Inhibitors for desktop parallelisation
نویسندگان
چکیده
Parallel computing is notoriously challenging, making it difficult to develop efficient and correct programs. With the arrival of multicore processors, desktop environments must be parallelised if they are to benefit from these new processors. However, the parallelisation of desktop environments entails even more challenges than that in a typical parallel program. This report outlines such challenges, suggesting possible areas and solutions to investigate.
منابع مشابه
Parallelisation of desktop environments
This report outlines the research undertaken which is aimed at easing the development of parallel desktop applications. Developing parallel applications is notoriously difficult, but developing parallel desktop applications is further complicated largely due to the interactive nature of desktop environments. With desktop processors widely becoming multicore, parallelising desktop applications i...
متن کاملFaster Facility Location and Hierarchical Clustering
We propose several methods to speed up the facility location, and the single link and the complete link clustering algorithms. The local search algorithm for the facility location is accelerated by introducing several space partitioning methods and a parallelisation on the CPU of a standard desktop computer. The influence of the cluster size on the speedup is documented. The paper further prese...
متن کاملInteractive supercomputing
MITMatlab is a system that provides users of high performance computers with an interactive, easy-to-use environment for solving their scientific and engineering problems. It is an effort to bridge the gap between scientific computing in the desktop and supercomputer worlds by providing users of parallel machines with a tool for manipulating and visualising large datasets. We believe that the m...
متن کاملSCBI_MapReduce, a New Ruby Task-Farm Skeleton for Automated Parallelisation and Distribution in Chunks of Sequences: The Implementation of a Boosted Blast+
Current genomic analyses often require the managing and comparison of big data using desktop bioinformatic software that was not developed regarding multicore distribution.The task-farm SCBI MapReduce is intended to simplify the trivial parallelisation and distribution of new and legacy software and scripts for biologists who are interested in using computers but are not skilled programmers. In...
متن کاملLoop parallelisation using a simple parallelising compiler with the RHODOS distributed system
The procedure of parallelising a sequential task can be very difficult and even with considerable experience in the area of parallelisation, it can still be quite monotonous. Automation of parallelisation through using a parallelising compiler is often an advantageous path to pursue due to the savings in both time and effort. A major component of automatic parallelisation is loop parallelisatio...
متن کامل