| Issue |
2014
SNA + MC 2013 - Joint International Conference on Supercomputing in Nuclear Applications + Monte Carlo
|
|
|---|---|---|
| Article Number | 04203 | |
| Number of page(s) | 7 | |
| Section | 4. Advanced Parallelism and HPC Strategies: b. Monte Carlo Methods, Parallelism and HPC | |
| DOI | https://doi.org/10.1051/snamc/201404203 | |
| Published online | 06 June 2014 | |
Thread Divergence and Photon Transport on the GPU (U)
LA-UR-13-27057
Los Alamos National Laboratory, P.O. Box 1663, Los Alamos, New Mexico 87545
* Corresponding Author, E-mail: rta@lanl.gov
Monte Carlo methods are commonly used to solve numerically the particle transport problems. A major disadvantage to Monte Carlo methods is the time required to obtain accurate solutions. Graphical Processing Units (GPUs) have increased in use as accelerators for improving performance in high-performance computing. Extracting the best performance from GPUs places requires careful consideration on code execution and data movement. In particular, performance can be reduced if threads diverge due to branching, and Monte Carlo codes are susceptible to branching penalties. We explore different schemes to reduce thread divergence in photonics transport and report on our performance findings.
Key words: Monte Carlo / Photon Transport / GPU / MCATK / CUDA
© Owned by the authors, published by EDP Sciences, 2014