P. Newton, J.C. Browne, The CODE 2.0 graphical parallel programming language, Proceedings of the Sixth ACM International Conference on Supercomputing, 1992, pp. 167--177.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....platform for capturing the parallel structure of a program as well as its associated workloads for di erent modeling techniques. Task graphs and their equivalents have proved to be a valuable abstraction for representing the execution of parallel programs in a number of di erent applications [3, 20, 24, 26, 15, 5, 14, 27, 23, 21]. Although the task graphs used in these systems di er in representation and semantics (e.g. whether task graph edges capture purely precedence constraints or also data ow requirements) there are close similarities. Perhaps most importantly, they all capture the parallel structure of a program ....

....parallel programs. The vast majority of performance models that use task graphs as inputs generally do not specify how the task graph should be constructed but assume that this has been done [3, 15, 20, 26] The various compiler based systems that use task graphs, namely PYRROS [27] CODE [23], HENCE [23] and Jade [21] construct task graphs by assuming some kind of information from the programmer. The PlusPYR project [12] has developed a task graph representation that has some similarities with ours along with compiler techniques to synthesize these task graphs. The key di erence ....

[Article contains additional citation context not shown here]

P. Newton and J. C. Browne. The CODE 2.0 Graphical Parallel Programming Language. In Proceedings of the 1992.

....in the same way as follows. procID[ x procID stride[i] 0] ncolumns) ncolumns] y procID stride[i] 1] The system provides textual changes and reduces users trial and error work. 6 Related Work There are some GUI systems for parallel computing. P. Newton and J. C. Browne proposed CODE [6], which is a visual parallelprogramming language. G. A. Geist proposes HeNCE [7] which is also a visual parallelprogramming language. Newton compared their characteristics in reference [8] We focus on interprocessor communications in the GRIX system because it represents the most characteristic ....

P. Newton and J. C. Browne. The code 2.0 graphical parallel programming language. In Proceedings of ACM International Conference on Supercomputing, July 1992.

....resembles the structure and execution of parallel programs. In the broadest perspective, they usually do not provide single graph formalism for all the activities. It can be observed that some of the more recent tools use a simplified and high level of abstraction to represent parallel programs [1, 2, 3, 4, 6]. They view parallel programming as a two step process, where the programmer first writes sequential components as text, and then utilizes graphics to organize the components into a graph, the so called large grain approach. Moreover, a system based exclusively on a control flow or data dependency ....

....generated from the execution data to identify the program or execution errors and to tune the performance. 5. RELATED WORK The approach of PCG to solving the complexity of parallel programming is similar to the approach taken by visual programming languages like HeNCE [2] Phred [3] and CODE [4]. PCG is, however, unique in its graph formalism that is designed to support program construction, debugging and performance tuning. Graphs in HeNCE and CODE can display only one type of information, either a control flow or a data flow, respectively. PCG is inherently a combination of the two. ....

Newton, P., and Browne, J. C. The CODE 2.0 Graphical Parallel Programming Language, Proceedings of ACM International Conference on Supercomputing. 1992, http://www.cs.utexas.edu/users/code

....visual language often depends on the editor used for graphics rendering and program composition. The editor enforces consistency and performs syntax checking. The ease of use has been recognized as a more important goal of visual programming environments. Newton conducted a case study of the CODE [6] parallel programming environment, and concluded that though important, ease of use is difficult to evaluate due to its subjective nature [7] It requires a comparative study of multiple languages and environments, such as the one reported in [8] conducted on many subjects. Based on these and ....

....were asked to implement a solution to a matrix multiplication problem. Figure 5. Poisson main program 4.2. The Lecture The lecture consisted of two presentations. In the first presentation, one subject talked about visual languages for parallel programming and PCG, and briefly compared PCG CODE [6] and HeNCE [14] This was his second presentation on the topic, since he gave one before to another group of students, with the idea to explain the PCG approach to other possibly interested students in the future. That presentation was also attended by the remaining three subjects who were given a ....

P. W. Newton & J. C. Browne (1992) The CODE 2.0 graphical parallel programming language. Proceedings of the ACM International Conference on Supercomputing, July.

