Energy Efficient Acceleration and Evaluation of Financial Computations towards Real-Time Pricing
Energy Efficient Acceleration and Evaluation of Financial Computations towards Real-Time Pricing
复制标题
实时定价金融计算的节能加速和评估
DOI：
--
复制
发表时间：
2011
期刊：
影响因子：
--
通讯作者：
R. Korn
中科院分区：
文献类型：
--
作者：
C. D. Schryver;Matthias Jung;N. Wehn;H. Marxen;A. Kostiuk;R. Korn
现代金融市场前所未有地活跃。如今，资产价格--以及所有相关金融产品的价格--在几毫秒内就会发生变化。然而，不仅仅是由于2008年的金融危机，为这些产品计算公平和有意义的价格比过去重要得多。为了获得可靠的价格，必须使用复杂的模拟模型。这些模型的定价一般具有非常高的计算复杂性，并且在许多情况下只能通过使用数值方法近似完成。另一方面，我们都知道能源成本在未来将变得越来越重要。增加的计算复杂度和消耗的能量之间的差距只能通过使用更定制的计算引擎来弥合，如专用硬件加速器或专用指令集处理器（ASIP）。在本文中，我们提出了一个全面的方法，最佳硬件加速器的有效设计和评估。我们给出两个案例研究：一个新的硬件随机数发生器的任意分布和一个专用的硬件加速器计算欧洲障碍期权价格。
Modern financial markets are as vivid as never before. Asset prices - and therefore the prices of all related financial products - change within several milliseconds nowadays. However, not only due to the financial crisis in 2008, calculating fair and meaningful prices for these products is much more important than in the past. In order to obtain reliable prices, sophisticated simulation models have to be used. Pricing in these models in general has a very high computational complexity and can in many cases only be approximately done by using numerical methods. On the other hand, we all know that energy costs will become more and more significant in the future. The gap between the increasing computational complexity and the consumed energy can only be bridged by using more tailored computation engines, like dedicated hardware accelerators or application specific instruction set processors (ASIPs). In this paper we present a comprehensive methodology for the efficient design of optimal hardware accelerators and the evaluation thereof. We give two case studies: a new hardware random number generator for arbitrary distributions and a dedicated hardware accelerator for calculating European barrier option prices.
影响因子：
2.7
作者：
Giles, Michael B.
通讯作者：
Giles, Michael B.