SAL Home Parallel Computing

[an error occurred while processing this directive]

Parallel computing extends to systems with more processors to obtain speedup in code execution. The efficiency and effectiveness of the parallelism are largely dependent on the problems to be solved with selected algorithms and hardware architectures. Based on load distribution and synchronization implementations, SIMD, MIMD, SPMD, and many other "fashions" have been categorized. Recently, multi-threading programming on SMP (Symmetrical Multiprocessor) architectures and message-passing programming on distributed memory systems (or even clustering systems) are becoming more and more popular. After all, scientific applications are still the major motivation behind the development of parallel computing technologies.

Search SAL:
Parallel Computing

Programming Languages & Systems   Charm/Charm++, CODE, Erlang, HPF, uC++ ...
Communication Libraries   BSPlib, LinuxThreads, MPI, Para++, PVM, TreadMarks ...
Tools & Utilities   NetSolve, PETSc, PIM, PLAPACK, ScaLAPACK ...

Other Resources:

High Performance Cluster Computing Resources

Beowulf Underground -- central source of information on Beowulf Computing.

PAPERS -- custom hardware that allows a cluster of unmodified PCs and/or workstations to function as a fine-grain parallel computer capable of MIMD, SIMD, and VLIW execution.

Avalon -- T-CNLS DEC/Alpha Cluster running Linux.

Extreme Linux

The Kluster Project -- to design and manufacture clusters of workstations for a broad range of applications.

xtreme MACHINES -- Commodity Supercomputing

SCL Cluster Cookbook

The Cluster Cookbook Home Page from The Portland Group, Inc. (PGI)

Linux Parallel Processing HOWTO: This document provides an overview of how to use multiple processors to speed-up execution of an individual program. Use of SMP Linux systems, clusters of networked Linux systems, and attached (parallel) processors hosted by Linux, are all discussed.

Parallel Processing using Linux:

NIH LoBoS Supercomputer Home Page --- an implementation of a Beowulf class computer, using Linux and 10,100 Gigabit Ethernet Drivers.

The SHRIMP (Scalable, High-Performance, Really Inexpensive Multi-Processor) Project. The computing nodes of SHRIMP are Pentium PCs (Linux), and the routing network is the same one used in the Intel Paragon.


The HERMES Linux Production Cluster

Internet Parallel Computing Archive

Parallel I/O Archive. This archive contains information about parallel I/O research. Most of the coverage relates to parallel file I/O.

Parallel Tools Projects Around the World: This page points to descriptions of parallel tool projects around the world.

Supercomputing and Parallel Computing Research Groups

High Performance Computing Archive

Concurrent Systems

A Comprehensive Bibliography of Distributed Shared Memory

ParaScope: A Listing of Parallel Computing Sites

Yahoo Links:

  SAL Home   |   Numerical Analysis   |   Scientific Data Processing & Visualization

Copyright © 1995-2001 by Herng-Jeng Jou
Copyright © 1997-2001 by Kachina Technologies, Inc.
All rights reserved.