....nodes re when all predecessors have red. HeNCE uses a shared name model: two or more nodes can access the same variable. The problem with HeNCE is that it fails to graphically 26 display the communication between nodes (tasks) since the graph shows control AEow and not dataAEow. 1.5. 2 CODE CODE [40] produces directed graph based programs that depict dataAEow between nodes of sequential programs. It also supports shared variables. A node res when the programmer supplied ring rules are satised. However, the nodes generally re when the data on all incoming arcs are present. Features of note ....

P. Newton and J. C. Browne, The CODE 2.0 Graphical Parallel Programming Language, Porceedings of the ACM International Conference on Supercomputing, July 1992.

....in, saving design time. Finally, modular specifications of systems are more suitable for automatic synthesis and simulation than non modular ones and have good scalability characteristics. This model has proved to be useful in the Unix operating system and in various parallel programming models [1, 8, 7, 13]. More recently it has been used to successfully model complicated hardware designs, where it has shown great promise in enabling very concise, clear specifications [2, 14] We want to underline once more that the loosely coupled modules connected by FIFOs model we use is not ....

P. Newton and J. C. Browne. The CODE 2.0 graphical parallel programming language. In Proceedings of the 1992 ACM International Conference on Supercomputing, Washington, DC, July 1992.

....of large distributed systems that distinguish a distributed system of other types of systems. Next the DADL language will be defined. DADL will then be used to specify several key architectures. Other related work includes Rapide [LV96] UniCon [SDK 94] ArTek [HRCP94] Wright [AG94] Code [NB92] Demeter [PXL95] Modechart [JM94] PSDL CAPS [LSBH93] Resolve [EHL 94] and Meta H [Ves94a] 10 3 Distributed Architecture Description Language 3.1 Overview There are two main, tightly coupled concepts in DADL which are the conversation and the participants. The conversation has an ....

Newton and Browne. The code 2.0 graphical parallel programming language. Proceedings, ACM International Conference on Super Computing, July 1992.

....of parallel programs, including quantitative analytical models [3, 19, 25 27] theoretical and abstract analytical models [14] and program simulation [5, 13] A second important use of task graphs is in parallel programming systems. Parallel programming environments such as PYRROS [28] CODE [24], HENCE [24] and Jade [20] have used task graphs at three di erent levels: as a programming notation for expressing parallelism, as an internal representation in the compiler for computation partitioning and communication generation, and as a runtime representation for scheduling and execution of ....

....programs, including quantitative analytical models [3, 19, 25 27] theoretical and abstract analytical models [14] and program simulation [5, 13] A second important use of task graphs is in parallel programming systems. Parallel programming environments such as PYRROS [28] CODE [24] HENCE [24], and Jade [20] have used task graphs at three di erent levels: as a programming notation for expressing parallelism, as an internal representation in the compiler for computation partitioning and communication generation, and as a runtime representation for scheduling and execution of parallel ....

[Article contains additional citation context not shown here]

P. Newton and J. C. Browne. The CODE 2.0 Graphical Parallel Programming Language. In Proceedings of the 1992 ACM International Conference on Supercomputing, Washington, DC, July 1992.

....models and dynamically gathered resource information. Visual programming based on the speci cation of applications and algorithms with directed graphs is the basis of the Heterogeneous Network Computing Environment (HeNCE) 4] and the Computationally Oriented Display Environment (CODE) [26]. Browne et al. have reviewed the use of visual programming in parallel computing and compared the approaches of HeNCE and CODE [7] Though a similar approach is used by the VCCE described in Section 2, HeNCE and CODE were designed for use at a ner level of algorithm design; thus, they require a ....

P. Newton and J. C. Browne, \The CODE 2.0 Graphical Parallel Programming Language", in Proceedings of the ACM International Conference on Supercomputing, July, 1992.

....[2] is a visual programming environment for composing and running PVM programs. It provides a GUI that keeps a consistent mental and visual image for the programmer while programming and running parallel programs, even though it uses two similar visual representations. On the other hand, CODE [11] is more concerned with program composition and compilation than with runtime analysis. It has a hierarchical graph and an object oriented architecture that facilitate the process of adding new target architecture to a PVM environment. In this paper we present a novel tool for MPI programming on ....

