Multiprocessor supercomputers for scientific/engineering applications
Kai Hwang
Abstract
Kai Hwang
Abstract
Supercomputers, the fastest computers available, are playing a major role in advancing human civilization. Based on today's technology, a computer is considered a supermachine if it can perform hundreds of milllions of floating-point operations per second (100 M-Flops) with a word length of approximately 64 bits and a main memory capacity of millions of words. The value of supercomputers has been identified in three general areas: (1) supercomputers aid in knowledge acquisition by processing complex data within a reasonable time. (2) supercomputers provide computational tractability in instrumentation, predictive simulations, energy resource exploration, general circulation modeling, weapons effects, and atmospheric testing. (3) supercomputers promote productivity, as in system optimization, computeraided design and manufacturing, and VLSI circuit design. The author briefly reviews the evolution of supercomputers, then discusses supercomputing in science and engineering. He considers multiprocessing supercomputers and exploratory research multiprocessors and finally considers future directions.
OpenAlex reports 36 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Supercomputers, the fastest computers available, are playing a major role in advancing human civilization. Based on today's technology, a computer is considered a supermachine if it can perform hundreds of milllions of floating-point operations per second (100 M-Flops) with a word length of approximately 64 bits and a main memory capacity of millions of words. The value of supercomputers has been identified in three general areas: (1) supercomputers aid in knowledge acquisition by processing complex data within a reasonable time. (2) supercomputers provide computational tractability in instrumentation, predictive simulations, energy resource exploration, general circulation modeling, weapons effects, and atmospheric testing. (3) supercomputers promote productivity, as in system optimization, computeraided design and manufacturing, and VLSI circuit design. The author briefly reviews the evolution of supercomputers, then discusses supercomputing in science and engineering. He considers multiprocessing supercomputers and exploratory research multiprocessors and finally considers future directions.
Key concepts: Computer science, Multiprocessing, Parallel computing, Supercomputer, Computer architecture, Software engineering, Operating system