next up previous contents index
Next: Parallel Computing on Clusters Up: Hardware and Software Requirements Previous: Basic Hardware and Software   Contents   Index

Multi-core CPUs and GPU Acceleration

VMD makes full use of multi-core processors and multiple GPUs for acceleration of the most computationally demanding visualization and analysis tasks. Multi-core CPUs accelerate features including interactive molecular dynamics [8,9], bond determination, ``within'' atom selections and derivatives, so-called streamline or field line visualizations [10], radial distribution functions [11], and high quality renderings using the built-in Tachyon [12,13,14] and OSPRay [15] ray tracing engines. VMD also supports GPU acceleration using CUDA, and takes advantage of both multi-core CPUs and GPUs for acceleration of electrostatics (i.e. ``volmap coulomb'', and ``volmap coulombmsm'') [16,17,18,19,20,21,22,23,24,25], implicit ligand sampling (i.e. ``volmap ils''), computation of radial distribution functions [11], quality-of-fit cross correlation calculation for hybrid fitting methods [26,27], trajectory clustering analyses [28], and computation and rendering of molecular orbitals [29,5,30,31,32] and molecular surfaces [33,34,14,35,36], The latest versions of VMD also incorporate a GPU-accelerated batch and interactive versions of the Tachyon ray tracing engine [35,36,37,38,7] based on NVIDIA OptiX and CUDA [39].