J. Merlin, S. Baden, S. Fink, and B. Chapman. Multiple Data Parallelism with HPF and KeLP. Future Generation Computing Systems, 15(3):393--405, May 1999.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....and communication among tasks. Linda is described in detail below. Therefore, HPF s tasking features are mainly suitable for algorithms which do not require intertask communication, such as divide and conquer. Several proposals for more advanced tasking facilities in HPF have been made [33, 51, 74, 84, 96, 139, 157]. Inter process communication in HPF is due to non local data accesses. An HPF compiler is in duty of generating appropriate code, in most cases message passing code, that takes care of providing non local data in place. All communication is handled transparently to the user. Linda [3, 45] is a ....

....In particular, the KeLP programmer may plug in existing serial numerical kernels that have known numerical properties and leverage mature compiler technology. While KeLP was originally designed for coordination and communication among sequential computations, an HPF interface (KeLP HPF [139]) has been developed recently, which allows HPF tasks to be invoked from within KeLP, but not vice versa. Moreover, support for clusters of SMPs is provided within KeLP2 [8] Heterogeneity. KeLP is based upon MPI and hence heterogeneous platforms providing a common MPI abstract machine may be ....

J. Merlin, S. Baden, S. Fink, and B. Chapman. Multiple Data Parallelism with HPF and KeLP. Future Generation Computing Systems, 15(3):393--405, May 1999.

....in addition, more complex animations can be obtained using the provided high level animation functions. In this way our system is able to deliver animated parallel algorithms to a group of users without comprising their original implementation. We have experimented our system with the Kelp[7] library, Kelp provides an MPI based implementation for several distributed data structures like n dimensional array. As a preliminary result we have built the animated version of the simplest but also most used distributed data structure, namely the vector. One of our future goals is to extend ....

J. H. Merlin, S.B. Baden, S.J. Fink and B.M. Chapman."Multiple data parallelism with HPF and KeLP ". HPCN 98 Amsterdam, April 1998. 3

....was originally introduced to support translation of a kind of nested parallelism that can occur in HPF programs when (for example) some intrinsically parallel Fortran 90 array expressions appear nested inside a forall construct. It has been used in practise to support multiple data parallelism [11], and also in work on a parallel database classification algorithm using a divide and conquer style of parallelism, combined with data parallelism [7] The ARMCI functions ARMCI Malloc, ARMCI free allocate raw shared or remotely accessible memory. In the original definition they were collective ....

J. H. Merlin, S. B. Baden, S. J. Fink, and B. M. Chapman. Multiple data parallelism with HPF and KeLP. In High-Performance Computing and Networking: Proceedings of HPCN Europe 1998, volume 1401 of Lecture Notes in Computer Science, pages 828--839. Springer, 1998.

....efficient sequential and parallel execution; the programmer indicates the potential for concurrency, and the system optimizes parallel granularity for the underlying machine. Both task parallelism and data parallelsim are supported. KeLP. The Kernel Lattice Parallelism (KeLP) programming system [35] is the successor of LPARX, both developed at the University of California, San Diego. KeLP is a software tool for efficiently implementing multi level and adaptive finite difference methods on distributed memory MIMD architectures. It is a domain specific C class library that supports dynamic ....

S. J. Fink J. H. Merlin, S. B. Baden and B. M. Chapman. Multiple data parallelism with HPF and KeLP. In Proc. HPCN '98, Amsterdam, 1998.

....provides more general, lower level primitives than KeLP, such as reductions and gather scatter. But, it does not help with data decomposition nor does it overlap communication with computation. Merlin et al. have successfully incorporated KeLP with SHPF, a data parallel HPF like Fortran dialect [38]. The NESL language [39] implements nested data parallelism, a model which supports hierarchical parallelism and data structures through vectors of vectors. NESL is an applicative language, and provides no constructs to control data decomposition or granularity of parallelism. Several ....

J.H. Merlin, S.B. Baden, S.J. Fink, and B.M. Chapman, "Multiple data parallelism with HPF and KeLP," Proc. HPCN '98, Apr. 1998.

....provides more general, lower level primitives than KeLP, such as reductions and gather scatter, but it does not help with data decomposition nor does it overlap communication with computation. Merlin et al. have successfully incorporated KeLP with SHPF, a data parallel HPF like Fortran dialect [38]. The NESL language [39] implements nested data parallelism, a model which supports hierarchical parallelism and data structures through vectors of vectors. NESL is an applicative language, and provides no constructs to control data decomposition or granularity of parallelism. Several ....

J. H. Merlin, S. B. Baden, S. J. Fink, and B. M. Chapman, "Multiple data parallelism with HPF and KeLP," in Proc. HPCN '98, (Amsterdam, Netherlands), Apr. 1998.

....having the run time system transparently interact with the operating system page coloring policy [37] Finally, interoperability with other programming models remains an outstanding KeLP issue. Merlin et al. have successfully incorporated KeLP with SHPF, a data parallel HPF like Fortran dialect [101]. Interaction with other systems, such as PETSc [16] and task oriented systems [74, 66] remains an unresolved issue. Chapter 3 Implementation 3.1 Introduction KeLP aims to provide a convenient programming model to express high performance block structured scientific calculations on SMP ....

....of Chemistry, UCSD) collaborating with Scott Baden, is using KeLP to parallelize a Fourier based planewave basis code for materials science applications. ffl John Merlin (Vienna Center for Parallel Computation) has implemented an interface between KeLP and the SHPF data parallel language [101]. This project combines the independent strengths of KeLP and HPF to provide a powerful programming model for multi block applications. ffl Michelle Ketcham (Department of Applied and Mechanical Engineering, UCSD) has developed a parallel version of the MIXLAYER code using KeLP [89] This program ....

J. H. Merlin, S. B. Baden, S. J. Fink, and B. M. Chapman. Multiple data parallelism with HPF and KeLP. In Proceedings HPCN '98, Amsterdam, Netherlands, April 1998.

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