T.A. Bratvold. Determining useful parallelism in higher order functions. In Proceedings of the 4th Int. Workshop on the Parallel Implementation of Functional Languages, 92.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....structure were specified and computed sequentially. Subsequent work by various groups has been addressing the complications that arise by allowing the composition and nesting of algorithmic skeletons. In each case, the computations have been expressed using a sequential language, such as ML[Bra95, Bra92] Haskell[RS93, Rab93] common Lisp[Ble95] Hope[JS91] C [Pel93, BK96b] and Fortran[DGTJ95] The following sections describe three of the above skeleton programming languages and highlight the main features and skeletons. We focus on the languages which use an imperative base language, i.e. ....

T.A. Bratvold. Determining useful parallelism in higher order functions. In Proceedings of the 4th Int. Workshop on the Parallel Implementation of Functional Languages, 92.

....[WMM90] This report presents a similar analysis of a Hough transform based segmentation algorithm, prototyped in SML and then implemented in Occam2. The process of prototyping a single algorithm and converting this prototype into a parallel implementation has been studied in detail [Bus93, Bra92] There has been little experience, however, in combining existing implementations into an overall system. The three individual components (edge detection, feature segmentation and model matching) were combined into a complete intensity based scene interpretation programme. This was carried ....

Tore A Bratvold. Determining Useful Parallelism in Higher Order Functions. In Proceedings of the 4th International Workshop on the Parallel Implementation of Functional Languages, Aachen, Germany, pages 213--226, September 1992.

....are a natural part of functional languages and it is straight forward to reason and manage about them and their parallel equivalents, especially where parallelism is restricted to a small set of HOFs. Finally, parallelising compilers are capable of exploiting eciently their implicit parallelism [18] and they fall under the semi implicit approach for parallelising functional languages with restrictions to the available parallelism they provide. Also, Skillicorn and Talia have classi ed algorithmic skeletons as a nothing explicit and parallelism implicit parallel computation model. They also ....

....results back. The e ect is to apply the function to every data item. There are a variety of topologies for process farms including linear lists and trees of workers. However these require relatively low level intercommunication and incur overheads for intermediate message passing between workers [18, 164]. Instead, we have taken advantage of MPI s exibility to avoid commitment to a concrete topology, and have implemented a broadcast farm. The farmer communicates directly with each worker through virtual connections and we rely upon the MPI implementation on each system to allocate processes to ....

T. Bratvold. Determining useful parallelism in higher order functions. In Proceedings of the 4th International Workshop on the Parallel Implementation of Functional Languages, Aachen, Germany, pages 213-226, September 1992.

....Backus, 7] was the first to introduce this idea. After a version of this idea was taken up by Cole, and named skeletons, a large amount of work was and is being done in this area, including program transformation using higher order functions [51] 9] their implementation [57] their compilation [10]; work on increasing their range [13] their use in costing programs [16] 15] and work on generalising them [53] In imperative languages, a common and increasing popular way to express parallelism is through the use of libraries of message passing functions. MPI and PVM are examples of these. ....

T.A. Bratvold. Determining useful parallelism in higher order functions. In Proceedings of the 4th Int. Workshop on the Parallel Implementation of Functional Languages, 92.

....we know, HOFs are a natural part of functional languages and a programmer can easily manage and reason about them. This results from the fact that parallelism is restricted only to a small set of HOFs. Also, parallelising compilers are capable of exploiting efficiently their implicit parallelism [3]. The algorithmic skeletons major advantage is portability of parallel programs written using this approach. This results from the separation of meaning and behaviour for each skeleton which was identified by Darlington et al. [4] A HOF in a functional language can be used to express the ....

T. Bratvold. Determining Useful Parallelism in Higher Order Functions. In Proceedings of the 4th International Workshop on the Parallel Implementation of Functional Languages, Aachen, Germany, pages 213--226, September 1992.

....we know, HOFs are a natural part of functional languages and a programmer can easily manage and reason about them. This results from the fact that parallelism is restricted only to a small set of HOFs. Also, parallelising compilers are capable of exploiting efficiently their implicit parallelism [8]. The algorithmic skeletons major advantage is portability of parallel programs written using this approach. This results from the separation of meaning and behaviour for each skeleton which was identified by Darlington et al. [9] A HOF in a functional language can Matrix Multiplication on ....

T. Bratvold. Determining Useful Parallelism in Higher Order Functions. In Proceedings of the 4th International Workshop on the Parallel Implementation of Functional Languages, Aachen, Germany, pages 213--226, September 1992.

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