A. Malony, B. Mohr, P. Beckman, S. Yang, F. Bodin. Performance Analysis of pC++: A Portable Dataparallel Programming System Scalable Parallel Computers. In Proceedings of the Eighth International Parallel Processing Symposium, pp. 75-85, 1994.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....systems provide, or could support the development of, program analysis tools. The Arcadia project and its ProDAG subsystem [12, 21] provide program analysis tools for use with Ada programs; these include tools for constructing control flow graphs and computing data and control dependencies. Sage [16] creates abstract syntax trees for C and C programs, and provides several utilities that can be used to create new analysis tools. Aria [6] generates testing and analysis tools for C and C programs from abstract syntax graph representations (an abstract syntax tree with semantic information) ....

A. Malony, B. Mohr, P. Beckman, D. Gannon, S. Yang, and F. Bodin. Performance analysis of pC++: A portable data-parallel programming system for scalable parallel computers. In 8th International Parallel Processing Symposium, pages 75--84, April 1994.

....system features, instead of providing portable, user level thread creation, synchronization and communication primitives. Among the first profiling tools for a user level thread library was Quartz [9] A target specific profiling environment for concurrent object oriented programs is pC [10]. pC is one of the few cases where the integrated performance analysis environment TAU [11] was implemented in concert with the language and runtime system. However, the design of pC TAU incorporates most of its profiling functionality into the preprocessor and runtime system, so extending the ....

A. Malony et al. Performance analysis of pC++: A portable data-parallel programming system for scalable parallel computers. In Proc. of the 8th International Parallel Processing Symp., pages 75--85, Cancun, Mexico, April 1994.

.... SP1, BBN TC2000, Sequent Symmetry, and a homogeneous networks of workstations (PVM) A lot of papers, the program files, and additional information are available from http: www.extreme.indiana.edu sage Email address: Dennis Gannon Gamma gannon cs.indiana.edu References: 31] 32] 143] 84] [153] 2.78 Parallel Computing Action Developer: Rank Xerox France, University of Paris VI Description: oo. Extension of C . Library. memory model. Hierarchical memory model. The programmer understands the difference between local objects and remote objects. This difference is apparent when an ....

A. Malony, B. Mohr, P. Beckman, D. Gannon, S. Yang, and F. Bodin. Performance analysis of pC++: A portable data-parallel programming system for scalable parallel computers. In H. J. Siegel, editor, Proc. of the 8th Int. Parallel Processing Symp. IPPS'94, Canc'un, Mexico, April 1994. IEEE Computer Society Press.

....2.2 Traditional performance tools There are many general purpose performance tools for parallel and sequential programs that provide performance measures in terms of the program s execution. Most of these tools for software measurement probes to be inserted in an application program s source code [43], or inserted by re compiling or re linking with an instrumentation library [55, 73, 70, 22, 54] or by re writing the binary [40] or dynamically at run time [47] Most of these tools can be classified as either profiling [22, 70, 43, 54] or event tracing tools [55, 73, 43, 25] Profiling tools ....

.... probes to be inserted in an application program s source code [43] or inserted by re compiling or re linking with an instrumentation library [55, 73, 70, 22, 54] or by re writing the binary [40] or dynamically at run time [47] Most of these tools can be classified as either profiling [22, 70, 43, 54] or event tracing tools [55, 73, 43, 25] Profiling tools typically insert instrumentation code to count and or time the execution of fragments of application code, run the application, and compute the total value of the performance metrics associated with different code fragments. One problem ....

[Article contains additional citation context not shown here]

A. Malony, B. Mohr, P. Beckman, D. Gannon, S. Yang, and F. Bodin. Performance Analysis of pC++: A Portable Data-Parallel Programming System for Scalable Parallel Computers. In Proceedings of the 8th International Parallel Processing Symposium (IPPS), pages 75--85, Cancun, Mexico, April 1994.

....of the original serial code. 2.6 Other Models and Implementations In this section, we survey and briefly compare the approaches represented in this work to nonactor models and implementations. Other actor approaches are considered in Section 3.6. 2.6. 1 pC Gannon and Lee [21] 22] 23] [24] have developed pC , an extension of C with support for distributed data structures similar to FORTRAN D. The pC language provides support for distributed collections, both array based and tree based, of arbitrary types and with full support for the C mechanisms of derivation and ....

