You really need to provide the *exact* process (and I mean verbatim)
for us to have any hope of figuring out how this happened, especially
since I don't think anyone has had this happen to them before (I know
it hasn't happened to me, and I compile Amber a lot!). For example,
here's exactly what I did to install a test version of Amber this
morning:
tar -jxf AmberTools14.tar.bz2
cd amber14
export AMBERHOME=`pwd`
./configure gnu
make -j4 install
Also any other relevant information you can provide (such as what
compiler version, what MPI library, etc etc) would be great. Thanks,
-Dan
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Apr 14 2015 - 18:00:02 PDT
