Pub Date : 1988-11-14DOI: 10.1109/SUPERC.1988.74146
D. A. Field, K. Yarnall
A software package for generation of tetrahedral finite-element mesh, whose kernel is a robust three-dimensional Delaunay triangulation algorithm, was ported to a CRAY X-MP for vector processing. The total execution time for the critical subroutines of the kernel decreased by a factor of six over scalar mode on the CRAY X-MP. The kernel is characterized by simple data structures and O(N/sup 2/) arithmetic operation counts, N being the number of finite-element nodes. Although the kernel is essentially a sequential algorithm, its simple data structures allow for key uses of vector processing and for streamlining sequential processing.<>
{"title":"Three dimensional Delaunay triangulation on a Cray X-MP","authors":"D. A. Field, K. Yarnall","doi":"10.1109/SUPERC.1988.74146","DOIUrl":"https://doi.org/10.1109/SUPERC.1988.74146","url":null,"abstract":"A software package for generation of tetrahedral finite-element mesh, whose kernel is a robust three-dimensional Delaunay triangulation algorithm, was ported to a CRAY X-MP for vector processing. The total execution time for the critical subroutines of the kernel decreased by a factor of six over scalar mode on the CRAY X-MP. The kernel is characterized by simple data structures and O(N/sup 2/) arithmetic operation counts, N being the number of finite-element nodes. Although the kernel is essentially a sequential algorithm, its simple data structures allow for key uses of vector processing and for streamlining sequential processing.<<ETX>>","PeriodicalId":103561,"journal":{"name":"Proceedings Supercomputing Vol.II: Science and Applications","volume":"59 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1988-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131789561","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1988-11-14DOI: 10.1109/SUPERC.1988.74129
S. Obayashi, S. Gavali
Different numerical techniques have been examined for simulating the underexpanded plumes in supersonic flows. The results revealed that the monotone differentiable limiter is superior to the TVD (total variation diminishing) minmod limiter when the MUSCL (monotonic upstream schemes for conservation laws) approach is used for Roe's upwind scheme. Roe's upwind scheme is found to be superior to the central difference scheme not only for capturing shock waves, but also for capturing shear flows.<>
{"title":"Three-dimensional simulation of underexpanded plumes using upwind algorithms","authors":"S. Obayashi, S. Gavali","doi":"10.1109/SUPERC.1988.74129","DOIUrl":"https://doi.org/10.1109/SUPERC.1988.74129","url":null,"abstract":"Different numerical techniques have been examined for simulating the underexpanded plumes in supersonic flows. The results revealed that the monotone differentiable limiter is superior to the TVD (total variation diminishing) minmod limiter when the MUSCL (monotonic upstream schemes for conservation laws) approach is used for Roe's upwind scheme. Roe's upwind scheme is found to be superior to the central difference scheme not only for capturing shock waves, but also for capturing shear flows.<<ETX>>","PeriodicalId":103561,"journal":{"name":"Proceedings Supercomputing Vol.II: Science and Applications","volume":"2010 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1988-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116947933","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1988-11-14DOI: 10.1109/SUPERC.1988.74147
E. O'Neil, C. G. Shaefer
The ARGOT strategy combines genetic algorithms with a mechanism providing a dynamically adaptive representation to form a robust optimization tool, as previously shown in the uniprocessor environment. For implementation of ARGOT on the BBN Butterfly multiprocessor, a parallel selection algorithm and a method of incremental payoff update were developed. These lead to enhanced parallelism and reduced the amount of computation needed by any genetic algorithms, including ARGOT. Experimental results on two matrix problems are presented, one a linear system from a FEM problem, and the other a nonlinear problem not well-behaved enough for consistent conjugate gradient results.<>
{"title":"The ARGOT strategy III: the BBN Butterfly multiprocessor","authors":"E. O'Neil, C. G. Shaefer","doi":"10.1109/SUPERC.1988.74147","DOIUrl":"https://doi.org/10.1109/SUPERC.1988.74147","url":null,"abstract":"The ARGOT strategy combines genetic algorithms with a mechanism providing a dynamically adaptive representation to form a robust optimization tool, as previously shown in the uniprocessor environment. For implementation of ARGOT on the BBN Butterfly multiprocessor, a parallel selection algorithm and a method of incremental payoff update were developed. These lead to enhanced parallelism and reduced the amount of computation needed by any genetic algorithms, including ARGOT. Experimental results on two matrix problems are presented, one a linear system from a FEM problem, and the other a nonlinear problem not well-behaved enough for consistent conjugate gradient results.<<ETX>>","PeriodicalId":103561,"journal":{"name":"Proceedings Supercomputing Vol.II: Science and Applications","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1988-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126791340","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1988-11-14DOI: 10.1109/SUPERC.1988.74140
A. Dubrulle
A block algorithm is presented for the Householder reduction of a matrix to Hessenberg form using the Bischof-Van Loan expression of a product of elementary matrices. Results of performance measurements on an IBM 3090 VG include a comparison with an alternate formulation considered for LAPACK. The algorithms based on the straightforward application of the Bischof-Van Loan formulations consistently appear to produce the best performance in all experiments conducted. It is likely that such behavior would be observed with other machines, but this conjecture remains to be tested. The algorithmic derivations presented are general enough to apply to other computational schemes based on similarity transformations, including those for the solution of the Hessenberg eigenvalue problem (QR).<>
