NAMD SCALABLE MOLECULAR DYNAMICS SOFTWARE

点击翻译按钮获取中文摘要

This subproject is one of many research subprojects utilizing the resources provided by a Center grant funded by NIH/NCRR. The subproject and investigator (PI) may have received primary funding from another NIH source, and thus could be represented in other CRISP entries. The institution listed is for the Center, which is not necessarily the institution for the investigator. NAMD (Nanoscale Molecular Dynamics, http://www.ks.uiuc.edu/Research/namd/) is a parallel molecular dynamics code designed for high performance simulation of large biomolecular systems [44, 113]. NAMD employs the prioritized messagedriven execution capabilities of the Charm++/Converse parallel runtime system (http://charm.cs.uiuc.edu/), allowing excellent parallel scaling on both massively parallel supercomputers and commodity workstation clusters. NAMD is distributed free of charge to over 15,000 registered users as both source code and convenient precompiled binaries. A new publication [44] documents the current capabilities, algorithms, and design of NAMD and provides examples of applications ranging from a tutorial exercise to the large multiscale simulation of [37]. NAMD 2.6b1 was released in July 2005 and has been downloaded by over 3600 users, 650 of whom are NIH-funded. This release added binaries for Linux on Itanium, Altix, and Opteron/Athlon 64/EMT64 and improved serial performance by 30% on IBM AIX and 70% on Mac OS X. Memory usage for simulations of large, highly bonded structures such as covalent crystals and for sparse simulations such as coarse-grained models has been greatly reduced without affecting the performance of typical biopolymer simulations. The new Adaptive Biasing Force method [115], implemented in Tcl, efficiently calculates the potential of mean force along a reaction coordinate by applying adaptive biasing forces to provide uniform sampling. With new Tcl-based boundary potentials, scripted forces may be efficiently applied individually to large numbers of atoms. Finally, the OPLS force field is now supported. The new 2000-processor Cray XT3 at the Pittsburgh Supercomputing Center (PSC) became available for NAMD porting in August, 2005. As the XT3 employs both a completely new interconnect and a completely new operating system, we anticipated (and encountered) difficulties, but with assistance from Cray and PSC staff NAMD now runs with most features (the XT3 lacks, e.g., the sockets needed for interactive simulations) and with good performance and scalability (tuning continues). Similarly, the single-rack BlueGene/L system at the San Diego Supercomputer Center (SDSC) also became available, allowing the completion of the NAMD port (previously used only for performance tuning) by Resource personnel. It will be impossible to distribute NAMD binaries for the Cray XT3 or BlueGene/L since due to the lack of dynamic linking on this platforms any system upgrade would require recompilation. Finally, NAMD has been ported to the new Mac OS for Intel processors and to the x86-64 version of Windows, both natively for full performance.