A. Malony, B. Mohr, P. Beckman, D. Gannon, S. Yang, , and F. Bodin, "Performance analysis of pC ++ : A portable data-parallel programming system for scalable parallel computers," in Proceedings, 8th International Parallel Processing Symposium, pp. 75--84, Apr. 1994.

....Java VM and the Java application, and accounted for these costs in terms of AP resources. With this data, we were easily able to determine what changes to make to the Java application to improve its performance. 5 Related Work There are many general purpose program performance measurement tools [3, 5, 7, 4] that can be used to measure the performance of the virtual machine. However, these are unable to present performance data in terms of the application program or in terms of the interaction between VM and AP. There are some performance tools that provide performance data in terms of AP s execution ....

A. Malony, B. Mohr, P. Beckman, D. Gannon, S. Yang, and F. Bodin. Performance Analysis of pC++: A Portable Data-Parallel Programming System for Scalable Parallel Computers. Proceedings of the 8th International Parallel Processing Symposium (IPPS), Cancun, Mexico, pages 75--85, April 1994.

....with the final effect to increase their overall execution time. Performance tools have to address all these issues. Hence, a tight integration with the compilation systems is required to relate the achieved performance to the source code. Examples of such an integration are presented in [1] [6], 7] This work was supported in part by the European Commission under the Esprit IV Long Term Research Project n. 21033 HPF . and [8] In [1] the performance of Fortran D programs is analyzed at the source level thanks to the integration of Fortran D compiler and the Pablo performance ....

....and [8] In [1] the performance of Fortran D programs is analyzed at the source level thanks to the integration of Fortran D compiler and the Pablo performance tool. The integration of TAU with the pC compiler and run time system allows instrumentation, profiling, and tracing support [6]. The source level evaluation of C, C , Fortran and HPF codes is addressed by the MPP Apprentice tool [8] and by the Visualization Tool [7] In this paper we present the integration of the VFC compilation system [3] and the Medea performance analysis tool [4] This work is part of the Esprit LTR ....

A.D. Malony, B.W. Mohr, P. Beckman, D. Gannon, S. Yang, and F. Bodin. Performance analysis of pC++: a portable data-parallel programming system for scalable parallel computers. In Proc. 8th Int. Parallel Processing Symposium, pages 75--84, Mexico, April 1994. IEEE Computer Society Press.

....as we have integrated all our necessary support directly into the distributed system. This allows us to cleanly provide novel features such as using logical event timestamping and the choice of monitoring support options at execution time. Another system that has an approach similar to ours is pC [9]. The pC system is a language extension to C coupled with a runtime system that provides a framework for data parallel computing. pC provides substantial integrated support for event tracing,and pC system specific analysis and visualization tools are under development. 7 Conclusion ....

A. Malony, B. Mohr, P. Beckman, D. Gannon, S. Yang, and Bodin F. Performance analysis of pC++: A portable data-parallel programming system for scalable parallel computers. In Proceedings of the 8th International Parallel Processing Symposium (IPPS '94), pages 75--84, Cancun, Mexico, April 1994.

....or ftp: cica.cica.indiana.edu in directory pub sage WWW: http: www.cica.indiana.edu sage Comments: May be of interest to C programmers who are excited by HPF. Report [22] gives pC examples for a parallel tridiagonal matrix solver and a fast Poisson solver. References: [22, 23, 95] 8 LANGUAGES AND LANGUAGE EXTENSIONS (EXCLUDING HPF) 34 8.2 Split C Name: Split C Description: Parallel extensions to the C language. Aims to provide the most useful features from message passing, shared memory, and dataparallel programming models whilst making efficient use of the underlying ....

A Malony, B Mohr, P Beckman, S Yang, and F Bodin. Performance analysis of pc++: A portable data-parallel programming system for scalable parallel computers. In Proceedings of the 8th International Parallel Processing Symposium (IPPS), Cancun, Mexico, April 1994. Available by anonymous ftp from ftp.cs.indiana.edu in directory /pub/sage.

....programming paradigm. General tracing tools are very flexible, but they do not include built in knowledge about the underlying Orca system and thus cannot provide feed back at the abstraction level of shared objects. Programming paradigm specific tracing tools, such as for data parallel programs [20], clearly cannot be used to accommodate the shared object model used in Orca. A tracing tool for performance debugging of Orca programs should fulfil the following requirements. First and foremost, the tracing tool should visualize the behavior of an Orca program in terms of objects and ....

