Accelerating Financial Code through Parallelisation and Source-Level Optimisation
نویسنده
چکیده
In this paper we summarise the experiences we obtained during past years in accelerating financial code through parallelisation and source-level optimisation. We have been focusing on developing optimised parallel programs to speedup financial computations where either binomial tree method or Monte Carlo simulation was applicable. The parallelisation was through explicit POSIX multi-threading on x86 sharedmemory multi-processor systems. The source-level optimisations we found most useful were data structure optimisation and elimination of common sub-expressions.
منابع مشابه
Compilation techniques for high-performance embedded systems with multiple processors
Despite the progress made in developing more advanced compilers for embedded systems, programming of embedded high-performance computing systems based on Digital Signal Processors (DSPs) is still a highly skilled manual task. This is true for single-processor systems, and even more for embedded systems based on multiple DSPs. Compilers often fail to optimise existing DSP codes written in C due ...
متن کاملStrategy Preserving Compilation for Parallel Functional Code
Graphics Processing Units (GPUs) and other parallel devices are widely available and have the potential for accelerating a wide class of algorithms. However, expert programming skills are required to achieving maximum performance. These devices expose low-level hardware details through imperative programming interfaces where programmers explicity encode device-specific optimisation strategies. ...
متن کاملResponse to the comments by Bernaschi et al. on 'Optimisation and parallelisation strategies for Monte Carlo simulation of HIV infection'
In recent years, the study of immune response behaviour through mathematical and computational models has attracted considerable efforts. The dynamics of key cell types, and their interactions, has been a primary focus in terms of building a picture of how the immune system responds to a threat. Discrete methods, based on lattice Monte-Carlo (MC) models, with their flexibility and relative simp...
متن کاملAccelerating a C++ Image Processing Library with a GPU
This paper presents work-in-progress towards a C++ source-to-source translator that automatically seeks parallelisable code fragments and replaces them with code for a graphics co-processor. We report on our experience with accelerating an industrial image processing library. To increase the effectiveness of our approach, we exploit some domain-specific knowledge of the library’s semantics. We ...
متن کاملA Guidance Tool to Improve Memory Locality Reuse and To Exploit Hidden Parallelism in Loop Nests
In recent years, methods for analyzing and parallelizing sequential code using dependence analysis and loop transformations have been developed. These techniques have proved successful, and have been used either to move from sequential to parallel codes, or to improve the e ciency of existing parallel codes. Our research focuses on Fortran code optimisation for parallelisation in Shared Memory ...
متن کامل