Newton, P., Browne, J. C.: The CODE 2.0 Graphical Parallel Programming Language. Proceedings of ACM International Conference on Supercomputing (1992) http://www.cs.utexas.edu/users/code

....portable than serial programs because their structure may critically depend on the hardware s specific architectural features. Since visual structures and relationships are much easier to reason about than similar structures described in plain text, many visualization systems have been proposed [1, 2, 3, 4, 5] to assist software development for parallel computing. These systems are graph based, because a multidimensional directed graph closely resembles the execution of parallel programs. When developing a program, we find that there are four main stages: problem partitioning, program construction, ....

....After the calculation is performed, the result is tested for convergence, against a highly accurate approximation of p. 5. Related Work The approach of PCG to solving the complexity of parallel programming is similar to the approach taken by visual languages like HeNCE [2] Phred [3] and CODE [4]. PCG is, however, unique in its graph formalism that is designed to support program construction, debugging and performance tuning. In this section we compare PCG with the mentioned languages. Graphs like HeNCE and CODE can display only one type of information, either control flow or data flow, ....

Newton P and Browne J C. The CODE 2.0 Graphical Parallel Programming Language, Proc. ACM International Conference. on Supercomputing, 1992, available from the Internet site: http://www.cs.utexas.edu/users/code.

....on the same computer. This problem is compounded for nondeterministic programs, because it is difficult to isolate programming mistakes from the nondeterminism caused by the runtime environment. There have been a number of tools built to help with these issues, like HeNCE [Beguelin 92] CODE [Newton 92] Phred [Beguelin 90] The primary role of Phred is to encourage writing deterministic rather than nondeterministic programs, even though the latter can yield better performance. HeNCE is targeted to solve problems with composing and running PVM programs. It provides a GUI that uses two different ....

Newton, P., and Browne, J. C. The CODE 2.0 Graphical Parallel Programming Language. Proceedings of ACM International Conference on Supercomputing. 1992, http://www.cs.utexas.edu/users/code.

....resembles the structure and execution of parallel programs. Older systems like Schedule [15] and Poker [16] dealt with low level issues like message routing and were usually tied to a particular environment or architecture. The more recent systems like Paralex [1] HeNCE [2] Phred [3] CODE [4], and VPE [5] have adopted higher level of abstraction providing programmers with tools that are tailored more towards computation rather than hardware specific issues. Activities in parallel programs are distributed across space and time. A computer program in its raw textual form is a ....

....the data buffers local to each end of a communication channel and the processors involved in a communication event: pR] sendBuffer [bI] pR] receiveBuffer [bI] where: pR is the processor rank bI is the buffer index. Its syntax is similar to the arc topology specification found in CODE [4], but it does not require dots to be used before the buffer names, because in Visper resource names are used to resolve the ownership. Type stanza is used to define the type of the data in a buffer and message protocol i.e. tag. Therefore, it takes two attributes: a type and a tag and the syntax ....

[Article contains additional citation context not shown here]

Newton P, Browne J C. The CODE 2.0 Graphical Parallel Programming Language, Proc. ACM International Conference. on Supercomputing, 1992, available from the Internet site: http://www.cs.utexas.edu/users/code.

....problem domain. Library systems such as PUL, archetypes and Cole s code skeletons [7] which provide libraries for particular parallel structures, generally fail this characteristic because of their limited set of structures. Visual tools and general libraries (such as PVM, MPI, HeNCE, and CODE [21]) meet this characteristic by virtue of their generality, which introduces other correctness problems discussed later. 6. Openness Parallel programming systems must exhibit openness, allowing programmers to access low level features and implementation details in their applications. In our case, ....

P. Newton and J. Browne. The CODE 2.0 graphical parallel programming language. In Proceedings of the 6th ACM International Conference on Supercomputing, pages 167-177, 1992.

....their problem domain. Library systems such as PUL, archetypes and Cole s code skeletons [7] which provide libraries for particular parallel structures, generally fail this characteristic because of their limited set of structures. Visual tools and general libraries (such as PVM, HeNCE, and CODE [20]) meet this characteristic by virtue of their generality, which introduces other correctness problems discussed later. 6. Openness Parallel programming systems must exhibit openness, allowing programmers to access low level features and implementation details in their applications. In our case, ....

P. Newton and J. Browne. The CODE 2.0 graphical parallel programming language. In Proceedings of the 6th ACM International Conference on Supercomputing, pages 167-177, 1992.

