Links

Molecular Modeling Software

• The primary modeling tool used on the cluster is the CHARMM (Chemistry at HARvard Molecular Mechanics) software package. Dr. Bernard Brooks, the head of the LCB, is one of the primary developers of CHARMM. The CHARMM Development Project involves a network of developers in the United States and elsewhere working with Professor Karplus and his group at Harvard to develop and maintain the package.
  
• Another piece of commercial software that performs quantum calculations is Q-Chem. Q-Chem uses efficient techniques to parallelize methods such as Hartree-Fock and DFT single-point energies, geometry optimizations, and frequency calculations. Q-Chem combines these advanced techniques into a single package for use in ab initio computations. Q-Chem is also capable of performing QM/MM calculations via a CHARMM QM/MM interface.
  
• For ab initio quantum mechanical calculations we have available GAMESS-US and GAMESS-UK. Both software packages provide means of performing QM/MM calculations via a CHARMM QM/MM interface.
  
• VMD is a multiplatform molecular visualization program. It supports many different coordinate and structure file formats, including those used by CHARMM, PQR, and AMBER. It also is able to read GAMESS log files. We use it to create beautifully rendered images of the molecular systems which we study (some of which can be found on this Web site).
  
• 
  

Compilers and Parallel Execution Tools

• Most of our compiling is done with the open-source GNU Compiler Collection. The version of MPICH that we currently use is compiled with version 3.3.3 of gcc.
  
• The Portland Group compiler suite is a set of high-end compilers for Fortran (77 and 90), C, and C++ that are designed to work in high performance computing environments. We compile some of our parallel code with these compilers.
  
• MPICH is a portable implementation of the MPI parallel computing standard. The software was developed at Argonne National Laboratory. We currently use the 1.2.5 series, but MPICH-2, which supports both the MPI-1 and MPI-2 standards, is now available. The parallel CHARMM jobs we run on LoBoS use MPICH to handle their MPI functionality.
  
• Open MPI: The Open MPI Project is an open source MPI-2 implementation that is developed and maintained by a consortium of academic, research, and industry partners. Open MPI is therefore able to combine the experise, technologies, and resources from all across the High Performance Computing community in order to build the best MPI library available. Open MPI offers advantages for system and software vendors, application developers and computer science researchers.
  
• LA-MPI is another implementation of the MPI standard, developed at Los Alamos National Laboratory. It is no longer in active development, but existing versions continue to be maintained. Development efforts are focusing on releasing Open MPI.
  
• We use Myricom products for low latency network communication, we also use the GM drivers to power our Myrinet communications. We're currently using a custom-built version of GM

Administrative and Monitoring Software

• The LoBoS job scheduler is the TORQUE Resource Manager disributed by Cluster Resources, Inc.. We currently use version 1.1.0p6 with a couple of small patches by Tim Miller and others. TORQUE is an improvement over OpenPBS and in tests it has been able to scale to managing over one thousand jobs simulataneously.
  
  
• We've found Nagios useful for general cluster monitoring and supervising important systems such as our RAID arrays. Nagios can automatically notify an administrator by e-mail when it detects a problem.
  
  
• For performance monitoring, we also have Ganglia configured. Ganglia is useful because it requires little configuration and scales well.
  
  

Miscellaneous

• We use GNUPlot for many of our graphing needs. This software makes it easy to create good looking graphs of data on the Unix command line.
  
  

Computational Biophysics Research Groups

• NIH Research Resource Center for the Development of Multiscale Modeling Tools for Structural Biology.
• The Center for Computational Chemistry at the University of Georgia.
• The Minnesota Computational Chemistry group.
• The Post group at Purdue University.

Other Computational Chemistry Resources (Books, Journals, etc.)

• The Center for Molecular Modeling at NIH.
• Biophysical Chemistry, the journal.
• Computational Chemistry List home page.
• Reviews in Computational Chemistry book series home page.
• The Journal of Theoretical and Computational Chemistry.

Web Pages and Institutions of Collaborators and Former Group Members

• Scott Feller's research group at Wabash University.
• Milan Hodoscek's page at the Center for Molecular Modeling at the Slovenian National Institute of Chemistry.
• Milan Hodoscek's Benchmark Page.
• The Center for Biologics Evaluation and Research at the FDA.

Crystal Data and Crystallography

• International Union of Crystallography.
• The RCSB Protein Data Bank.
• The Nucleic Acid Database at Rutgers.
• Lawrence Livermore National Laboratory Crystallographic Web Applets.

Structure and Prediction

• Los Alamos HIV Seq. Data, Nucleic Acid Modeling Tool (NAMOT)
• UCSF Neural Net Structure Prediction
• UCLA/DOE protein fold recognition server
• EMBL Heidelberg WWW Services
• Sugar Momento estimation of sucrose properties

Local Links Pages

• The CHARMM links page.
• The Laboratory of Computational Biology links page
• The LoBoS links page contains links to many useful computational chemistry and cluster resources.

Parallel Computing Resources

• Cluster Hardware and Design
• Beowulf.org has an extensive list of cluster resources, as well as a guide to building your own Beowulf cluster.
• Engineering a Beowulf-style Compute Cluster is an online book with good information on cluster design.
• Myricom is a company that produces Myrinet, a high-speed, highly-scalable cluster interconnect technology.
• InfiniBand is an emerging set of network standards for high-performance computing.
• Topspin produces InfiniBand network gear.
• Foundry Networks provides the high-performance ethernet switches which LoBoS uses.
• Parallel Programming Libraries and Other Resources
• MPI is a robust, open interface specification for parallel computing.
• Scyld Software provides a custom OS kernel and other resources for use in high-performance computing.
• The Extreme Linux section of Linux Magazine has many excellent articles about cluster programming.
• An MPI tutorial hosted at Lawrence-Livermore National Laboratory.
• The Web page for a course on Parallel Computing taught by Dr. Chen Yu Zong at the University of Singapore has several MPI tutorials.

Cluster Software

• Application Software
• CHARMM is a customizable, parallelized, multi-platform molecular modelling program.
• MPICH is an open-source implementation of MPI for Unix and Windows systems.
• PVM is another set of parallel communications libraries, similar to MPI.
• For benchmarking, HPL is a freely-available implementation of Linpack.
• Administration and Monitoring Software
• LoBoS uses Torque, an extension of OpenPBS, as its batch processing system.
• NAGIOS is a flexible, extensible distributed monitoring tool.
• Ganglia is a distributed monitoring tool designed specifically for clusters.

Other Clusters of Note

• Biowulf is another cluster at NIH. Some old LoBoS nodes are now part of Biowulf.
• Vrana at the Slovenian National Institute of Chemistry, which is run by Milan Hodoscek, who did early work on LoBoS.
• The Beowulf Project at the Center for Advanced Computing Research at CalTech.
• The Hoffman Cluster at UCLA.
• Avalon is a Beowulf at Los Alamos National Laboratory which uses HP Alphas instead of Intel processors.
• MadDog at the Theoretical Physics group at the University of Manchester.