A parellising programming language that enables tasks to be distributed over multiple DEC VAX computers to simulate a parallel processing supercomputer will soon be available from Scientific Computing Associates, New Haven, Connecticut. The language, called Linda, was developed by researchers working at Yale University in New Haven, in partnership with engineers from the Sandia National Laboratory. Use of Linda is claimed to have enabled a scientist at the Sandia laboratory to create a complex simulation of a rocket exhaust plume using a network of 14 VAX computers at two locations more than 1,000 miles apart, and the simulation took 143 minutes to produce, half the time it took on the lab’s ageing Cray-1 supercomputer. The developers are promoting the idea that where multiple similar machines, such as VAXes, are installed in different departments of an organisation, they should be kept busy for a much higher proportion of the time by harnessing their spare moments to participate in co-operative parallel processing tasks. United Press International reports that researchers have begun developing applications for the networked VAXes, including a monitoring system for patients in cardiac intensive care units and faster searches through huge computer databases. As it was developed with government funding, Linda is not copyrightable, and is at least theoretically in the public domain to any US company.