....The dataflow corresponding to these two states simply connects a string process (the String class is predefined) to the print process. The code generated for this example can be found in Appendix A. 7 Comparison with existing tools In this section, we compare Visifold and other tools like CODE2 [NB92] Phred [BN94] and HeNCE2 [Ba94] Many visual programming tools tend to make the distinction between coordination and computation. Typically, what these tools represent is intrinsically multidimentional and reflects the coordination of sequential processes. Consequently what most designers of ....

Peter Newton and James C. Browne. The code 2.0 graphical parallel programming language. In Proc. ACM Int. Conf. on Supercomputing, July 1992. Appendix A: Code generated for the Fibonacci sequence

....of parallel programs, including quantitative analytical models [30, 22, 29, 28, 2] theoretical and abstract analytical models [15] and program simulation [14, 4] A second important use of task graphs is in parallel programming systems. Parallel programming environments such as PYRROS [31] CODE [27], HENCE [27] and Jade [23] have used task graphs at three di erent levels: as a programming notation for expressing parallelism, as an internal representation in the compiler for computation partitioning and communication generation, and as a runtime representation for scheduling and execution of ....

....programs, including quantitative analytical models [30, 22, 29, 28, 2] theoretical and abstract analytical models [15] and program simulation [14, 4] A second important use of task graphs is in parallel programming systems. Parallel programming environments such as PYRROS [31] CODE [27] HENCE [27], and Jade [23] have used task graphs at three di erent levels: as a programming notation for expressing parallelism, as an internal representation in the compiler for computation partitioning and communication generation, and as a runtime representation for scheduling and execution of parallel ....

[Article contains additional citation context not shown here]

P. Newton and J. C. Browne. The CODE 2.0 Graphical Parallel Programming Language. In Proceedings of the 1992 ACM International Conference on Supercomputing, Washington, DC, July 1992.

....dependency, i.e. when none of the input variables to the second procedure of the composition are outputs from the first procedure. These two rules allow the computation graph to be interpreted as an expression of parallelism much as the graphs in parallel programming environments, such as CODE [12, 49], and reengineering tools such as E SP [61] There are more subtle kinds of parallelism that are not so easily deduced from the computation graphs. In particular, it is often useful to unroll recursion and execute the different depths of a recursive procedure in parallel. Furthermore, the division ....

Newton, Peter, and Browne, James C., The CODE 2.0 Graphical Parallel Programming Language, Proc. ACM International Conf. on Supercomputing, July, 1992

....style. One example is Sisal (see [Feo90] which was designed for high performance computing and seems to deliver performance comparable to Fortran programs. Various parallelizing Sisal compilers 13 have been developed. There is also a visual dataflow programming tool called Code (see [NB92]) 1.3 Dynamic load balancing The static mapping and scheduling of task graphs is not sufficient if the task graph is not known completely at start time. In some cases, the task graph is not known completely at start time. One example are iterative solvers for linear systems (e.g. Conjugate ....

P. Newton and J.C. Browne, The code 2.0 graphical parallel programming language, Proceedings of the ACM International Conference on Supercomputing, July 1992.

No context found.

Newton P. and Browne J. C., The CODE 2.0 Graphical Parallel Programming Language, in Proceedings of the ACM International Conference on Supercomputing, July 1992.

No context found.

P. Newton, J.C. Browne, The CODE 2.0 graphical parallel programming language, Proceedings of the Sixth ACM International Conference on Supercomputing, 1992, pp. 167--177.

No context found.

Browne, J.C. and P. Newton. The CODE 2.0 Graphical Parallel Programming Language. in ACM Int. Conf. 8 on Supercomputing. 1992. Washington DC: ACM Press.

No context found.

P. Newton, J.C. Browne, The CODE 2.0 Graphical parallel programming language, in: Proceedings of the 1992.

No context found.

P. Newton and J. C. Browne. The CODE 2.0 Graphical Parallel Programming Language. In Proceedings of the 1992.

No context found.

Newton, P. and Browne, J. C., \The CODE 2.0 Graphical Parallel Programming Language," ACM Intn'l. Conference on Supercomputing, July 1992.

First 50 documents  Next 50

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