{"title":"On block Householder algorithms for the reduction of a matrix to Hessenberg form","authors":"A. Dubrulle","doi":"10.1109/SUPERC.1988.74140","DOIUrl":"https://doi.org/10.1109/SUPERC.1988.74140","url":null,"abstract":"A block algorithm is presented for the Householder reduction of a matrix to Hessenberg form using the Bischof-Van Loan expression of a product of elementary matrices. Results of performance measurements on an IBM 3090 VG include a comparison with an alternate formulation considered for LAPACK. The algorithms based on the straightforward application of the Bischof-Van Loan formulations consistently appear to produce the best performance in all experiments conducted. It is likely that such behavior would be observed with other machines, but this conjecture remains to be tested. The algorithmic derivations presented are general enough to apply to other computational schemes based on similarity transformations, including those for the solution of the Hessenberg eigenvalue problem (QR).<<ETX>>","PeriodicalId":103561,"journal":{"name":"Proceedings Supercomputing Vol.II: Science and Applications","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1988-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133413941","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1988-11-14DOI: 10.1109/SUPERC.1988.74142
J. Stampen, J. R. Davis
The authors explore voting alignments in six US Senates between 1959 and 1981. The authors' thesis is that the Senate is organized by four recurring multiissue coalitions that combine in various ways to form larger voting coalitions. The four-coalition hypothesis was tested with the use of complete-link hierarchical clustering on a matrix of senators over contested bills. The significance of clustering into four rather than some other number of groups was established by computing gamma goodness-of-fit statistics for each possible clustering, a computation that was practical only with the aid of a supercomputer. Beyond the basic clustering result, descriptive statistics were used to measure membership consistency over time. Membership in the four coalitions persisted even across Senates spanning the supposed Reagan realignment. This finding and others support the conclusion that the multiissue coalition is the appropriate unit in roll-call analysis.<>
{"title":"Multi-issue coalitions in the Congress (voting alignments on supercomputer)","authors":"J. Stampen, J. R. Davis","doi":"10.1109/SUPERC.1988.74142","DOIUrl":"https://doi.org/10.1109/SUPERC.1988.74142","url":null,"abstract":"The authors explore voting alignments in six US Senates between 1959 and 1981. The authors' thesis is that the Senate is organized by four recurring multiissue coalitions that combine in various ways to form larger voting coalitions. The four-coalition hypothesis was tested with the use of complete-link hierarchical clustering on a matrix of senators over contested bills. The significance of clustering into four rather than some other number of groups was established by computing gamma goodness-of-fit statistics for each possible clustering, a computation that was practical only with the aid of a supercomputer. Beyond the basic clustering result, descriptive statistics were used to measure membership consistency over time. Membership in the four coalitions persisted even across Senates spanning the supposed Reagan realignment. This finding and others support the conclusion that the multiissue coalition is the appropriate unit in roll-call analysis.<<ETX>>","PeriodicalId":103561,"journal":{"name":"Proceedings Supercomputing Vol.II: Science and Applications","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1988-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131384857","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1988-11-14DOI: 10.1109/SUPERC.1988.74145
R. Hessel, M. Myszewski, G. Brussino, J. Swanson, L. Wagner
A solver kernel, LSOLVE, has been added to ANSYS, a finite-element analysis code. LSOLVE is a rank-n update routine that can be implemented efficiently on parallel and vector machines with a hierarchical memory structure. An implementation of LSOLVE on a parallel computer is described by examining the Fortran code used on the Alliant parallel computer. The utility of LSOLVE is demonstrated by performance on standard ANSYS benchmarks.<>
{"title":"Tuning the ANSYS kernel LSOLVE for a parallel computer","authors":"R. Hessel, M. Myszewski, G. Brussino, J. Swanson, L. Wagner","doi":"10.1109/SUPERC.1988.74145","DOIUrl":"https://doi.org/10.1109/SUPERC.1988.74145","url":null,"abstract":"A solver kernel, LSOLVE, has been added to ANSYS, a finite-element analysis code. LSOLVE is a rank-n update routine that can be implemented efficiently on parallel and vector machines with a hierarchical memory structure. An implementation of LSOLVE on a parallel computer is described by examining the Fortran code used on the Alliant parallel computer. The utility of LSOLVE is demonstrated by performance on standard ANSYS benchmarks.<<ETX>>","PeriodicalId":103561,"journal":{"name":"Proceedings Supercomputing Vol.II: Science and Applications","volume":"144 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1988-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124342708","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1988-11-14DOI: 10.1109/SUPERC.1988.74143
O.V. Burton
The author discusses the history and role of the US Census. He identifies some of the problems of handling and manipulating such large amounts of data. He describes the Supercomputer History Project, in which the speed of a supercomputer is combined with the versatility of a microcomputing environment to deal with census data.<>
