Objective:

The students will be introduced to the use of cluster computers, and web service technology, to solve computationally intensive problems that arise in financial modeling and quantitative risk management. Included will be the use of Excel to drive the computations as well as the use of the .Net environment.

Background:

It is assumed the students are familiar with the basics of numerical computing and mathematical modeling that leads to computational problems. Familiarity with finance is also beneficial though not required. Students should have some computing and computer science background.

Schedule:

Time: 1:30pm – 5:00pm, 26th May, 2006

1.      Introduction to financial problems requiring large computational resources: Risk management, hedging, portfolio optimization, Monte Carlo simulation 2 hours

2.      (Optional with time available) Serial code optimization: profiling, timing, flops, benchmarks 30 minutes.

3.      Introduction to parallel programming (Part I): Scale-up, scale-out, network of workstations (cluster), grid computing, SIMD, MIMD, Speedup, Coarse and fine grained applications / architectures 1 hour

Time: 1:30pm – 5:00pm, 27th May, 2006

1.      Introduction to parallel programming (Part II): Scale-up, scale-out, network of workstations (cluster), grid computing, SIMD, MIMD, Speedup, Coarse and fine grained applications / architectures 2 hours

2.      Demo: Example 3- Excel front end with Visual studio tools for office, Pricing callable bonds on a cluster. With some on-site programming and benchmarking tasks. 1 hours

Time: 1:30pm – 5:00pm, 28th May, 2006

1.      Introduction to MPI- basics, point-to-point communications, collective communications 1.5 hours

2.      Demo: MPI based computation? Pricing portfolios of exotics. With some on-site programming and benchmarking tasks. 1.5 hours