Abstract Using 1995 results from the NASA Parallel Benchmarks, this paper compares the sustained performance (as opposed to peak performance) and cost per unit of sustained performance for state-of-the--art Parallel p...Abstract Using 1995 results from the NASA Parallel Benchmarks, this paper compares the sustained performance (as opposed to peak performance) and cost per unit of sustained performance for state-of-the--art Parallel processing systems that use MicroProcessors(MPP) versus those that use Custom-designed Processors (PVP). This comparison shows that MPPs approach, but seldom surpass, PCPs with respect to either metric. This paper also discusses the merits and demerits of 'clusters' of workstations Ironically, recently introduced low--cost, parallel systems can also be viewed as 'virtual workstation clusters' that offer comparable, if not superior, capabilitics. Finally, this paper explores the question of whether it will be economically feasible for computer manut'acturers to continue building PVPs, or are MPPs the only economically viable option for building parallel systems?展开更多
In wind tunnel tests of tall building models comprised of rounded or streamlined surfaces, Reynolds Number Effects (RNE) may be remarkable. In most researches and projects, owing to the presence of sharp edged corners...In wind tunnel tests of tall building models comprised of rounded or streamlined surfaces, Reynolds Number Effects (RNE) may be remarkable. In most researches and projects, owing to the presence of sharp edged corners at model, RNE was often treated as negligible, and data colleCted in subcritical flow regime were applied directly to designs which were sure to be supercritical without any modification. But it was proved necessary to take RNE into consideration when a model had predominant rounded smooth surface(s). Some research activities were devoted to the mentioned condition and two experimental techniques accompanied by two wind tunnel model tests were introduced in the article. The authors also presented some amusing phenomenon such as extremely low pressure coefficients caused by the separating bubbles when now past sharp edged corners, unsymmetrical pressure coefficient distributions commenced by biased gap f'low (seen Figs. 3, 4) while models were placed symmetrically in the approaching flow.展开更多
文摘Abstract Using 1995 results from the NASA Parallel Benchmarks, this paper compares the sustained performance (as opposed to peak performance) and cost per unit of sustained performance for state-of-the--art Parallel processing systems that use MicroProcessors(MPP) versus those that use Custom-designed Processors (PVP). This comparison shows that MPPs approach, but seldom surpass, PCPs with respect to either metric. This paper also discusses the merits and demerits of 'clusters' of workstations Ironically, recently introduced low--cost, parallel systems can also be viewed as 'virtual workstation clusters' that offer comparable, if not superior, capabilitics. Finally, this paper explores the question of whether it will be economically feasible for computer manut'acturers to continue building PVPs, or are MPPs the only economically viable option for building parallel systems?
文摘In wind tunnel tests of tall building models comprised of rounded or streamlined surfaces, Reynolds Number Effects (RNE) may be remarkable. In most researches and projects, owing to the presence of sharp edged corners at model, RNE was often treated as negligible, and data colleCted in subcritical flow regime were applied directly to designs which were sure to be supercritical without any modification. But it was proved necessary to take RNE into consideration when a model had predominant rounded smooth surface(s). Some research activities were devoted to the mentioned condition and two experimental techniques accompanied by two wind tunnel model tests were introduced in the article. The authors also presented some amusing phenomenon such as extremely low pressure coefficients caused by the separating bubbles when now past sharp edged corners, unsymmetrical pressure coefficient distributions commenced by biased gap f'low (seen Figs. 3, 4) while models were placed symmetrically in the approaching flow.
