Pharmaceutical research demands fast, efficient simulations to predict how molecules interact, speeding up drug discovery. Jiqun Tu, a senior developer technology engineer at NVIDIA, and Ellery Russell, tech lead for the Desmond engine at Schrödinger, explore advanced GPU optimization techniques designed to accelerate molecular dynamics simulations.
In this NVIDIA GTC 2024 session, they present practical strategies for improving workload efficiency and throughput, giving pharmaceutical researchers the tools to enhance computational drug discovery. Building on existing CUDA workflows, they cover innovations such as CUDA Graphs, C++ coroutines, and mapped memory to overcome scaling challenges and bottlenecks.
Follow along with a PDF of the session, which equips attendees with actionable techniques to optimize performance, minimize latency, and fully harness GPU capabilities for molecular simulations. Topics include:
CUDA Graphs: How grouping kernel launches into dependency trees reduces overhead and enables more efficient execution.
GPU throughput optimization: Focus on throughput by scheduling multiple independent simulations on the same GPU to mask serial bottlenecks.
Mapped memory: Using direct memory access between host and device to eliminate data transfer delays.
C++ coroutines: Strategies to overlap computations and yield control across multiple simulations, improving GPU utilization without complex code restructuring.
FEP+ and Desmond engine performance: Case studies on how these tools are used in Schrödinger’s molecular dynamics engine, achieving up to 2.02x speedup in key workloads.
Watch the session Accelerating Drug Discovery: Optimizing Dynamic GPU Workflows with CUDA Graphs, Mapped Memory, C++ Coroutines, and More, explore more videos on NVIDIA On-Demand, and gain valuable skills and insights from industry experts by joining the NVIDIA Developer Program.
This content was partially crafted with the assistance of generative AI and LLMs. It underwent careful review and was edited by the NVIDIA Technical Blog team to ensure precision, accuracy, and quality.
