 |
|
White PapersWhite Paper: TCP Linda
TCP Linda is a unique programming tool that enables programmers to develop parallel versions of existing applications efficiently and quickly to run on standard distributed memory multiprocessors, clusters, or networks of computers.
Virtual Shared Memory and the Paradise System
This paper provides an example of implementing a distributed computing solution for a pricing system for a Wall Street investment firm for complex financial instruments such as fixed income derivative products.
A Hierarchical Environment for Virtual Supercomputing
A presentation-style paper that discusses strategies for assembling cost-effective High Perfomance Computing platforms using hierarchical architectures and heterogeneous components.
Paradise Improves Network Parametric Computing
Many large-scale computations require running a single application many times using different input data. this is a time-consuming process when done sequentially on a single processor.
Paradise for Use With Message Passing
Paradise Virtual Shared Memory is designed to be used in conjunction with message passing systems, such as MPI and PVM, to enhance communication capabilities.
Paradise for Programmers Using Message Passing
Paradise Virtual Shared Memory middleware is designed to be used in conjunction with message passing systems, such as MPI and PVM, to enhance communication capabilities.
Paradise for Batch Queuing Systems
Batch Queuing systems enable the users of computer clusters to submit jobs without intervention. The Paradise Virtual Shared Memory enhances batch queuing systems such as LSF and PBS to permit different applications on networks.
Using a Parallel Main Memory Database for Microarray Data
by Professor Kei-Hoi Cheung
At the Yale Center for Medical Informatics (http://ycmi.med.yale.edu), we have been developing a variety of biological databases and tools to help biomedical researchers manage and analyze their experimental data. Among these databases and tools, Yale Microarray Database is a University-wide database for archiving, managing, and analyzing microarray gene expression data on a large scale. Given the large data volume and high complexity of such data and their analysis, we run into a performance problem when using a traditional database approach (Oracle). To address this problem, we have been exploring the use of parallel main memory (PAMM) database approach (implemented using LindaTM) to augment the relational database approach in terms of performing complex queries of microarray data. We have done a benchmark comparison between the two approaches. The results suggest that a significant speedup can be gained using the PAMM approach. |
|
||
Copyright 2024 Scientific Computing Associates, Inc.