R. Hempel, The ANL/GMD Macros (PARMACS) in FORTRAN for Portable Parallel Programming Using the Message Passing Programming Model, User's Guide and Reference Manual. Gesellschaft fur Mathematik und Datenverarbeitung mbH (GMD), november 1991.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....exploiting the computing power of otherwise idling workstations. 30] estimates that 90 of Lawrence Livermore National Laboratory s workload could be solved with this technology. There is already a large number of research prototypes like PVM [15] Parform [10] Picl [14] P4 [6] Parmacs [17], Lips [25] or POSYBL [27] and there are even commercialised solutions like LINDA [1] or Express [23] Examples demonstrate, that for special applications the hypercomputer is a cheaper [22] and faster [8] choice than a traditional supercomputer, and international prizes have been awarded for the ....

R. Hempel, The ANL/GMD Macros (PARMACS) in FORTRAN for Portable Parallel Programming Using the Message Passing Programming Model, User's Guide and Reference Manual. Gesellschaft fur Mathematik und Datenverarbeitung mbH (GMD), november 1991.

....a global communication among all processes in order to compute the minimal time step and to broadcast such information to all concurrent processes. The first parallel implementation of shell force evaluation kernel it was based on the portable message passing interface PARMACS (release 5. 1) [2]. This parallel interface is particularly useful for architectures where the concept of a 2 hierarchical tree processor organization has some implications. At the contrary, in a parallel system as a cluster of workstations, the PVM message passing interface appears more adapted to an efficient ....

R. Hempel, The ANL/GMD Macros (PARMACS) in FORTRAN for Portable Parallel Programming using the Message Passing Programming Model, Technical Report, GMD, St. Augustin, Germany, November 27, 1991.

....ffl gccg uses orthomin, diagonal scaling ffl xccg uses orthomin, modified lu preconditioning For the first two solvers hand coded message passing programs based on domain decomposition were also available. Both versions were realized by using the portable message passing interface PARMACS [Hem91]. The Gauss solver uses red black relaxation. 3 Data Structures for Unstructured Grids All solvers are iterative ones that work on unstructured grids 2. While structured grids are represented by two or three dimensional arrays, unstructured grids are represented by onedimensional arrays to ....

R. Hempel. The ANL/GMD Macros (PARMACS) in FORTRAN for Portable Parallel Programming using the Message Passing Programming Model. User's guide and reference manual, Gesellschaft fur Mathematik und Datenverarbeitung mbH, November 1991.

....program. On the other hand, the PARMACS ANL GMD Macros provide a portable programming model for distributed memory architectures. They are available for a variety of architectures (Cray Y MP, Intel iPSC 2, iPSC 860, nCUBE 2, SUPRENUM, Meiko, Transputers) as well as for networks of workstations [13]. PARMACS provides portable point to point message passing between processes arranged in a host node configuration. The macros were originally based on a message passing interface developed at Argonne National Laboratory for the C programming language. Four implementations of PARMACS were used ....

R. Hempel. The ANL/GMD Macros (PARMACS) in FORTRAN for Portable Parallel Programming using the Message Passing Programming Model. User's Guide and Reference Manual. Pallas GmbH, 1991.

....successfully scaled up to 64 processors and we expect they will scale well for a larger number of processors. A. Trace Generation The procedure to be followed in order to generate the necessary trace data for a DLS is straightforward. A small trace library has been implemented on top of PARMACS [26], PVM [27] and MPI [28] both for FORTRAN and C) on the following architectures: ffl PARMACS version: CM 5, PARSYS Supernode, and MEIKO CS 1. ffl PVM version: CM 5, Meiko CS 2, and Sun workstation network. ffl MPI version: Sun workstation network. 6 PUBLISHED IN CPE, APRIL 1996. REVISED ....