.... system uses instrumentation of the binary to retrieve this information, and a tool NV to present the performance data at the desired level of abstraction [15] pC knows specialized object types for data parallel arrays, and their tool uses this data function mapping to achieve the same goal [20]. There is a correspondence between data parallel languages and Orca. Both use language level constructs (hidden library calls in dataparallel languages, object operations in Orca) to express parallelism. Both have data types in which parallelism is encapsulated (data parallel arrays and shared ....

A. Malony and B. Mohr et.al. Performance analysis of pC++: A portable data-parallel programming system for scalable parallel computers. In Proceedings of the 8th IPPS, pages 75--85, Canczon, Mexico, April 1994.

.... and Compilers for Parallel Computing Sage is a toolkit for building source to source translators from Indiana University [6] The authors foresee optimizing translators, simulation of language extensions, language preprocessors, and code instrumentation as possible applications of Sage [7, 46, 26, 5]. Sage is written in C and provides parses for C, C , pC , Fortran 77, and Fortran 90. Because Sage is a toolkit instead of a compiler, it is not limited to particular hardware architectures. SUIF from Stanford University is a compiler framework that can be used as either a source to C ....

A. Malony et al. Performance analysis of pC++: A portable data-parallel programming system for scalable parallel computers. In Proceedings of the 8th International Parallel Processing Symposium, 1994.

....for implementing logical clocks for all processes. Again, our approach differs as there is no need for a centralized monitor that all nodes must interact with. Instead, all processes automatically manage their own logical clocks. Another system that has an approach similar to ours is pC [10]. The pC system is a language extension to C coupled with a runtime system that provides a framework for data parallel computing. pC provides substantial integrated support for event tracing, and pC system specific analysis and visualization tools are under development. 7 Conclusion ....

A. Malony, B. Mohr, P. Beckman, D. Gannon, S. Yang, and Bodin F. Performance analysis of pC++: A portable data-parallel programming system for scalable parallel computers. In Proceedings of the 8th International Parallel Processing Symposium (IPPS '94), pages 75--84, Cancun, Mexico, April 1994.

....The ports include the Kendall Square Research KSR 1, Intel Paragon, TMC CM 5, IBM SP 1, and homogeneous clusters of UNIX workstations using PVM; work on porting pC to the Cray T3D and Meiko CS 2 is in progress. More details about the pC language and runtime system can be found in [8, 9, 14, 15]. 2.1 An Observability and Analysis Model for pC Collection definition and use are the key aspects for program analysis in the pC system. Programmers using pC require support for observing the collection data structures with respect to both their object oriented definition and their ....

....that allows the user to locate the corresponding routine in the callgraph or source text by simply clicking on the related measurement result objects. Before presenting the dynamic analysis tools racy and easy, we briefly describe the approach used to implement profiling and tracing in pC (see [15] for more details) 5.1 Portable Profiling for C and Languages based on C A very valuable tool for program tuning is function profiling. Here, special instrumentation code is inserted at all entry and exit points of each function to capture data that can be used to calculate the number of ....

A. Malony, B. Mohr, P. Beckman, D. Gannon, S. Yang, F. Bodin, Performance Analysis of pC++: A Portable Data-Parallel Programming System for Scalable Parallel Computers, Proc. 8th Int. Parallel Processing Symb. (IPPS), Canc'un, Mexico, Apr. 1994.

....behavior may be too unpredictable to guide the extrapolation process. 3 A Performance Extrapolation Technique for pC We have developed a performance extrapolation technique that allows performance information and metrics to be predicted for data parallel programs written in the pC language [1,8,9,16]. In particular, we investigated the problem of extrapolating from a 1 processor execution of a n thread parallel program to a n processor execution for an environment where certain architectural and system parameters are configurable. In general, our approach is to execute a n thread pC program ....

....for barrier synchronization. The message transfer time will contribute to the barrier time. 0 do not use actual messages for barrier synchronization. 1 BarrierMsgSize Size of a message used for barrier synchronization. 128 TABLE 1. Parameters for the Barrier Model 9 important component [16]. Second, we wanted to verify that modifying simulation parameters of interest resulted in observable and expected effects in extrapolated benchmark performance behavior. Although ExtraP provides an ability to do performance debugging of portable pC programs, to apply extrapolation confidently ....

