Automatic framework to generate reconfigurable accelerators for option pricing applications
Research output: Contribution to book/conference proceedings/anthology/report › Conference contribution › Contributed › peer-review
Contributors
Abstract
Option Pricing is a fundamental application in most financial institutions dealing with derivative market. It frequently requires huge computational effort and low latency demand. Therefore, a number of different Option Pricing implementations have been developed on FPGA-based platform. However, none of the existing works cover more than one models or different types of options, which yields problem of productively implementing several hardware accelerators for different models. To fill in the gap, we propose a design flow for generating efficient hardware accelerators for option pricing applications with different models and option types. The framework boosts the designers productivity and enables quick prototyping on FPGA platform by providing general template architecture for option pricing applications. The architecture comes along with a prebuilt design library, which covers a wide range of popular financial models. Experimental results for four models show that the accelerators generated from our design flow outperform their counterpart software implementation with two order of magnitude speedup. While comparing with existing hardware designs for the same models, our framework can produce the accelerators that overcome most of manual designed engines.
Details
Original language | English |
---|---|
Title of host publication | 2016 International Conference on Reconfigurable Computing and FPGAs, ReConFig 2016 |
Editors | Peter Athanas, Rene Cumplido, Claudia Feregrino, Ron Sass |
Publisher | IEEE, New York [u. a.] |
ISBN (electronic) | 9781509037070 |
Publication status | Published - 2016 |
Peer-reviewed | Yes |
Conference
Title | International Conference on Reconfigurable Computing and FPGAs 2016 |
---|---|
Abbreviated title | ReConFig 2016 |
Duration | 30 November - 2 December 2016 |
City | Cancun |
Country | Mexico |