AMBER at MCSR

The Amber computational chemistry package is available in three versions on two platforms: Amber 7, 8, and 9 on redwood and Amber 8 on mimosa. MCSR has a site license for Amber 8/9. Let us know if you would like a copy of the CD. MCSR also has Amber 10 installed on Sequoia.

Amber 7/8/9 on Redwood

Amber versions 7, 8, and 9 are installed on redwood at /usr/local/appl/amber7, /usr/local/appl/Amber8, and /usr/local/appl/amber9. You should set the AMBERHOME environment variable to /usr/local/appl/amber7, /usr/local/appl/Amber8, or /usr/local/appl/amber9. The executables are in $AMBERHOME/exe, and the examples are in $AMBERHOME/examples. You may browse this directory and copy example files into your own working directory for execution. You should ensure that the $AMBERHOME/exe/ directory is in your PATH environment variable.

Amber version 7 example on Redwood

To run a parallel Amber7 program, use mpirun, and specify the number of CPUs/processes to run on with the -np option to mpirun. For example, to run the Amber "nucgen" program on 4 CPUs, using the sample "nuc.in" input file supplied in the cd /usr/local/appl/amber7/tutorial/polyA-polyT directory:

1. Ensure AMBERHOME environment variable is set to /usr/local/appl/amber7
   
2. Ensure PATH environment variable includes $AMBERHOME/exe/
   
3. Copy the nuc.in file from the $AMBERHOME/tutorial/polyA-polyT directory into your working directory.
   
4. Type: mpirun -np 4 $AMBERHOME/exe/nucgen -O -i nuc.in -o nuc.out -d $AMBERHOME/dat/leap/parm/nucgen.dat -p nuc.pdb
   Make sure it created a new nuc.out
   
5. Verify results in view nuc.out
   
6. For more optimal use of memory, try using the dplace utility:
   Type: mpirun -np $NCPUS dplace -c1-$NCPUS -s1 mpirun -np 4 $AMBERHOME/exe/nucgen -O -i nuc.in -o nuc.out -d $AMBERHOME/dat/leap/parm/nucgen.dat -p nuc.pdb
   where $NCPUS has been defined as the number of CPUs you want to run on.
7. You should use PBS to submit your Amber jobs. Here is a sample script that submits a 4-processors Amber 7 job from the user's /ptmp directory using the dplace utility. The script requests 1000 hours and 3 GB to run.
   #PBS -S /bin/csh
#PBS -l cput=1000:00:00
#PBS -l ncpus=4
#PBS -l mem=3GB
setenv NCPUS 4
setenv AMBERHOME /usr/local/appl/amber7
set path = ( $path $AMBERHOME/exe )
cd /ptmp/$USER/
mpirun -np $NCPUS dplace -c1-$NCPUS -s1 $AMBERHOME/exe/nucgen -O -i nuc.in -o nuc.out -d $AMBERHOME/dat/leap/parm/nucgen.dat -p nuc.pdb


Amber version 8 example on Redwood

To run a parallel Amber 8 program, use mpirun, and specify the number of CPUs/processes to run on with the -np option to mpirun. For example, to run the Amber "dna_pol" program on 8 CPUs, using the input files supplied in the /usr/local/appl/Amber8/test/dna_pol/ directory:

1. Ensure AMBERHOME environment variable is set to /usr/local/appl/Amber8
   
2. Ensure PATH environment variable includes $AMBERHOME/exe/
   
3. Ensure LD_LIBRARY_PATH environment variable is set to /opt/cmplrs/8.0.046/lib/:/opt/cmplrs/8.0.066/lib
   
4. Copy the files from the $AMBERHOME/tutorial/polyA-polyT directory into your working directory as different names shown below.
   cp $AMBERHOME/test/dna_pol/mdin dna_pol_mdin
cp $AMBERHOME/test/dna_pol/inpcrd dna_pol_inpcrd
cp $AMBERHOME/test/dna_pol/prmtop dna_pol_prmtop
cp $AMBERHOME/test/dna_pol/inpdip inpdip

5. Type: mpirun -verbose -np 8 $AMBERHOME/exe/sander -O -i dna_pol_mdin -o dna_pol.out.v8_8p -c dna_pol_inpcrd -p dna_pol_prmtop -r dna_pol.restrt -ref dna_pol_inpcrd
   Make sure it created a new dna_pol.out.v8_8p
   
6. Verify results in view dna_pol.out.v8_8p
   
7. For more optimal use of memory, try using the dplace utility after setting NCPUS environment variable to 8 and TOP_CPU variable to `expr $NCPUS - 1`:
   Type: mpirun -np $NCPUS dplace -c0-$TOP_CPU -s1 $AMBERHOME/exe/sander -O -i dna_pol_mdin -o dna_pol.out.v8_8p -c dna_pol_inpcrd -p dna_pol_prmtop -r dna_pol.restrt -ref dna_pol_inpcrd
   where $NCPUS has been defined as the number of CPUs you want to run on.
8. You should use PBS to submit your Amber jobs. Here is a sample script that submits a 8-processor Amber 8 job from the user's /ptmp directory using the dplace utility. The script requests 1000 hours and 3 GB to run.
   #PBS -S /bin/csh
#PBS -l cput=1000:00:00
#PBS -l ncpus=8
setenv LD_LIBRARY_PATH /opt/cmplrs/8.0.046/lib/:/opt/cmplrs/8.0.066/lib
setenv NCPUS 8
setenv TOP_CPU=`expr $NCPUS - 1`
setenv AMBERHOME /usr/local/appl/Amber8
set path = ( $path $AMBERHOME/exe )
cd /ptmp/$USER/
mpirun -np $NCPUS dplace -c0-$TOP_CPU -s1 $AMBERHOME/exe/sander -O -i dna_pol_mdin -o dna_pol.out.v8_8p -c dna_pol_inpcrd -p dna_pol_prmtop -r dna_pol.restrt -ref dna_pol_inpcrd


For support running Amber on any MCSR system where it is installed, see the FAQ below, or email assist@mcsr.olemiss.edu.

