Presentation 2010/11/23
FPGA Acceleration of Default Intensity Model
Takaaki YOKOYAMA, Hideharu AMANO,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) From the financial crisis, pricing of securitized paper needs to be simulated more accurately and rapidly. In computational finance, pricing of financial product is generally simulated with Monte-Carlo method, since the pricing includes probability process. Although it is also applied to securitized paper, Monte-Carlo method requires from several tens of thousands to half a million trials and so huge simulation time. On the other hand, these simulation steps are independent each other enabling FPGA devices to perform it in parallel. In this paper, CDO pricing "default intensity model" is implemented on FPGA device. Default intensity model is the simulation that can express not only ranking, but also default call. There is no study such type of model, since it is a heavy-loaded application with a lot of parameters. Here, we implement Duffie and Garleanu's model that is one of the default intensity model on an FPGA device. In this paper, a pipeline module without stall as possible is implemented with independence of each simulation of Monte-Carlo method. By using Xilinx Virtex5VSX50T, it achieves 59.8 times speed, compared with the software execution on Xeon 3.2GHz.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) FPGA / Computational Finance / Monte-Carlo / Box-Muller / CDO / Securitized Paper
Paper # RECONF-2010-52
Date of Issue

Conference Information
Committee RECONF
Conference Date 2010/11/23(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Reconfigurable Systems (RECONF)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) FPGA Acceleration of Default Intensity Model
Sub Title (in English)
Keyword(1) FPGA
Keyword(2) Computational Finance
Keyword(3) Monte-Carlo
Keyword(4) Box-Muller
Keyword(5) CDO
Keyword(6) Securitized Paper
1st Author's Name Takaaki YOKOYAMA
1st Author's Affiliation Faculty of Science and Technology, Keio University()
2nd Author's Name Hideharu AMANO
2nd Author's Affiliation Faculty of Science and Technology, Keio University
Date 2010/11/23
Paper # RECONF-2010-52
Volume (vol) vol.110
Number (no) 319
Page pp.pp.-
#Pages 6
Date of Issue