Remember me

Register  |   Lost password?


Introduction to QuantLib Development - Intensive 3-day Training Course - September 10-12th, 2018 - Download Registration Form Here

 
Bookmark & Share

 

 

GPUs, Monte Carlo Simulation and Kooderive with Professor Mark Joshi

Start date

End date

February 25, 2015 February 27, 2015

Where

TBA, London [Map it]

Event price

£2300 (exc VAT)

Details

SUPPORTED BY

Nvidia

10% Discount Before November 30th

Group Discounts Also Available

Download Registration Form Here

A case study of how to obtain 140 times speed-ups over C++ code for your Monte Carlo pricing models using the open source project Kooderive and a Tesla K20 graphics card.

This is a practical course for financial market participants and limited places are available. To reserve your place and see further information on the course content please download the registration form above.

GPUs promise much in terms of raw computing power. The challenge lies in how to harness that power for computing derivatives prices. In this intensive 3-day course, we teach how to use the Kooderive code to develop millions of Monte Carlo paths for realistic complex cases in seconds. In particular, we look at how to implement the LIBOR market model and least-squares regression in a flexible efficient manner. The course demonstrates optimal design choices and how to achieve maximum performance from the GPU, while describing how these choices are expressed in the Kooderive code. At the end of the course, delegates will be familiar with the workings of Kooderive and be in a position to adapt it to their own ends.

What do you learn?

  • How to program in CUDA.
  • The differences in design philosophy between CUDA and C++.
  • How to implement highly parallel Monte Carlo simulations using GPUs.
  • How to avoid memory transfer delays.
  • The effective use of existing code in Kooderive.
  • The utilization of Thrust to minimize coding and memory issues.
  • How to implementation least-squares regressions on the GPU for pricing early exercisable derivatives.

Prerequisites

We assume that the attendee has experience of modern C++ and finance. They should have previously implemented Monte Carlo pricing models for derivatives using C++.

About the Speaker

Mark Joshi obtained a B.A. in mathematics (top of year) from the University of Oxford in 1990, and a Ph.D. in pure mathematics from the Massachusetts Institute of Technology in 1994. He was an Assistant Lecturer in the department of pure mathematics and mathematical statistics at Cambridge University from 1994 to 1999. Following which he worked for the Royal Bank of Scotland from 1999 to 2005 as a quantitative analyst at a variety of levels designing and implementing derivatives pricing models, finishing as the Head of Quantitative Research for Group Risk Management. He joined Melbourne University in November 2005 as an Associate Professor and is now a full Professor.

Mark has written several books on financial mathematics. His The Concepts and Practice of Mathematical Finance, CUP 2003 has become a standard introductory text in the area, and his programming book C++ Design Patterns and Derivatives Pricing, CUP 2004, has also proved popular. He has published twenty pure mathematics papers, as well as writing over fifty papers on financial mathematics, many of which deal with the practical aspects of implementing market models. He is a developer for the QuantLib C++ open source project and was the driving force behind its implementation of the LIBOR market model.