....do loop. If the number of iterations is very large, the trace can be generated only for a set of iterations (e.g. each 100 iterations) in order to reduce the amount of information produced. The hardware platform consisted of a 64 T800 Parsys Supernode and the program was written in C using PARMACS [26]. A.1 Results for a 100x100 matrix on 32 processors The DLS of figure 9 represents the matrix multiplication algorithm for two matrices of dimension 100x100, running on 32 processors (initialization and terminating values have been removed in order to get a clearer picture) The values ....

R. Hempel, The ANL/GMD Macros (PARMACS) in FORTRAN for Portable Parallel Programming using the Message Passing Programming Model. User's Guide and Reference Manual, Pallas GmbH, 1991.

....recv step[ 4] 522 [ us ] 366: send 367: compstats 385: compstats 412: recv step no: 4 step min: 522 [ us ] step max: 536 [ us ] step sum: 2.113 [ us ] step mean: 528 [ us ] step med: 528 [ us ] step var: 35 [ us)2 ] Figure 2. 17: Results using the fact tool 40 than PICL (PARMACS [43] and Express [11] and implementing a version of PICL to run on transputers. For more details about the original versions of ParaGraph and PICL, see [9, 10, 44] Details about the modified versions of ParaGraph and PICL can be found in [45, 4] This section will be based on the modified version ....

....processes) and 1 in the best case (communication time) Observations on codes for other algorithms, including an LU parallel decomposition algorithm, are described in [24] 3. 2 ANDES Enhancements This section describes the latest improvements of ANDES including a version running with PARMACS [43] and the implementation of VisANDES, a prototype of a visualization tool for ANDES based on the Unix facility gnuplot. 3.2.1 ANDES on PARMACS ANDES has been successfully extended in order to run with the PARMACS communication macros [43] This extension allows ANDES to analyze programs running ....

[Article contains additional citation context not shown here]

R. Hempel, The ANL/GMD Macros (PARMACS) in FORTRAN for Portable Parallel Programming using the Message Passing Programming Model. User's Guide and Reference Manual, Pallas GmbH, 1991.

....translators, which would allow comparative performance results. Another interesting and useful comparison is between the time measurements when compiling with the prototype HPF systems and the time measurements from the hand written code with explicit message passing primitives (PARMACS macros [13] in our particular case) The benchmark measurements were taken on three different iPSC 860 machines the 64 node configuration in Daresbury (U.K. the 32 node hypercube in Julich (Germany) and the small development platform at K.U. Leuven (Belgium) Although the convenient and easy access has ....

R. Hempel. The ANL/GMD Macros (PARMACS) in FORTRAN for Portable Parallel Programming using the Message Passing Programming Model. User's Guide and Reference Manual, Version 5.1. GMD, November 1991.

....such networks have much unused compute power 37 to be exploited. In practice, the large latencies involved in communicating among workstations make them low performance parallel computers. Models for workstation message passing include systems such as PVM [24 27, 94, 188, 189] Parmacs [113, 114], and p4 [43] Such models are exactly the same as inter multiprocessor message passing systems, except that they typically have much larger grain processes to help conceal the latency, and they must address heterogeneity of the processors. For example, PVM (Parallel Virtual Machine) has gained ....

R. Hempel. The ANL/GMD macros (PARMACS) in Fortran for portable parallel programming using the message passing programming model -- Users' Guide and Reference Manual. Technical report, GMD, Postfach 1316, D-5205 Sankt Augustin 1, Germany, November 1991.

....existing message passing systems, rather than selecting one of them and adopting it as the standard. Thus, MPI1 has been strongly influenced by work at the IBM T. J. Watson Research Center by Bala, Kipnis, Snir and colleagues [1,2] Intel s NX 2 [17] Express [16] nCUBE s Vertex [14] and PARMACS [11,12]. Other important contributions have come from Zipcode [18,19] Chimp [6,7] PVM [8,20] and PICL [9] One of the objectives of this paper is to promote a discussion within the concurrent computing research community of the issues that must be addressed in establishing a practical, portable, and ....

