Session: IA^3 2020: 10th Workshop on Irregular Applications: Architectures and Algorithms

Parallelizing Irregular Computations for Molecular Docking

Wednesday, November 11, 2020

11:20 AM – 11:40 AM ET

Channel: Track 4

Author/Presenter(s)

LS

Leonardo Solis-Vasquez

Technical University Darmstadt
Darmstadt, Germany
DS

Diogo Santos-Martins

Scripps Research Institute, United States of America
Andreas F. Tillack

Scripps Research, United States of America
AK

Andreas Koch

Technical University Darmstadt, Germany
JE

Jérôme Eberhardt

Scripps Research Institute
SF

Stefano Forli

Scripps Research Institute, United States of America

AUTODOCK is a molecular docking software widely used in computational drug design. Its time-consuming executions have motivated the development of AUTODOCK-GPU, an OpenCL-accelerated version that can run on GPUs and CPUs. This work discusses the development of AUTODOCK-GPU from a programming perspective, detailing how our design addresses the irregularity of AUTODOCK while pushing towards higher performance. Details on required data transformations, re-structuring of complex functionality, as well as the performance impact of
different configurations are also discussed. While AUTODOCK-GPU reaches speedup factors of 373x on a Titan V GPU and 58x on a 48-core Xeon Platinum 8175M CPU, experiments show that performance gains are highly dependent on the molecular complexity under analysis. Finally, we summarize our preliminary experiences when porting AUTODOCK onto FPGAs.