Live Webcast 15th Annual Charm++ Workshop

The structure and function of biomolecular machines are the foundation on which living systems are built. Genetic sequences stored as DNA translate into chains of amino acids that fold spontaneously into proteins that catalyze chains of reactions in the delicate balance of activity in living cells. Interactions with water, ions, and ligands enable and disable functions with the twist of a helix or rotation of a side chain. The fine machinery of life at the molecular scale is observed clearly only when frozen in crystals, leaving the exact mechanisms in doubt. One can, however, employ molecular dynamics simulations to reveal the molecular dance of life in full detail. Unfortunately, the stage provided is small and the songs are brief. Thus, we turn to petascale parallel computers to expand these horizons. Biomolecular simulations are challenging to parallelize. Typically, the molecular systems to be studied are not very large in relation to the available memory on computers: they contain ten thousand to a few million atoms. Since the size of basic protein and DNA molecules to be studied is fixed, this number does not increase in size significantly. However, the number of time steps to be simulated is very large. To simulate a microsecond in the life of a biomolecule, one needs to simulate a billion time steps. The challenge posed by biomolecules is that of parallelizing a relatively small amount of computation at each time step across a large number of processors, so that billions of time steps can be performed in a reasonable amount of time. In particular, an important aim for science is to effectively utilize the machines of the near future with tens of petaFLOPs of peak performance to simulate systems with just a few million atoms. Some of these machines may have over a million processor cores, especially those designed for low power consumption. One can then imagine the parallelization challenge this scenario poses. NAMD is a highly scalable and portable molecular dynamics (MD) program used by thousands of biophysicists. We show in this chapter how NAMD's parallelization methodology is fundamentally well-suited for this challenge, and how we are extending it to achieve the goals of scaling to petaFLOP machines. We substantiate our claims with results on large current machines like IBM's Blue Gene/L and Cray's XT3. We also talk about a few biomolecular simulations and related research being conducted by scientists using NAMD.