R. Hempel, "The ANL/GMD Macros (PARMACS) in Fortran for Portable Parallel Programming Using the Message Passing Programming Model -- Users' Guide and Reference Manual," GMD, Postfach 1316, D-5205 Sankt Augustin 1, Germany (November 1991).

....coordination and communication routines that can be invoked from within FORTRAN or C application programs. Over the past several years, a large number of programming environments and communication libraries for parallel computers have been developed, including PVM [3] LINDA [5] PICL [9] PARMACS [11], Zipcode [13] Express [15] the NCUBE 2 library [12] the CM 5 library [14] and the iPSC 860 library [16] The design of the EUI borrows some of the popular communication concepts that already exist in many of these libraries, and in addition, provides several novel operations. The primary goal ....

R. Hempel, "The ANL/GMD macros (PARMACS) in FORTRAN for portable parallel programming using the message passing programming model, user's guide and reference manual", Technical Memorandum, Gesellschaft fur Mathematik und Datenverabeitung mbH, West Germany.

....provide hardware or low level system support for, thereby enhancing scalability. The functionality that mpi is designed to provide is based on current common practice, and is similar to that provided by widely used message passing systems such as Express [12] NX 2 [13] Vertex, 11] parmacs [8, 9], and P4 [10] In addition, the flexibility and usefulness of mpi has been broadened by incorporating ideas from more recent and innovative message passing systems such as chimp [4, 5] Zipcode [14, 15] and the IBM External User Interface [7] The general design philosophy followed by mpi is that ....

R. Hempel. The ANL/GMD macros (PARMACS) in fortran for portable parallel programming using the message passing programming model -- users' guide and reference manual. Technical report, GMD, Postfach 1316, D-5205 Sankt Augustin 1, Germany, November 1991.

....to another. ffl FortNet: FortNet [1] is a message passing harness developed for use in applications. It is integrated with a number of other parallel tools including a graphical execution monitoring package and a parallel programming development tool. ffl PARMACS: PARMACS (PARallel MACros) [60, 61] has been developed by the GMD Birlinghoven as a portable message passing system for Fortran and C programs running on a variety of machines. ffl PICL: PICL (Portable Instrumented Communications Library) 47, 48] was developed at Oak Ridge National Laboratory. It is a subroutine library that ....

R. Hempel. The ANL/GMD macros (PARMACS) in Fortran for portable parallel programming using the message passing programming model -- users' guide and reference manual. Technical Report, GMD, Postfach 1316, D-5205 Sankt Augustin 1, Germany, 1991.

....using lightweight pre emptive processes while retaining portability. These features are required for our research into the use of optimistic mechanisms, based on the virtual time paradigm [13] in general purpose parallel computing 1 . Our reasons for choosing p4 (which is related to PARMACS [9]) rather than other message passing systems such as PVM [18] were based on performance [8] and the suitability of the p4 programming model for our research. We are closely following the efforts to define a message passing interface standard [7] but this work is still at a draft stage. 1 The ....

R. Hempel. The ANL/GMD macros (PARMACS) in fortran for portable parallel programming using the message passing programming model -- users' guide and reference manual. Technical report, GMD, Sankt Augustin, Germany, 1991.

....routines provided by a third party. Calls to these routines are then either translated into native calls on each machine or into a message passing system provided by the third party. These strategies have been used in many packages including Chameleon [8] Express [14] p4 [4] PARMACS [9] and PVM [2] A third alternative is to take a code written for one native message passing system and use a look alike library to allow usage of the same code on another native message passing system (cf. 8, 12] These solutions have several drawbacks. First, the library writers must either port ....

Hempel, R., The ANL/GMD macros (PARMACS) in Fortran for portable parallel parogramming using the message passing programming model - users' guide and reference manual, GMD (November 1991).

....translators, which would allow comparative performance results. Another interesting and useful comparison is between the time measurements when compiling with the prototype HPF systems and the time measurements from the hand written code with explicit message passing primitives (PARMACS macros [12] in our particular case) The benchmark measurements were taken on four different iPSC 860 machines the 64 node configuration in Daresbury (U.K. the 32 node hypercube in J ulich (Germany) a 32 node machine available to APR, Inc. U.S.A. and a small development platform at K.U. Leuven ....

