| GPU下的二叉树定价模型 |
| |
| 引用本文： | 桂叶晨,刘涛,彭蓉,文高进.GPU下的二叉树定价模型[J].华中科技大学学报(自然科学版),2011,39(Z1):102-105. |
| |
| 作者姓名： | 桂叶晨 刘涛 彭蓉 文高进 |
| |
| 作者单位： | 中国科学院深圳先进技术研究院先进计算与数字工程研究所,广东深圳,518055 |
| |
| 基金项目： | 国家高技术研究发展计划资助项目(2006AA01A114,2007AA120502); 深圳市科技计划资助项目(SY200806300211A) |
| |
| 摘 要： | 针对二叉树定价模型算法计算量大、耗时长的问题,利用CUDA架构对该模型进行了并行改造,在GPU上对该算法进行了加速测试与相应的性能分析.实验结果表明:在单个GTX295节点上,对于数据规模为16 K的期权,GPU相对于其4核Xeon E5520的加速比已达约200倍,GPU每s所能处理的期权数量达到了24 852个,符合实时商业引擎的要求.另外,通过衡量一些关键指标,考察了该算法在GPU上的扩展性及其计算精度对结果的影响.
|
| 关 键 词： | 二叉树模型 期权定价 统一计算架构 图形处理单元 加速比 |
GPU implementation of binomial tree model for option pricing |
| |
| Gui Yechen,Liu Tao,Peng Rong,Wen Gaojin.GPU implementation of binomial tree model for option pricing[J].JOURNAL OF HUAZHONG UNIVERSITY OF SCIENCE AND TECHNOLOGY.NATURE SCIENCE,2011,39(Z1):102-105. |
| |
| Authors: | Gui Yechen Liu Tao Peng Rong Wen Gaojin |
| |
| Institution: | Gui Yechen Liu Tao Peng Rong Wen Gaojin(Institute of Advanced Computing and Digital Engineering,Shenzhen Institutes of Advanced Technology,Chinese Academy of Science,Shenzhen 518055,Guangdong China) |
| |
| Abstract: | As the binomial tree model involves constructing a new tree and then working backward through it,and thus is time-consuming.To overcome these problem,the binomial tree model was parallelized according to the CUDA(compute unified device architecture) model so that it could be accelerated on GPU(graphic processing unit).Our results on single GTX 295 GPU chip and 4-core Xeon E5520 chips shows that,for an volume option-data set of 16 K combined with 512 time steps for each option pricing,GPU implementation with... |
| |
| Keywords: | binomial tree model option pricing CUDA GPU speedup ratio |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
