Hardware Accelerator for Computer-Aided Drug Design

2020/07 - Present

• A hardware-accelerated approach for AutoDock Vina has been developed, leading to deployments on GPU and FPGA platforms with average speedups of 19.8x and 3.7x, respectively.
• An FPGA-based heterogeneous accelerator has been introduced for predicting GPCR ligand biological activity values. The system operates 54.5x faster than a CPU counterpart and achieves an energy efficiency that is 35.2x superior to GPU implementations.
• Publications:
  1. “Vina-FPGA-Cluster: Multi-FPGA Based Molecular Docking Tool with High-Accuracy and Multi-Level Parallelism”, IEEE Trans. BioCAS, 2024
  2. “FPGA Accelerating Multi-source Transfer Learning with GAT for Bioactivities of Ligands Targeting Orphan G Protein-coupled Receptors”, in FPL, 2023.
  3. “Vina-FPGA: A Hardware-Accelerated Molecular Docking Tool With Fixed-Point Quantization and Low-Level Parallelism,” IEEE Trans. VLSI. Syst, 2023.
  4. “Biological Activity Prediction of GPCR-targeting Ligands on Heterogeneous FPGA-based Accelerators,” in FCCM, 2022.
  5. “Accelerating AutoDock Vina with GPUs,” Molecules, 2022.