R. Hempel. The ANL/GMD Macros (PARMACS) in FORTRAN for Portable Parallel Programming using the Message Passing Programming Model. User's Guide and Reference Manual, Version 5.1. GMD, November 1991.

....model of multi stage interconnection networks and point to point networks with virtual cut through like routing algorithms. In fact, this model has been widely used by several researchers (see [24] and has been adopted by numerous communication libraries, such as Express [21] PARMACS [25], PICL [23] Zipcode [31] Venus [5] and CCL [4] Two complexity measures are frequently used in evaluating communication algorithms in such a model. The first measure, denoted by C 1 , is the number of communication rounds required by an algorithm. Notice that in each round, the messages sent ....

R. Hempel, The ANL/GMD macros (PARMACS) in FORTRAN for portable parallel programming using the message passing programming model, user's guide and reference manual, Technical Memorandum, Gesellschaft fur Mathematik und Datenverabeitung mbH, West Germany.

....feature is not distributed shared memory; but rather, a portable mechanism for shared address space programming in true shared memory multiprocessors. Parmacs is a project that is closely related to the P4 effort. Essentially, Parmacs is a set of macro extensions to the P4 system developed at GMD [10]. It originated in an effort to provide Fortran interfaces to the P4 system, but is now a significantly enhanced package that provides a variety of high level abstractions, mostly dealing with global operations. Parmacs provides macros for logically configuring a set of P4 processes; for example, ....

R. Hempel, The ANL/GMD MAcros (Parmacs) in Fortran for Portable Parallel Programming Using Message Passing, GMD Technical Report, November 1991.

....to write and run distributed computer applications. Tools to start and monitor distributed applications are usually integrated into these systems. Many of these communication environments for distributed computing provide constructs to facilitate message passing (PVM [26] P4 [5] 6] PARMACS [13], Parform [7, 8] Another environment, Network Linda [9] provides a virtual shared memory called tuple space through which data is transferred between processes or from which idle processors can fetch processes (tasks) to be executed. High Performance Fortran [11] and Vienna Fortran [10] are ....

R. Hempel. The ANL/GMD Macros (PARMACS) in FORTRAN for Portable Parallel Programming using the Message Passing Programming Model, User's Guide and Reference Manual. Gesellschaft f¨ur Mathematik und Datenverarbeitung mbH, November 1991.

....recently. Most of CCL is implemented as part of the parallel application programming interface of SP1. Over the past few years, a large number of programming environments and communication libraries for parallel computers have been developed, including PVM [8] Linda [16] PICL [29] PARMACS [31], Zipcode [35] Express [37] the nCUBE 2 library [34] the CM 5 library [36] and the iPSC 860 library. The design and implementation of the CCL adopts some of the popular communication concepts that already exist in many of these libraries, and, in addition, it provides several novel aspects. In ....

....environments. These trends are evident in systems such as Thinking Machines CM 5 [36] Intel s Paragon [38] NCUBE s nCUBE 2 [34] MIT s J Machine [20] IBM s Vulcan [10, 39] and the recently announced IBM s Scalable POWERparallel System 1 (SP1) and in environments such as Express [37] PARMACS [31], PICL [29] Zipcode [35] and Venus [4] These systems and environments generally ignore the specific structure and topology of the communication network and assume a fully connected collection of processes, in which each process can communicate directly with any other process by sending and ....

R. Hempel, "The ANL/GMD macros (PARMACS) in FORTRAN for portable parallel programming using the message passing programming model, user's guide and reference manual", Technical Memorandum, Gesellschaft fur Mathematik und Datenverabeitung mbH, West Germany.

