Intel® Parallel Computing Center at the San Diego Supercomputer Center, University of California San Diego

Principal Investigators:

Ross C. Walker
Associate Research Professor, San Diego Supercomputer Center
Adjunct Associate Professor, Department of Chemistry and Biochemistry

Prof. Walker serves as an Associate Research Professor at the San Diego Supercomputer Center at the University of California, San Diego (UCSD), as an Adjunct Associate Professor of Chemistry and Biochemistry at UCSD and as biosciences lead of the Scientific Applications Computing group at the San Diego Supercomputing Center. He is a principal developer of the AMBER Molecular Dynamics and AmberTools Software Suites and has extensive experience in developing and teaching classical and QM/MM molecular dynamics techniques. Prof. Walker holds a Ph.D. and Diploma of Imperial College in Computational Chemistry, a Masters in Science degree in Chemistry from the UK's Imperial College of Science, Technology and Medicine as well as an Associateship of the Royal College of Science. Prior to joining UCSD Prof. Walker worked as an applications scientist and postdoctoral research scholar at The Scripps Research Institute in La Jolla, CA.

Amit Majumdar
Director, Data Enabled Scientific Computing Division, San Diego Supercomputer Center
Associate Professor, Department of Radiation Medicine and Applied Sciences, School of Medicine, UC San Diego

Dr. Majumdar directs the Data Enabled Scientific Computing Division at the San Diego Supercomputer center and also holds an Affiliated Associate Professorship in the Department of Radiation Medicine and Applied Sciences at the UC San Diego School of Medicine. He works in the fields of high performance computing, computational science and science gateways. He has over 20 years of expertise in parallel computing and collaborates with numerous researchers on tuning for HPC architectures. Dr. Majumdar holds a Ph.D. from the University of Michigan in Nuclear Engineering and Scientific Computing as well as a Masters degree in Nuclear Engineering from Idaho State University.

Andreas Goetz
Project Scientist, San Diego Supercomputer Center

Dr. Goetz is an Assistant Project Scientist at the San Diego Supercomputer Center (SDSC). His work combines aspects of (bio) chemistry, physics, numerical mathematics, software development and high performance computing. He is a contributing author to the ADF quantum chemistry software and the AMBER and AmberTools software packages for biomolecular simulation. Both are widely used in academic and industrial research. Dr. Goetz collaborates on a variety of research projects in molecular simulation, computational enzymology and drug design. Dr. Goetz holds a Ph.D. in Theoretical Chemistry and a Masters Degree in Chemistry from the Friedrich-Alexander University in Erlangen, Germany. Prior to joining SDSC in 2009 Dr. Goetz was a postdoctoral scholar in quantum chemistry at the VU University in Amsterdam.

Perri Needham
Postdoctoral Fellow

Dr. Needham is a postdoctoral research associate working on advanced molecular dynamics simulations. Her specific focus is on GPU accelerated molecular dynamics as well as Intel funded work to explore the use of Intel Many Integrated Core technologies to accelerate molecular dynamics within the AMBER software package. Dr. Needham holds a Ph.D. in Computational Chemistry from the University of Manchester and a Masters degree in High Performance Computing from Edinburgh University.

Description:

The Intel® Parallel Computing Center (Intel® PCC) at the San Diego Supercomputer Center (SDSC) is led by Prof. Ross Walker, Dr. Amit Majumdar and Dr. Andreas Goetz and focuses on three distinct projects.

Prof. Walker leads work on advanced many- and multi-core optimization of classical all atom Molecular Dynamics simulations within the life sciences arena. Specifically work focuses on improving the performance of the AMBER (http://ambermd.org) Molecular Dynamics software on Intel Xeon® and Intel Xeon Phi™ architectures for both single and multi-node configurations. Work conducted as part of this center feeds directly back into the AMBER code base and also to the Molecular Dynamics community as a whole having impact on the fields of drug discovery, bio-materials and bio-catalysis. The AMBER software is used in all major US and International Research Universitys as well as broadly across the pharmaceutical and biotech industries. Optimization and porting work in this project will thus directly benefit a large number of scientists.

Dr. Majumdar leads work on many- and multi-core optimization of the neuronal simulation software package NEURON (http://www.neuron.yale.edu). He is collaborating with Profs. Michael Hines and Ted Carnevale at Yale University. Together they are working to optimize spiking neuronal simulations on Intel Xeon® and Intel Xeon Phi™ architectures. Ultimately this work will enhance the performance of complex neuronal network simulations providing insight into neuro-degenerative diseases, brain function and computational neuroscience. Dr. Majumdar is also involved in developing the NeuroScience Gateway (http://www.nsgportal.org) which provides software such as NEURON running on HPC resources to the larger computational neuroscience community. The work carried out as part of this Intel® PCC project will allow such gateways to have a broader reach.

Dr. Goetz and Prof. Walker jointly lead work on optimization of multi-scale hybrid Quantum Mechanical / Molecular Mechanical (QM/MM) techniques for the direct simulation of reactions in enzymes and solution. Their work is focusing on overcoming the diagonalization and Fock matrix manipulation bottlenecks in semi-empirical and DFTB QM/MM calculations as well as improved serial and parallel performance on Intel® architectures. The goals are to enable the use of the Intel Xeon Phi™ architecture as an accelerator for QM/MM/MD and Adaptive QM/MM/MD simulations as well as extensive parallel optimization for the Intel Xeon® architecture. Ultimately this work will be widely distributed within the OpenSource AmberTools software suite (http://ambermd.org/#AmberTools) as well as the semi-empirical QM/MM module within the Charmm MD package (http://www.charmm.org) benefiting tens of thousands of users. It will improve the rate at which small molecules are parameterized for drug discovery applications and also enable accurate simulation of enzymatic reactions on the atomic scale. The ability to directly model chemical reactions has utility in many fields, one such example being work to design advanced catalysts for more efficient bio-fuel production.

The SDSC Intel® PCC also acts as a conduit for offering training in parallel and accelerator programming to researchers and students at UCSD and the broader San Diego biotech arena.

Publications:

Related websites:

The Amber Molecular Dynamics Software: http://ambermd.org
The NEURON Software: http://www.neuron.yale.edu
The NeuroScience Gateway: http://www.nsgportal.org
The AmberTools Software Suite: http://ambermd.org/#AmberTools
The San Diego Supercomputer Center: http://www.sdsc.edu
The University of California San Diego: http://www.ucsd.edu

For more complete information about compiler optimizations, see our Optimization Notice.