Improving computation efficiency using input and architecture features for a virtual screening application

TL;DR

This paper presents GPU-based optimizations using input and architecture features that nearly double the performance of virtual screening in drug discovery, enabling faster candidate evaluation.

Contribution

It introduces input and architecture feature-based optimizations for GPU virtual screening, demonstrating significant performance improvements with SYCL and CUDA implementations.

Findings

Performance nearly doubled on modern GPU nodes.

SYCL implementation provides consistent benefits with CUDA.

Enhanced efficiency allows screening more candidates in less time.

Abstract

Virtual screening is an early stage of the drug discovery process that selects the most promising candidates. In the urgent computing scenario it is critical to find a solution in a short time frame. In this paper, we focus on a real-world virtual screening application to evaluate out-of-kernel optimizations, that consider input and architecture features to improve the computation efficiency on GPU. Experiment results on a modern supercomputer node show that we can almost double the performance. Moreover, we implemented the optimization using SYCL and it provides a consistent benefit with the CUDA optimization. A virtual screening campaign can use this gain in performance to increase the number of evaluated candidates, improving the probability of finding a drug.