H. Hong, A. Neubacher, and W. Schreiner. The Design of the SACLIB/ PACLIB Kernels. Journal of Symbolic Computation, 19:111--132, 1995.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....C [28] our design incorporates data types parameterized by arbitrary coefficient domains and generic algorithms such as homomorphic imaging of black boxes. Thus FOXBOX can be compiled with an imported underlying domain arithmetic, and in fact we currently plug into the arithmetic of SACLIB 1. 1 [16], GNU s MP [13] and Victor Shoup s fast modular polynomial arithmetic package NTL 2.0 [30] C template classes allow us to define a precise interface while compiling FOXBOX and the imported packages in a seamless and efficient fashion. A second methodology incorporated into FOXBOX is an ....

HONG, H., NEUBACHER, A., AND SCHREINER, W. The design of the SACLIB/PACLIB kernels. In Design and Implementation of Symbolic Computation Systems (Heidelberg, Germany, 1995), A. Miola, Ed., vol. 722 of Lect. Notes Comput. Sci., Springer Verlag, pp. 288-- 302.

....Winkler 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 y 1.5 1 0.5 0 0.5 1 1.5 x 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 y 8 6 4 2 0 2 4 6 8 x Figure 1. 1 The Tacnode Curve and its Transformed Version has evolved from our own experience in the development of parallel computer algebra environments [3] and from related work of other researchers [9, 2, 1] On the basis of Distributed Maple, we have parallelized several algorithms from algebraic geometry, mostly in the frame of the diploma thesis of the second author [4] In this paper, we present one of these parallel solutions. 2. RESOLUTION ....

....the task represented by t has terminated and returns its result. Multiple tasks may independently wait for and retrieve the result of the same task t. This parallel programming model is essentially based on functional principles which is sufficient for many kinds of computer algebra algorithms [3]. In addition, the environment supports a non deterministic form of task synchronization for speculative parallelism and shared data objects which allow tasks to communicate via a global store. The later facility was applied in the implementation of the algorithm presented in this paper for ....

H. Hong, A. Neubacher, and W. Schreiner. The Design of the SACLIB/ PACLIB Kernels. Journal of Symbolic Computation, 19:111--132, 1995.

.... to allocate heap memory simultaneously and efficiently, methods are used that do not enforce serialization by a central lock [27] Since memory is allocated by concurrent tasks at a higher total rate, also the garbage collector is parallelized in order to reclaim memory at a corresponding speed [28, 21]. On most large scale multiprocessors and computer networks, several distributed garbage collection techniques have been studied [33] The efficient use of global references relies on the ability to ensure locality of data access. This introduces an overhead due to the serialization (marshaling, ....

....safety of libraries for multi threaded environment) and to define a parallel programming. To bypass this problems first parallel computer algebra systems have been designed on the integration of a parallel runtime support into a modified computer algebra library (e.g. PARSAC [27] and PACLIB [21], respectively pD [38] for shared memory machines and SAC 2 [40] DSC [13, 7] PAC Givaro [36, 23] Sturm [22] REDUCE [32] DTS [6] CABAL [20] and CALYPSO [20] for distributed systems) The integration of complete computer algebra systems such as Mathematica or Maple faces the practical ....

H. Hong, A. Neubacher, and W. Schreiner. The design of the Saclib / Paclib kernels. In A. Miola, editor, DISCO'93, Gmunden, Austria, LNCS 722. SpringerVerlag, 1993.

....algorithms on a shared memory multi processor demonstrates the elegance and efficiency of this approach. 1. Introduction In our Ph.D. thesis (Schreiner 1994b) we have developed a para functional language pD as a high level programming interface for the parallel computer algebra kernel PACLIB (Hong et al. 1993). PACLIB represents a multiprocessor variant of the runtime system for SACLIB (Buchberger et al. 1992) a widely used C library of about 650 functions that efficiently implements many computer algebra algorithms. In pD, parallelism may be expressed by several forms of annotations; the compiler ....

....version of a PASCO 94 paper (Schreiner 1994a) 2. The Runtime System The generated target code is linked with a runtime system that provides first order tasks and automatic memory management (garbage collection) Currently, we use the PACLIB kernel that has been described elsewhere in detail (Hong et al. 1993); in this paper we only give a short sketch: The PACLIB kernel is a shared memory multi processor variant of the runtime system of the computer algebra library SACLIB developed at our institute (Buchberger et A Para Functional Programming Interface 3 al. 1992) It provides automatic memory ....

Hong, Hoon, Neubacher, Andreas, Schreiner, Wolfgang (1993). The Design of the SACLIB/PACLIB Kernels. In Miola, Alfonso, editor, DISCO '93 --- International Symposium on Design and Implementation of Symbolic Computation Systems (invited paper), volume 722 of Lecture Notes in Computer Science, Gmunden, Austria, September 15--17. Springer, Berlin.

....Section 6 presents several application results. 2 The Runtime System The generated target code is linked with a runtime system that provides first order tasks and automatic memory management (garbage collection) Currently, we use the PACLIB kernel that has been described elsewhere in detail [8, 18]; in this paper we only give a short sketch: The PACLIB kernel is a parallel variant of the runtime system of the computer algebra library SACLIB [2] for shared memory multi processors. It provides automatic memory management with an interface that is upward compatible to the SACLIB kernel; ....

H. Hong, A. Neubacher, and W. Schreiner. The Design of the SACLIB/PACLIB Kernels. In DISCO 93 --- Int. Symposium on Design and Implementation of Symbolic Computation System, vol. 722 of LNCS, Gmunden, Austria, Sept. 15--17, 1993.

No context found.

H. Hong, A. Neubacher, and W. Schreiner. The design of the SACLIB / PACLIB kernels. In A. Miola, editor, Design and Implementation of Symbolic Computation Systems, volume 722 of LNCS, pages 288--302. Springer-Verlag, 1993.

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