Posted on 2007-04-10 10:57
kk 閱讀(743)
評(píng)論(0) 編輯 收藏 引用 所屬分類:
IT
DeinoMPI is an implementation of MPI-2 for Microsoft Windows.
DeinoMPI是微軟windows下MPI-2的一個(gè)實(shí)現(xiàn)。
注：MPI-2是Message Passing Interface, http://www.mpi-forum.org/。What is MPI? MPI is a library of functions and macros which is intended for use in programs that exploit the existence of multiple processors by message-passing. 也就是說，MPI是一套規(guī)范，定義了一些函數(shù)和宏，它被用來在程序中通過消息傳遞來發(fā)揮多處理器的處理能力。
DeinoMPI is an implementation of the MPI-2 standard for parallel computing or more generally it is system level middle-ware for high performance parallel computing on Microsoft Windows systems. It supports Win32 and Win64 machines.
DeinoMPI是MPI-2的一個(gè)實(shí)現(xiàn)，MPI-2是for并行計(jì)算的。更通用地說，DeinoMPI是windows系統(tǒng)上的一個(gè)高性能并行計(jì)算的系統(tǒng)level中間件。它有兩個(gè)版本，分別是for win32和for win64的。
1) DeinoMPI provides the libraries necessary for software developers to write parallel applications that conform to the MPI-2 standard for parallel computing. Software developers who wish to develop new parallel applications or wish to add parallel capabilities to existing software would benefit from using DeinoMPI. The libraries provided support a wide range of C, C++ and Fortran compilers.
2) DeinoMPI also provides a process manager used to start processes on multiple machines in a cluster remotely. Microsoft Windows does not provide native capability to start user applications on remote machines. DeinoMPI provides a secure mechanism to do just that. The primary purpose of the DeinoMPI process manager is to set up the environment and launch processes used in MPI jobs but it is not restricted to this functionality. It can launch any application remotely on behalf of the user.
The following users will benefit from DeinoMPI: Businesses that develop software and wish to add parallel capabilities to increase the performance of their software. Research institutions that have parallel software codes that they want to run on their Windows machines. Universities or other educational institutions that teach courses on parallel computing. University or research labs that run computationally intensive software and wish to use parallel versions of their software to better utilize their computer resources. Students or hobbyists that wish to write parallel programs.