The molecular electrostatic potential (MEP) is a useful tool to design and
develop drugs. However, the evaluation of this property using quantum chemistry
methods presents a challenge for molecules of medium or large size since this is
computationally expensive. In this chapter, we showed two implementations of this
property over graphics processing units (GPU). In the first instance, we discussed some
details that must be considered when GPUs are involved in high-performance
computing. After this step, the algorithms considered to evaluate MEP over GPUs are
exposed to observe the main differences between a method with minimal
approximations and another one where usual approximations are implemented in many
quantum chemistry codes. The benefits provided by these graphics cards are evidenced
when our implementations are applied over molecules of considerable size like those
found in protein-ligand complexes, where usually the electrostatic potential is modeled
by a set of point charges.
Keywords: Cuda-C, GPUs, Molecular Electrostatic Potential, Multipolar
Expansion, Proteins.