....the software development within this project. The toolset will be portable. The portability layer for the cross development tools will be the PCTE (Portable Common Tool Environment) database system [17, 3] and the portability layer for the run time system will be the PARMACS message passing macros [10, 16]. It recognised that internationally accepted standards which define stable interfaces between sub systems are the key to assuring both portability and inter operability: ffl Language standards form the interface between applications and their human developers. The primary application programming ....

R. Hempel, The ANL/GMD Macros (PARMACS) in FORTRAN for Portable Parallel Programming using the Message Passing Programming Model, User's Guide and Reference Manual, Version 5.1, Gesellschaft fur Mathematik und Datenverarbeitung mbH, Postfach 1240, 5205 St. Augustin 1, Germany.

....hardware or low level system support for, thereby enhancing scalability. The functionality that mpi is designed to provide is based on current common practice, and is similar to that provided by widely used message passing systems such as Express [15] PVM [2] NX 2 [16] Vertex, 14] parmacs [10, 11], and P4 [4, 13] In addition, the flexibility and usefulness of mpi has been broadened by incorporating ideas from more recent and innovative message passing systems such as chimp [6, 7] Zipcode [17, 18] and the IBM External User Interface [8] The general design philosophy followed by mpi is ....

R. Hempel. The ANL/GMD macros (PARMACS) in fortran for portable parallel programming using the message passing programming model -- users' guide and reference manual. Technical report, GMD, Postfach 1316, D-5205 Sankt Augustin 1, Germany, November 1991.

....provide hardware or low level system support for, thereby enhancing scalability. The functionality that MPI is designed to provide is based on current common practice, and is similar to that provided by widely used message passing systems such as Express [11] NX 2 [12] Vertex, 10] parmacs [7, 8], and P4 [9] In addition, the flexibility and usefulness of MPI has been broadened by incorporating ideas from more recent and innovative message passing systems such as chimp [4, 5] Zipcode [13, 14] and the IBM External User Interface [6] The MPI standardization effort involved about 60 ....

R. Hempel. The ANL/GMD macros (PARMACS) in fortran for portable parallel programming using the message passing programming model -- users' guide and reference manual. Technical report, GMD, Postfach 1316, D-5205 Sankt Augustin 1, Germany, November 1991.

....Up to version 5.1, the PARMACS package consisted of a set of macro definitions and a library of Fortran routines. The programmer had to insert macro calls into his Fortran program that were expanded by the m4 macro processor [2] into Fortran code containing calls to the library routines (see [1]) In this way, by inserting a reference to a global variable or a machine specific system call for a PARMACS macro, the overhead of a subroutine or function call could often be avoided. Furthermore, this approach made it easy to produce an event trace useful for parallel debuggers and ....

Rolf Hempel. The ANL/GMD Macros (PARMACS) in FORTRAN for Portable Parallel Programming using the Message Passing Programming Model --- User's Guide and Reference Manual. GMD, Postfach 1240, D--5205 St. Augustin 1, Germany, version 5.0 edition, August 1991.

....message passing systems, rather than selecting one of them and adopting it as the standard. Thus, MPI1 has been strongly influenced by work at the IBM T. J. Watson Research Center by Bala, Kipnis, Snir and colleagues [2, 3] Intel s NX 2 [18] Express [17] nCUBE s Vertex [15] and PARMACS [11, 13]. Other important contributions have come from Zipcode [19, 20] Chimp [6, 7] PVM [8, 21] and PICL [9] One of the objectives of this paper is to promote a discussion within the concurrent computing research community of the issues that must be addressed in establishing a practical, portable, ....

R. Hempel. The ANL/GMD macros (PARMACS) in fortran for portable parallel programming using the message passing programming model -- users' guide and reference manual. Technical report, GMD, Postfach 1316, D-5205 Sankt Augustin 1, Germany, November 1991.

No context found.

R. Hempel, The ANL/GMD Macros (PARMACS) in FORTRAN for Portable Parallel Programming Using the Message Passing Programming Model, User's Guide and Reference Manual. Gesellschaft fur Mathematik und Datenverarbeitung mbH (GMD), november 1991.

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