Amber 8 on Mimosa

Amber version 8 example on Mimosa

To run a parallel Amber 8 program, use mpirun, and specify the number of CPUs/processes to run on with the -np option to mpirun. For example, to run the Amber "dna_pol" program on 8 CPUs, using the input files supplied in the /usr/local/apps/working/amber8/mimosa_test directory:

1. Ensure AMBERHOME environment variable is set to /usr/local/apps/working/amber8
   
2. Ensure PATH environment variable includes $AMBERHOME/exe/
   
3. Ensure LD_LIBRARY_PATH environment variable is set to " /usr/local/apps/pgi-5.2/linux86/5.2/lib/:/usr/local/apps/pgi-5.2/linux86/5.2/liblf "
   
4. Copy the files from the $AMBERHOME/mimosa_test directory into your working directory shown below.
   cp $AMBERHOME/test/dna_pol/dna_pol_mdin dna_pol_mdin
cp $AMBERHOME/test/dna_pol/dna_pol_inpcrd2p dna_pol_inpcrd2p
cp $AMBERHOME/test/dna_pol/dna_pol_prmtop2p dna_pol_prmtop2p
cp $AMBERHOME/test/dna_pol/inpdip inpdip

5. You should use PBS to submit your Amber jobs. Here is a sample script that submits a 2-processor(nodes) Amber 8 job from the user's /ptmp directory. The script requests 2 nodes to run.
   #PBS -S /bin/csh
#PBS -l nodes=2
setenv LD_LIBRARY_PATH /usr/local/apps/pgi-5.2/linux86/5.2/lib/:/usr/local/apps/pgi-5.2/linux86/5.2/liblf
setenv PGI /usr/local/apps/pgi-5.2
setenv AMBERHOME /usr/local/apps/working/amber8
set path = ( $path $AMBERHOME/exe )
cd /ptmp/$USER/
/usr/local/apps/pgi-5.2/linux86/5.2/bin/mpirun -verbose -np 1 $AMBERHOME/exe/sander -O -i dna_pol_mdin -o dna_pol.out.v8_2p -c dna_pol_inpcrd2p -p dna_pol_prmtop2p -r dna_pol.restrt2p -ref dna_pol_inpcrd2p


For support running Amber on any MCSR system where it is installed, see the FAQ below, or email assist@mcsr.olemiss.edu.

MCSR/Ole Miss Frequently Asked Questions about Amber (under construction)

1. What versions of Amber are installed on what systems at MCSR?
2. How do I run a parallel Amber job on redwood?
3. How do I run a batch Amber job on redwood?
4. Where are example Amber files on redwood?
5. Which system do you suggest to use in terms of performance?

MCSR/Ole Miss Frequently Asked Questions about Amber (under construction)

1. What versions of Amber are installed on what systems at MCSR?
   

   There are two versions of Amber installed on redwood and mimosa. These versions are Amber versions 7 and 8.
   

2. How do I run a parallel Amber job on redwood?
   

   To run a parallel Amber job on Redwood, you need to use PBS script to submit your jobs, please refer to the examples above for each version.
   

3. How do I run a batch Amber job on redwood?
   

   You can run up to 8 processor batch job in Redwood, however, you need to setup the environment variables as shown in the examples above. These environment variables are AMBERHOME, NCPUS, LD_LIBRARY_PATH and PATH.

   In csh/tcsh:
   setenv NCPUS 8
setenv TOP_CPU `expr $NCPUS - 1`
setenv LD_LIBRARY_PATH /opt/cmplrs/8.0.046/lib/:/opt/cmplrs/8.0.066/lib
setenv AMBERHOME /usr/local/appl/Amber8
set path = ( $path $AMBERHOME/exe )


   In bash/ksh/sh:
   export NCPUS=8
export TOP_CPU=`expr $NCPUS - 1`
export LD_LIBRARY_PATH='/opt/cmplrs/8.0.046/lib/:/opt/cmplrs/8.0.066/lib'
export AMBERHOME=/usr/local/appl/Amber8
export PATH="$PATH:$AMBERHOME/exe"

   

4. Where are example Amber files on redwood?
   

   The example files are in the folders of test and examples under the /usr/local/appl/amber7 and /usr/local/appl/Amber8 directories. If you would like to use an example or test file, don't forget to copy the files to your current directory since you cannot run these files from Amber's home directories.

5. Which system do you suggest to use in terms of performance?
   

   Since Redwood is a shared memory NUMAlink architecture designed by SGI, it is optimized for faster and better performances. For detailed performance analysis and results, please go to our Parallelogram news article, Amber Version 8 on Redwood . However, due to heavy load on the redwood system, running 4 cpu Amber jobs on Redwood and higher cpu Amber jobs on mimosa would be strongly suggested to decrease the waiting time in the queue.