GPU parallel computing on detector simulation

Geant4是一个全面的粒子物理模拟包，可以跟踪粒子穿过物质并与物质相互作用。由于在不同的应用领域中需要更详细和复杂的模拟，因此对并行体系结构调整代码非常感兴趣。这项工作是核心Geant4算法的CUDA端口，并在GPU上并行跟踪许多粒子。GPU架构的功能为增强Geant4应用程序的性能提供了机会。基准测试显示，与单线程CPU版本相比，速度提高了30-40。

Geant4 is a comprehensive particle physics simulation package that tracks particles passing through and interacting with matter . As more detailed and complex simulations are required in different application domains, there is much interest in adapting the code for parallel architectures. This work is a CUDA port of the core Geant4 algorithm, and tracks many particles in parallel on the GPU. The capabilities of GPU architectures provide an opportunity to enhance the performance of Geant4 applications. Benchmark tests shows 30-40 speedup compared to the single-threaded CPU version.