GPU accelerated implementation of NCI calculations using promolecular density

Abstract

The NCI approach is a modern tool to reveal chemical noncovalent interactions. It is particularly attractive to describe ligand-protein binding. A custom implementation for NCI using promolecular density is presented. It is designed to leverage the computational power of NVIDIA graphics processing unit (GPU) accelerators through the CUDA programming model. The code performances of three versions are examined on a test set of 144 systems. NCI calculations are particularly well suited to the GPU architecture, which reduces drastically the computational time. On a single compute node, the dual-GPU version leads to a 39-fold improvement for the biggest instance compared to the optimal OpenMP parallel run (C code, icc compiler) with 16 CPU cores. Energy consumption measurements carried out on both CPU and GPU NCI tests show that the GPU approach provides substantial energy savings. © 2017 Wiley Periodicals, Inc.

Accurately extracting the signature of intermolecular interactions present in the NCI plot of the reduced density gradient versus electron density

“…All computations were carried out with the NCIgpu code for pro-molecular ED. 20 Full geometry optimizations were performed at the B3LYP/6-311++G** level of theory in gas phase using the Gaussian 09 package. 21 All minima were carefully characterized through harmonic frequency analysis (stationary points with no imaginary frequencies).…”

New Way for Probing Bond Strength

“…17 Such grid computational scheme could leverage the massively parallel GPU technology for the acceleration of δg pair calculation, as demonstrated with NCI in a previous study. 18 Just like in the NCI approach, this δg pair signature can be transposed in the real space taking the form of isosurfaces colored according to the value of signed ED (sign(λ 2 )ρ) (see Figure 1). Clearly, δg pair accounts for the tendency of electrons to be shared between the two atoms.…”

GPUs as boosters to analyze scalar and vector fields in quantum chemistry

“…In our laboratory, we have developed the Graphics Processing Units for Atoms and Molecules (GPUAM) code to evaluate several scalar and vector quantum chemistry fields coupled with the Atoms in Molecules (AIM) analysis . It is worth to mention that there are several codes with some of these capabilities . However, GPUAM has been designed to use heterogeneous computational resources.…”

“…The main reason to implement derivatives of the electron density obtained by semiempirical methods stems from reports, where these methods evaluate the NCI and consequently short times are needed to study large systems. Thus, GPUAM can be a useful tool for this purpose.…”

“…[30] It is worth to mention that there are several codes with some of these capabilities. [31][32][33][34][35][36] However, GPUAM has been designed to use heterogeneous computational resources. The design of this code is simple since for the evaluation of a field we associate one thread of a GPU with one point in the grid.…”