A. D. Malony, B. Mohr, P. Beckman, D. Gannon, S. Yang, F. Bodin, Performance Analysis of pC++: A Portable Data-Parallel Programming System for Scalable Parallel Computers, Proc. IPPS International Parallel Processing Symposium, pages 75-85, April 1994.

....performance analysis tools developed partly by other groups [6, 7, 8, 9] The toolset provides support for accessing static information about the program and for querying and analyzing dynamic data obtained from program execution. A more detailed discussion of the tools can be found in [10, 11, 12]. 4 A Tour of the Sort Program: Static Analysis The pC implementation of the Sort program is easily derived from the class object defined above. A pC collection of Pvector objects is created. The member functions of the Pvector class are divided into two categories: those functions that ....

A. Malony, B. Mohr, P. Beckman, D. Gannon, S. Yang, F. Bodin, Performance Analysis of pC++: A Portable Data-Parallel Programming System for Scalable Parallel Computers, Proc. 8th Int. Parallel Processing Symb. (IPPS), Canc'un, Mexico, IEEE Computer Society Press, pp. 75--85, Apr. 1994.

....from OACIS, and by IBM R D contract MHVU3704. y Department of Computer and Information Science, University of Oregon, Eugene, Oregon 97403, fdarrylb, hacks, malony, mohrg cs.uoregon.edu. 2 Malony et al. of the static and dynamic analysis tools have been covered in detail elsewhere (see [9, 10]) Instead of reiterating that discussion (a brief description is given in x2.1 and x2.2) we focus on the latest tool for breakpoint based program analysis in x3. Future directions are discussed in x4. 2 The pC Program Analysis Environment We advocate the development of language specific ....

....the Kendall Square Research KSR 1, Intel Paragon, TMC CM 5, IBM SP 1, Sequent Symmetry, BBN TC2000, and homogeneous clusters of UNIX workstations using PVM; work on porting pC to the Cray T3D and Meiko CS 2 is in progress. More details about the pC language and runtime system can be found in [8, 9]. The Environment 3 The tools are implemented as graphical hypertools. While they are distinct tools, they act in concert as if they were a single application. Each tool implements some well defined tasks. If one tool needs a feature of another one, it sends a message to the other tool ....

[Article contains additional citation context not shown here]

A. Malony, B. Mohr, P. Beckman, D. Gannon, S. Yang, F. Bodin, Performance Analysis of pC++: A Portable Data-Parallel Programming System for Scalable Parallel Computers, Proc. 8th Int. Parallel Processing Symb. (IPPS), Canc'un, Mexico, IEEE Computer Society Press, pp. 75--85, Apr. 1994.

No context found.

A. Malony, B. Mohr, P. Beckman, D. Gannon, S. Yang, and F. Bodin. Performance Analysis of pC++: A Portable Data-Parallel Programming System for Scalable Parallel Computers. In H.J. Siegel, editor, Proc. Eighth International Parallel Processing Symposium. IEEE Computer Society Press, April 1994.

....that y could be used instead, since it is redefined later. A very different application of Sage is in constructing a tool for automatic instrumentation of user code. A team at the University of Oregon has used Sage to automatically add instrumentation for tracking class member function calls [9]. They use Sage to add an instrumentation object to the definition of each member function to be tracked. When the program is run, invocation of any of the instrumented methods calls the constructor of the instrumentation object, which can log the method s invocation. Upon exit from the method, ....

A. Malony, B. Mohr, P. Beckman, D. Gannon, S. Yang, F. Bodin, "Performance Analysis of pC++: A Portable Data-Parallel Programming System for Scalable Parallel Computers", In Proc. 8th Int. Parallel Processing Symb. (IPPS), Canc'un, Mexico, Apr. 1994.

....mode on UNIX workstations; supported thread systems are Awesime [6] Pthreads, LWP, and the AT T task library. This enables the testing and pre evaluation of parallel pC applications in a familiar desktop environment. More details about the pC language and runtime system can be found in [1,12]. t provides a collection of tools with user friendly graphical interfaces to help a programmer analyze the performance of pC programs. Elements of the t graphical interface represent objects of the pC programming model: collections, classes, methods, and functions. These language level ....

.... browser called racy (Routine and data ACcess profile displaY) an event trace browser called easy (Event And State displaY) and a breakpoint debugger called breezy (BReakpoint Executive Environment for visualiZation and data displaY) A more detailed discussion of the t tools can be found in [4,12,14]. 3 ExtraP A Performance Extrapolation Tool for pC ExtraP is a performance extrapolation system for pC that has been integrated into t in the guise of speedy. This section explains the modeling approach of extrapolation and the techniques used by ExtraP. 3.1 Performance Extrapolation ....

