Rajeev Thakur, William Gropp, and Ewing Lusk, I/O in parallel applications: The weakest link, in The International Journal of High Performance Computing Applications, Vol. 12, No. 4, Winter 1998, pp. 389-395.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....parameters are also outside the scope of b e io. 3. The major programming interface parameter is speci cation of which I O interface is used: Posix I O bu ered or raw, special lesystem I O of the vendor s lesystem, or MPI I O, which is a standard designed for high performance parallel I O [12] and therefore used in b e io. 4. MPI I O de nes the following orthogonal parameters: a) access methods, i.e. rst writing of a le, rewriting or reading, b) positioning method, c) collective or noncollective coordination, d) synchronism, i.e. blocking or not. For b e io there is no ....

Rajeev Thakur, William Gropp, and Ewing Lusk, I/O in parallel applications: The weakest link, in The International Journal of High Performance Computing Applications, Vol. 12, No. 4, Winter 1998, pp. 389-395.

....Unix like API with variations that are nonportable. Furthermore, the Unix API is not an appropriate API for parallel I O: it lacks some of the features necessary to express access patterns common in parallel programs, such as noncontiguous accesses and collective I O, resulting in poor performance [35]. To overcome these limitations, the MPI Forum defined a new API for parallel I O (commonly referred to as MPI IO) as part of the MPI 2 standard [19] MPI IO is a comprehensive API with many features intended specifically for I O parallelism, portability, and high performance. Implementations of ....

R. Thakur, E. Lusk, and W. Gropp. I/O in Parallel Applications: The Weakest Link. International Journal of High Performance Computing Applications, 12(4):389--395, Winter 1998.

....that Java offers as a programming language. To be useful as a language for high performance computing, however, Java must not only have good support for computation but must also be able to provide high performance file I O, as many scientific applications have significant I O requirements [6, 22, 34]. In this paper, we investigate the I O capabilities of Java for high performance computing and provide suggestions for relatively simple changes to the Java I O model that can improve performance significantly. We first examine several approaches that attempt to provide high performance I O in ....

....and Computer Science Division, Argonne National Laboratory, Argonne, IL 60439. thakur mcs.anl.gov 1 1. 1 I O in High Performance Computing Many computationally intensive scientific applications also need to access large amounts of data, and I O is often the bottleneck in such applications [6, 22, 34]. A common I O requirement is as follows. The application has some large data structures, say multidimensional arrays, distributed among processes in some fashion. The arrays must be read from or written to a file containing the global array. The program may begin by reading in an input array and ....

Thakur, R., Lusk, E., and W. Gropp. I/O in Parallel Applications: The Weakest Link. International Journal of High Performance Computing Applications, 124:389--395, Winter 1998.

....the Java I O model in Section 6. Related work is discussed in Section 7, followed by conclusions in Section 8. 2. I OINHIGH PERFORMANCECOMPUTING Many computationally intensive scientific applications also need to access large amounts of data, and I O is often the bottleneck in such applications [3, 13, 22]. A common I O requirement is as follows. The application has some large data structures, say multidimensional arrays, distributed among processes in some fashion. The arrays must be read from or written to a file containing the global array. The program may begin by reading in an input array and ....

Thakur, R., Lusk, E., and W. Gropp. I/O in Parallel Applications: The Weakest Link. International Journal of High Performance Computing Applications, 124:389--395, Winter 1998.

....suggestions for improving the Java I O model in Section 6. In Section 7 we discuss related work, and we conclude in Section 8. 2 I O in High Performance Computing Many large scientific applications also need to access large amounts of data, and I O is often the bottleneck in such applications [3, 13, 21]. A common I O requirement is as follows. The application has some large data structures, say multidimensional arrays, distributed among processes in some fashion. The arrays must be read from or written to a file containing the global array. The program may begin by reading in an input array and ....

Thakur, R., Lusk, E. and W. Gropp. I/O in Parallel Applications: The Weakest Link. International Journal of High Performance Computing Applications, 124:389--395, Winter 1998.

....Unix like API with variations that are nonportable. Furthermore, the Unix API is not an appropriate API for parallel I O; it lacks some of the features necessary to express access patterns common in parallel programs, such as noncontiguous accesses and collective I O, resulting in poor performance [34]. To overcome these limitations, the MPI Forum defined a new API for parallel I O (which we call MPI IO) as part of the MPI 2 standard [17] MPI IO is a comprehensive API with many features intended specifically for I O parallelism, portability, and high performance. Implementations of MPI IO, ....

R. Thakur, E. Lusk, and W. Gropp. I/O in Parallel Applications: The Weakest Link. International Journal of High Performance Computing Applications, 1998. To appear.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC