Installation of GaMD in Amber

GaMD in Amber16

GaMD is fully supported in AMBER16, which was released on Apr 30, 2016 and can be obtained at http://ambermd.org.

GaMD patch for Amber14

"amber14_GaMD_patch_confidential.txt":
GaMD patch for Amber 14 + AmberTools 15
Authors: Yinglong Miao & Ross C. Walker
Date: 29th July 2015
Programs: pmemd and pmemd.cuda
Description: This patch adds temporary GaMD support to AMBER 14.
It is built against AMBER 14 + AmberTools 15 + Amber 14 updates to 12.
This patch is provided for convenience prior to AMBER 16 release which
will feature GaMD support. It is made available by request only and subject
to your agreement not to redistribute this patch or to provide access to
a GaMD patched copy of AMBER without the written permission of
Ross C. Walker or Yinglong Miao.

By applying this patch you accept these terms and certify that you have
been granted permission to use this with Amber 14 + AmberTools 15.


Suppose you have installed "AMBER 14 + AmberTools 15" (http://ambermd.org).
You can apply the GaMD patch and compile the code as the following:

# Uncompress the patch
tar -xvzf amber14-gamd-patch.tgz

# Apply the GaMD patch
# export AMBERHOME to your amber home directory
cp -v amber14_GaMD_patch_confidential.txt $AMBERHOME
cd $AMBERHOME
./update_amber --update
patch -p0 < amber14_GaMD_patch_confidential.txt

# Re-compile the AMBER code
cd $AMBERHOME
./configure gnu
cd $AMBERHOME/src/pmemd/src
make install

cd $AMBERHOME
./configure -cuda gnu
cd $AMBERHOME/src/pmemd/src
make cuda


GaMD-Amber @ UCSD

The GPU version of GaMD-Amber (pmemd.cuda) has been installed on the following computing clusters. You can use it by including the following lines in your job submission script:

On XSEDE Comet (comet.sdsc.xsede.org):
export MODULEPATH=/share/apps/compute/modulefiles/mpi:$MODULEPATH
module load openmpi_ib/1.8.8.g447
module load cuda/6.5
export AMBERHOME=/home/yin/amber
export LD_LIBRARY_PATH=$AMBERHOME/lib:$LD_LIBRARY_PATH
export PATH=$AMBERHOME/bin:$PATH

On Triton Shared Computing Cluster (tscc-login.sdsc.edu):
module load cuda
module load openmpi-x86_64
export AMBERHOME=/home/yin/amber
export LD_LIBRARY_PATH=$AMBERHOME/lib:$LD_LIBRARY_PATH
export PATH=$AMBERHOME/bin:$PATH

On NBCR computing cluster (kivid.ucsd.edu):
module load amberGaMD14/gnu
export CUDA_VISIBLE_DEVICES=`/share/apps/bin/showVisGpu 1`

On Workstations in the McCammon group:
module load cuda/6.5
module load openmpi
module load amber/14-GaMD

 


Copyright Reserved © Yinglong Miao
Last updated: Tue, November 22, 2016