[Article contains additional citation context not shown here]

A. Malony, B. Mohr, P. Beckman, D. Gannon, S. Yang, F. Bodin, Performance Analysis of pC++: A Portable Data-Parallel Programming System for Scalable Parallel Computers, Proc. 8th Int. Parallel Processing Symb. (IPPS), Mexico, IEEE, pp. 75-85, April 1994.

....mode on UNIX workstations; supported thread systems are Awesime [9] Pthreads, LWP, and the AT T task library. This enables the testing and pre evaluation of parallel pC applications in a familiar desktop environment. More details about the pC language and runtime system can be found in [1,16]. 4 t provides a collection of tools with user friendly graphical interfaces to help a programmer analyze the performance of pC programs. Elements of the t graphical interface represent objects of the pC programming model: collections, classes, methods, and functions. These ....

....called breezy (BReakpoint Executive Environment for visualiZation and data displaY) To generate the dynamic execution data for these tools, profiling and tracing instrumentation and measurement support has been implemented in pC . A more detailed discussion of the t tools can be found in [5,16,18]. 3 ExtraP A Performance Extrapolation Tool for pC ExtraP is a performance extrapolation system for pC that has been integrated into t in the form of speedy. This section explains the modeling approach of extrapolation and the techniques used by ExtraP. 3.1 Performance Extrapolation ....

[Article contains additional citation context not shown here]

A. Malony, B. Mohr, P. Beckman, D. Gannon, S. Yang, F. Bodin, Performance Analysis of pC++: A Portable Data-Parallel Programming System for Scalable Parallel Computers, Proc. 8th Int. Parallel Processing Symb. (IPPS), Mexico, IEEE Computer Society Press, pages 75-85, April 1994.

....of portability. The ports include the Intel Paragon, TMC CM 5, IBM SP 1, Kendall Square Research KSR 1, Sequent Symmetry, BBN TC2000, and workstations clusters using PVM; ports to the Cray T3D and Meiko CS 2 are in progress. More details about the pC language and runtime system can be found in [8, 9]. 2.1 An Observability and Analysis Model for pC Collection definition and use are the key aspects for program analysis in the pC system. Programmers using pC require support for observing the collection data structures with respect to both their object oriented definition and their ....

....that allows the user to locate the corresponding routine in the callgraph or source text by simply clicking on the related measurement result objects. Before presenting the dynamic analysis tools racy and easy, we briefly describe the approach used to implement profiling and tracing in pC (see [9] for more details) 5.1 Portable Profiling for C and Languages Based on C A very valuable tool for program tuning is function profiling. Here, special instrumentation code is inserted at all entry and exit points of each function to capture data that can be used to calculate the number of ....

[Article contains additional citation context not shown here]

A. Malony, B. Mohr, P. Beckman, D. Gannon, S. Yang, F. Bodin, Performance Analysis of pC++: A Portable Data-Parallel Programming System for Scalable Parallel Computers, Proc. 8th Int. Parallel Processing Symb. (IPPS), Canc'un, Mexico, IEEE Computer Society Press, pp. 75--85, Apr. 1994.

No context found.

A. Malony, B. Mohr, P. Beckman, S. Yang, F. Bodin. Performance Analysis of pC++: A Portable Dataparallel Programming System Scalable Parallel Computers. In Proceedings of the Eighth International Parallel Processing Symposium, pp. 75-85, 1994.

No context found.

A. Malony, B. Mohr, P. Beckmen, D. Gannon, S. Yang, and F. Bodin. Performance analysis of pC++: A portable data-parallel programming system for scalable parallel computers. In Proceedings of the 8th International Parallel Processing Symposium, 1994.

No context found.

A. Malony, B. Mohr, P. Beckman, D. Gannon, S. Yang, and F. Bodin. Performance Analysis of pC++: A Portable DataParallel Programming System for Scalable Parallel Computers. In H.J. Siegel, editor, Proc. Eighth International Parallel Processing Symposium. IEEE Computer Society Press, April 1994.

No context found.

A. Malony et al. Performance analysis of pC++: A portable data-parallel programming system for scalable parallel computers. In Proceedings of the 8th International Parallel Processing Symposium, 1994.

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