{"title":"基于www的声学匹配场处理的高性能计算支持","authors":"D. Gever, D.J. Fabozzi","doi":"10.1109/OCEANS.2001.968401","DOIUrl":null,"url":null,"abstract":"Under the sponsorship of the Center of Excellence in Research Ocean Sciences (CEROS), the Defense Advanced Research Projects Agency (DARPA), and the US Navy, Science Applications International Corporation (SAIC) has developed several ocean acoustic data modeling and analysis utilities. Execution of these programs was limited however, because the implementations (which include broad-band matched field processing) exceeded the computing capacity of desktop computers, and the interface was burdened by the management of large quantities of parameter and output files. Further, the migration of these algorithms to powerful processors resulted in additional constraints such as difficulty in resource access, user authentication, run-time security, and non-standardized interfaces. This project addressed these concerns through simultaneous development of a World Wide Web (WWW)-based interface, and the migration of the core algorithms to the Maui Supercomputer Center (MSC) and a SAIC Linux cluster. Operators can now use common gateway interface (CGI) scripts to input parameters, launch jobs, view graphical display results, and monitor job status from the massively parallel processors (MPP). After which, the run-time parameters and output files are managed via a Web-based relational database management system (RDBMS). This paper describes the implementation details of the parallel computing infrastructure to support broad-band MFP, the Web-based interface, user authentication, job launch, the RDMS, and a cost/benefit analysis of the migration to a multiuser high performance computing (HPC) environment. Also discussed is the security measures selected to insure resource and user protection. To serve as a feasibility baseline for the future adaptation of other applications and platforms, performance metrics are presented for the migration and processing on the MSC MPP system and the SAIC Linux cluster.","PeriodicalId":326183,"journal":{"name":"MTS/IEEE Oceans 2001. An Ocean Odyssey. Conference Proceedings (IEEE Cat. No.01CH37295)","volume":"184 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2001-11-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":"{\"title\":\"WWW-based high performance computing support of acoustic matched field processing\",\"authors\":\"D. Gever, D.J. Fabozzi\",\"doi\":\"10.1109/OCEANS.2001.968401\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Under the sponsorship of the Center of Excellence in Research Ocean Sciences (CEROS), the Defense Advanced Research Projects Agency (DARPA), and the US Navy, Science Applications International Corporation (SAIC) has developed several ocean acoustic data modeling and analysis utilities. Execution of these programs was limited however, because the implementations (which include broad-band matched field processing) exceeded the computing capacity of desktop computers, and the interface was burdened by the management of large quantities of parameter and output files. Further, the migration of these algorithms to powerful processors resulted in additional constraints such as difficulty in resource access, user authentication, run-time security, and non-standardized interfaces. This project addressed these concerns through simultaneous development of a World Wide Web (WWW)-based interface, and the migration of the core algorithms to the Maui Supercomputer Center (MSC) and a SAIC Linux cluster. Operators can now use common gateway interface (CGI) scripts to input parameters, launch jobs, view graphical display results, and monitor job status from the massively parallel processors (MPP). After which, the run-time parameters and output files are managed via a Web-based relational database management system (RDBMS). This paper describes the implementation details of the parallel computing infrastructure to support broad-band MFP, the Web-based interface, user authentication, job launch, the RDMS, and a cost/benefit analysis of the migration to a multiuser high performance computing (HPC) environment. Also discussed is the security measures selected to insure resource and user protection. To serve as a feasibility baseline for the future adaptation of other applications and platforms, performance metrics are presented for the migration and processing on the MSC MPP system and the SAIC Linux cluster.\",\"PeriodicalId\":326183,\"journal\":{\"name\":\"MTS/IEEE Oceans 2001. An Ocean Odyssey. Conference Proceedings (IEEE Cat. No.01CH37295)\",\"volume\":\"184 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2001-11-05\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"1\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"MTS/IEEE Oceans 2001. An Ocean Odyssey. Conference Proceedings (IEEE Cat. No.01CH37295)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/OCEANS.2001.968401\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"MTS/IEEE Oceans 2001. An Ocean Odyssey. Conference Proceedings (IEEE Cat. No.01CH37295)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/OCEANS.2001.968401","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
WWW-based high performance computing support of acoustic matched field processing
Under the sponsorship of the Center of Excellence in Research Ocean Sciences (CEROS), the Defense Advanced Research Projects Agency (DARPA), and the US Navy, Science Applications International Corporation (SAIC) has developed several ocean acoustic data modeling and analysis utilities. Execution of these programs was limited however, because the implementations (which include broad-band matched field processing) exceeded the computing capacity of desktop computers, and the interface was burdened by the management of large quantities of parameter and output files. Further, the migration of these algorithms to powerful processors resulted in additional constraints such as difficulty in resource access, user authentication, run-time security, and non-standardized interfaces. This project addressed these concerns through simultaneous development of a World Wide Web (WWW)-based interface, and the migration of the core algorithms to the Maui Supercomputer Center (MSC) and a SAIC Linux cluster. Operators can now use common gateway interface (CGI) scripts to input parameters, launch jobs, view graphical display results, and monitor job status from the massively parallel processors (MPP). After which, the run-time parameters and output files are managed via a Web-based relational database management system (RDBMS). This paper describes the implementation details of the parallel computing infrastructure to support broad-band MFP, the Web-based interface, user authentication, job launch, the RDMS, and a cost/benefit analysis of the migration to a multiuser high performance computing (HPC) environment. Also discussed is the security measures selected to insure resource and user protection. To serve as a feasibility baseline for the future adaptation of other applications and platforms, performance metrics are presented for the migration and processing on the MSC MPP system and the SAIC Linux cluster.