No label defined (Q3729987)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 3960955
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | No label defined |
scientific article; zbMATH DE number 3960955 |
Statements
1986
0 references
Supercomputers
0 references
high-speed computation
0 references
parallelism
0 references
vector and array processing
0 references
functional and logic programming languages
0 references
high-performance architectures
0 references
0.8156175017356873
0 references
To obtain cost-effective systems assuring high-speed computation, a combination of techniques at different levels is necessary, ranging from fast circuit hardware architectures to special algorithms and programming languages. The major element for the next generation of supercomputers will be a continued evolution of architectures involving parallelism. Vector and array processing are evolutionary steps on the road of parallelism, demanding more intelligent hardware and new algorithmic structures. A direction is to extract parallelism within the classical procedural languages (like FORTRAN). Since FORTRAN reflects closely the von Neumann computer rather than the human problem solving process, functional and logic programming languages, in particular data-flow languages (like LUCID) are investigated as promising alternatives for high-performance architectures. The exploitation of these architectures through new algorithms, operating systems and compiler technology represents the key for the next generation software development. The present book is a collection of twelve interesting articles concerning the hardware organization, software problems and general characteristics of the following computers: CRAY-1, CRAY X-MP, CRAY-2 (U.S.A.): 64 bit word length, 256 Million bytes memory, 1-4 processors (CPU), 4.1 nanosec clock cycle time, UNIX Operating system; STAR-100, CYBER 200/205 (U.S.A.): 64 word length, 32-256 Mbytes, 8-10 nsec, 200-400 Million floating point oper/sec; CYBERPLUS (U.S.A.): 64 bits, 64 processors, 50 MFlops; HITACHI S-810/20 (Japan): 256 Mbytes memory, 630 MFlops; VP-400 Fujitsu (Japan): 256 Mbytes, 1.14 GFlops; NEC SX (Japan): 256 Mbytes \(+ 2\) Gbytes (ext), 1.3 GFlops; MPP (U.S.A.): 64 Mbytes, ultra-high speed immage processing. All these supercomputers are endowed with vector and/or array multiprocessing capabilities and their application software is mainly oriented on FORTRAN, PASCAL, C specialized compilers (due to their present scientific or specialized destination). They represent real, concrete steps on the wonderful (even illusive) purpose to touch the boundary between real and artificial intelligence.
0 references