{"title":"Complementary processing: a supercomputer/personal computer US census database project","authors":"O.V. Burton","doi":"10.1109/SUPERC.1988.74143","DOIUrl":"https://doi.org/10.1109/SUPERC.1988.74143","url":null,"abstract":"The author discusses the history and role of the US Census. He identifies some of the problems of handling and manipulating such large amounts of data. He describes the Supercomputer History Project, in which the speed of a supercomputer is combined with the versatility of a microcomputing environment to deal with census data.<<ETX>>","PeriodicalId":103561,"journal":{"name":"Proceedings Supercomputing Vol.II: Science and Applications","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1988-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115182326","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1988-11-14DOI: 10.1109/SUPERC.1988.74135
R. Matzner
The study of large-scale fluid or field-theoretical problems is greatly assisted by a graphical presentation to make computed results intelligible. The author discusses the minimal research tools and ideal state-of-the-art pedagogical tools for such visualization, using research on cosmic strings as an example. In carrying out this research it became necessary to view the result of 3-dimensional 64/sup 3/ field simulation. The day-to-day research used NCAR graphics programme EZISOSURE to contour the fields. Single-color frames were captured on a multitasking computer (Amiga 1000) running a Tektronix terminal emulator plus a screensave program. Multiple contours of a single frame were saved and then combined in a multicolor image; these images can be reloaded at about a 3-per-second frame rate to provide a movie capability.<>
{"title":"The interaction of cosmic strings-a study of research and teaching graphics","authors":"R. Matzner","doi":"10.1109/SUPERC.1988.74135","DOIUrl":"https://doi.org/10.1109/SUPERC.1988.74135","url":null,"abstract":"The study of large-scale fluid or field-theoretical problems is greatly assisted by a graphical presentation to make computed results intelligible. The author discusses the minimal research tools and ideal state-of-the-art pedagogical tools for such visualization, using research on cosmic strings as an example. In carrying out this research it became necessary to view the result of 3-dimensional 64/sup 3/ field simulation. The day-to-day research used NCAR graphics programme EZISOSURE to contour the fields. Single-color frames were captured on a multitasking computer (Amiga 1000) running a Tektronix terminal emulator plus a screensave program. Multiple contours of a single frame were saved and then combined in a multicolor image; these images can be reloaded at about a 3-per-second frame rate to provide a movie capability.<<ETX>>","PeriodicalId":103561,"journal":{"name":"Proceedings Supercomputing Vol.II: Science and Applications","volume":"65 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1988-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123802165","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1988-11-14DOI: 10.1109/SUPERC.1988.74133
T. Aki
Supercomputing efforts in computational fluid dynamics (CFD) are described. The environment supporting supercomputing is introduced with a brief historic retrospective on efforts to obtain supercomputation in the CFD discipline over the past decade. The present status of supercomputing efforts, including the graphics, is described. Results obtained from two studies typical of nonstationary flows are discussed. Finally, some requirements on future supercomputation and computers are noted from the author's personal point of view.<>
{"title":"Supercomputing efforts on CFD for nonstationary flow analysis at NAL","authors":"T. Aki","doi":"10.1109/SUPERC.1988.74133","DOIUrl":"https://doi.org/10.1109/SUPERC.1988.74133","url":null,"abstract":"Supercomputing efforts in computational fluid dynamics (CFD) are described. The environment supporting supercomputing is introduced with a brief historic retrospective on efforts to obtain supercomputation in the CFD discipline over the past decade. The present status of supercomputing efforts, including the graphics, is described. Results obtained from two studies typical of nonstationary flows are discussed. Finally, some requirements on future supercomputation and computers are noted from the author's personal point of view.<<ETX>>","PeriodicalId":103561,"journal":{"name":"Proceedings Supercomputing Vol.II: Science and Applications","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1988-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132723347","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1988-11-14DOI: 10.1109/SUPERC.1988.74136
D. Duke, R. Salvador, D. Sandee
A discussion is presented of multitasking various scientific programs on the ETA 10 supercomputer. The architecture of the ETA 10 is described, especially the features relevant for understanding the authors' strategy for multitasking the Ising model. This strategy will also be useful for a wide class of physical systems that have local interactions. The authors also summarize their results for solving large, dense, unsymmetric systems of linear equations, i.e. the LINPACK benchmark. They summarize the multitasking software that will soon be available for general use on the ETA 10.<>
{"title":"Multitasking the Ising model on the ETA 10 supercomputer","authors":"D. Duke, R. Salvador, D. Sandee","doi":"10.1109/SUPERC.1988.74136","DOIUrl":"https://doi.org/10.1109/SUPERC.1988.74136","url":null,"abstract":"A discussion is presented of multitasking various scientific programs on the ETA 10 supercomputer. The architecture of the ETA 10 is described, especially the features relevant for understanding the authors' strategy for multitasking the Ising model. This strategy will also be useful for a wide class of physical systems that have local interactions. The authors also summarize their results for solving large, dense, unsymmetric systems of linear equations, i.e. the LINPACK benchmark. They summarize the multitasking software that will soon be available for general use on the ETA 10.<<ETX>>","PeriodicalId":103561,"journal":{"name":"Proceedings Supercomputing Vol.II: Science and Applications","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1988-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133066704","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：