An extended set of FORTRAN basic linear algebra subprograms
ACM Transactions on Mathematical Software (TOMS)
ACM Transactions on Mathematical Software (TOMS)
Improving performance of linear algebra algorithms for dense matrices, using algorithmic prefetch
IBM Journal of Research and Development
Techniques for Optimizing Applications: High Performance Computing
Techniques for Optimizing Applications: High Performance Computing
An updated set of basic linear algebra subprograms (BLAS)
ACM Transactions on Mathematical Software (TOMS)
Architecture, algorithms and applications for future generation supercomputers
FRONTIERS '96 Proceedings of the 6th Symposium on the Frontiers of Massively Parallel Computation
Performance Evaluation of Parallel Algorithms for Pricing Multidimensional Financial Derivatives
ICPPW '02 Proceedings of the 2002 International Conference on Parallel Processing Workshops
Architecture independent parallel binomial tree option price valuations
Parallel Computing
Computer Architecture, Fourth Edition: A Quantitative Approach
Computer Architecture, Fourth Edition: A Quantitative Approach
Parallel binomial valuation of american options with proportional transaction costs
APPT'11 Proceedings of the 9th international conference on Advanced parallel processing technologies
| Hi-index | 0.01 |
An option contract is a financial instrument that gives right to its holder to buy or sell a financial asset at a specified price, referred to as strike price, on or before the expiry date. Determining the value of an option contract with high accuracy is a computationally intensive task. Earlier implementations of binomial model on a parallel computer have a big gap between the realized performance and the peak performance of the parallel computer. This is mainly due to the implementation not considering the memory hierarchy available in today's computers. We propose two algorithms based on a hierarchical model of memory that maximize locality for data access. We implement these algorithms on a single processor and a shared memory multiprocessor. The proposed algorithms outperform the earlier reported algorithms by a factor of 20 on uniprocessor; and the speedup varies from 5 to 7.4 on a Sun SMP.