Direct simulation Monte Carlo: The quest for speed

Key Points

• Assess computational acceleration techniques, algorithmic optimizations, and parallel architecture strategies for Direct Simulation Monte Carlo (DSMC) methods in rarefied flow regimes.
• Evaluate algorithmic improvements in particle tracking, spatial grid indexing, and collision partner selection.
• Analyze parallel computing approaches, including domain decomposition, GPU acceleration, and hybrid continuum-kinetic frameworks.
• Algorithmic restructuring and optimized collision sampling significantly decrease per-timestep computational overhead.
• Modern parallel and GPU implementations yield orders-of-magnitude execution speedups, expanding DSMC feasibility to transitional and continuum-edge regimes.