Hellmich, Stephan: GPU accelerated n-body integrators for long-term simulations of planetary systems. 2017
Inhalt
- 1 Introduction
- 1.1 Minor planets in the solar system
- 1.2 Dynamical evolution of minor planets
- 1.3 Motivation for this work
- 2 Solving the n-body problem
- 2.1 The n-body problem
- 2.2 Numerical integration
- 2.3 Tuning the method for solar system like n-body problems
- 2.3.1 Mixed Variable Symplectic method
- 2.3.2 Integrating minor bodies of infinitesimal mass
- 2.3.3 Encounters between minor bodies and planets
- 2.3.4 Limitations due to machine precision
- 2.4 The Yarkovsky Effect
- 3 General-Purpose Computation on Graphics Processing Units
- 3.1 The History of Graphics Processing Units (GPUs)
- 3.2 General-Purpose Computation on Graphics Processing Units
- 3.3 CUDA Computing architecture on Nvidia GPUs
- 4 cuSwift - a library of GPU based n-Body Integrators
- 4.1 Included Integration Methods
- 4.1.1 Wisdom-Holman-Mapping (WHM)
- CPU implementation
- GPU Implementation
- Data layout
- Memory management
- Thread divergence
- Dynamic memory reallocation
- 4.1.2 RMVS
- 4.2 Yarkovsky effect
- 5 Validation and Benchmarking
- 5.1 Test setup for WHM
- 5.2 Test setup for RMVS
- 5.2.1 Monitoring the Jacobi constant in the restricted three body problem
- 5.2.2 Creating a set of particles frequently involved in close encounters
- 5.2.3 Validating cuRMVS
- 5.3 Test setup for the Yarkovsky effect
- 5.4 Performance
- 6 Impact of the Yarkovsky effect on the Jupiter Trojan Asteroids
- 6.1 Jupiter's Trojan asteroids
- 6.2 Does the Yarkovsky force affect Jupiter Trojans?
- 6.3 Results and discussion
- 7 Future work
- 7.1 Improve and Implement more non-gravitational effects
- 7.1.1 Improve the Yarkovsky effect and include YORP Effect
- 7.1.2 Include YORP Effect
- 7.1.3 Acceleration due to cometary activity
- 7.1.4 Collisions and Fragmentation
- 7.2 Implement additional Integration Methods
- 8 Bibliography