test Browse by Author Names Browse by Titles of Works Browse by Subjects of Works Browse by Issue Dates of Works

Advanced Search
& Collections
Issue Date   
Sign on to:   
Receive email
My Account
authorized users
Edit Profile   
About T-Space   

T-Space at The University of Toronto Libraries >
School of Graduate Studies - Theses >
Master >

Please use this identifier to cite or link to this item: http://hdl.handle.net/1807/18317

Title: Accelerating an Analytical Approach to Collateralized Debt Obligation Pricing
Authors: Gupta, Dharmendra
Advisor: Chow, Paul
Department: Electrical and Computer Engineering
Keywords: FPGA Application
Credit Derivative
Issue Date: 19-Jan-2010
Abstract: In recent years, financial simulations have gotten computationally intensive due to larger portfolio sizes, and an increased demand to perform real-time risk analysis. In this paper, we propose a hardware implementation that uses a recursive analytical method to price the Collateralized Debt Obligations. A novel convolution approach based on FIFOs for storage is implemented for the recursive convolution. It is also used to address one of the main drawbacks of the analytical approach. The FIFO-based convolution approach is compared against two different convolution approaches outperforming them with a much smaller memory usage. The CDO core designed with the FIFO-based convolution method is implemented and tested on a Virtex-5 FPGA and compared against a C implementation, running on a 2.8GHz Intel Processor, resulting in a 41-fold speed up. A brief comparison against a Monte Carlo based hardware implementation for structured instruments yields mixed results.
URI: http://hdl.handle.net/1807/18317
Appears in Collections:Master
The Edward S. Rogers Sr. Department of Electrical & Computer Engineering - Master theses

Files in This Item:

File Description SizeFormat
Gupta_Dharmendra_P_200911_MASc_thesis.pdf1.97 MBAdobe PDF

This item is licensed under a Creative Commons License
Creative Commons

Items in T-Space are protected by copyright, with all rights reserved, unless otherwise indicated.