Supercomputing in the biological sciences: Toward Zettascale and Yottascale simulations.

Area of Science:

• Computational biology
• Molecular dynamics
• Biophysics

Background:

• Molecular simulations in biology are highly compute-intensive due to electrostatic forces and extensive time steps.
• Current state-of-the-art simulations reach microseconds to milliseconds, insufficient for physiological processes.
• Simulating biological systems at relevant physiological timescales (seconds to days) remains a significant computational challenge.

Purpose of the Study:

• To highlight the computational demands of molecular simulations in biological systems.
• To underscore the gap between current simulation capabilities and physiologically relevant timescales.
• To emphasize the potential for biological sciences to drive advancements in computing power.

Main Methods:

• Analysis of computational requirements for molecular dynamics simulations.
• Comparison of simulation timescales with physiological timescales.
• Illustration of computational needs using an exascale supercomputer example.

Main Results:

• Biological simulations are orders of magnitude more demanding than those in materials science or astrophysics.
• Even exascale computers are vastly underpowered for simulating biological processes over physiological durations.
• A 10 billion-fold increase in speed is needed to simulate a virus for 3 hours.

Conclusions:

• There is an insatiable need for increased computing power in molecular simulations of biological systems.
• The growing field of computational drug design positions biological sciences as a key driver for future computing advancements.
• Bridging the gap between simulation and physiological timescales is critical for biological and pharmaceutical research.