𝔖 Bobbio Scriptorium
✦   LIBER   ✦

Special Issue on Workstation Clusters and Network-Based Computing

✍ Scribed by Dhabaleswar K. Panda; Lionel M. Ni


Publisher
Elsevier Science
Year
1997
Tongue
English
Weight
107 KB
Volume
40
Category
Article
ISSN
0743-7315

No coin nor oath required. For personal study only.

✦ Synopsis


makes the problem much harder. In order to build costeffective NOW systems which can deliver MPP-like performance, several important problems need to be solved. Examples of such problems include designing efficient messaging layers; network interface design for low-latency and high-throughput communication; architectural and operating systems support for fast communication and synchronization; collective communication support; distributed memory implementation with PVM and MPI; memory models and architectural supports for implementing distributed shared memory paradigm; programming environments; load balancing; scheduling; fault tolerance; run time support; and performance evaluation.

In this special issue we have attempted to include some of the state-of-the-art research results, which lead the parallel processing community towards the realization of the NOW paradigm. We received 53 manuscripts in response to our call for papers. Due to many high quality submissions, we had to go through an extremely difficult selection process. We are spreading out the selected papers over two issues. This first issue, with the theme of hardware and software, contains eight regular papers and two research notes. A set of additional papers, emphasizing performance evaluation, scheduling, and fault-tolerance, is tentatively scheduled for the June 1997 issue. A brief summary of the papers appearing in this issue is as follows:

M. Lauria and A. Chien provide a new design approach for low level communication layers by providing a closer integration with the upper layer. They describe their design with respect to MPI-FM, a high performance implementation of MPI on networks of workstations connected by a Myrinet network.

J. Bruck et al. present an efficient design and implementation of the collective communication part in MPI that is optimized for clusters of workstations. They have implemented their system on a collection of IBM RS/6000 workstations connected via a 10MBit Ethernet LAN.

I. Foster et al. propose techniques that allow multiple


πŸ“œ SIMILAR VOLUMES


SCIPVM: Parallel distributed computing o
✍ Zoraja, Ivan; Hellwagner, Hermann; Sunderam, Vaidy πŸ“‚ Article πŸ“… 1999 πŸ› John Wiley and Sons 🌐 English βš– 142 KB

Workstation and PC clusters interconnected by SCI (scalable coherent interface) are very promising technologies for high-performance cluster computing. Using commercial SBus to SCI interface cards and system software and drivers, a two-workstation cluster has been constructed for initial testing and

Special issue on granular computing and
✍ Tsau Y. Lin; Lotfi A. Zadeh πŸ“‚ Article πŸ“… 2004 πŸ› John Wiley and Sons 🌐 English βš– 29 KB πŸ‘ 1 views

The program of the 2002 World Congress on Computational Intelligence (WCCI 2002) included six sessions on "Granular Computing and Data Mining," reflecting the growing interest in application of granular computing to data mining. In these sessions, a broad range of issues were addressed. This Special