The Rise of Parallelism (and Other Computing Challenges)

International Science Grid This Week (12/12/07) El Baz, Didier Parallelism is no longer restricted to high-performance or high-speed computing, as it is used in PCs, cellular phones, and numerous other electronic devices, writes Didier El Baz, head of the Distributed Computing and Asynchronism team, LAAS-CNRS. El Baz says the arrival of grid computing and parallelism have raised numerous questions in computer science and numerical computing. The combination of parallel and distributed computing could potentially change the nature of computer science and numerical computing. To ensure efficient use of new parallel and distributed architectures, new concepts on communication, synchronization, fault tolerance, and auto-organization are needed and must be widely accepted. Manufacturers agree that future supercomputers will have massively parallel architectures that will need to be fault tolerant and well suited to dynamicity, which will require some type of auto-organization as controlling these large systems efficiently will not be possible entirely from the outside. Parallel and distributed algorithms will also have to be more adapt at coping with the asynchronous nature of communication networks and the faults in the system. These problems are attracting more and more attention, particularly from scientists working on communication libraries, and will need to be addressed to find solutions and drive the evolution of computing.

Click Here to View Full Articlehttp: http://www.isgtw.org/?pid=1000812