Research in models for massively parallel computing
Workstation-Based Supercomputing: Supercomputing in the DoE has been traditionally performed on von Neumann vector pipeline machines. In the past few years, a belief has developed that advances in supercomputing capability in the future must come from massively parallel architectures. When considering such architectures, many questions remain which are active areas of research. Examples of research subjects for massively parallel computing paradigms are communication structures, memory structures, node processor architectures, latency hiding, algorithm development, switching schemes, communication bandwidth and support software. Gigabit Networks: A problem which has existed in computing since networks were invented is that the quality of service received by those using a computing resource is inversely proportional to the user`s physical distance from the server. The goal of the CASA Gigabit project is to remove this restriction by inventing gigabit speed networks with support protocols, switches, etc. which will operate at global distances.
- Research Organization:
- Los Alamos National Lab., NM (United States)
- Sponsoring Organization:
- USDOE, Washington, DC (United States)
- DOE Contract Number:
- W-7405-ENG-36
- OSTI ID:
- 10112041
- Report Number(s):
- LA-UR-92-3495; CONF-9210229-1; ON: DE93003748
- Resource Relation:
- Conference: IEEE scientific supercomputing subcommittee meeting,Livermore, CA (United States),25 Oct 1992; Other Information: PBD: 21 Oct 1992
- Country of Publication:
- United States
- Language:
- English
Similar Records
QsNetII : Defining High-Performance Network Design
Evaluation of DEC`s GIGAswitch for distributed parallel computing