J. May, F. Berman, "Panorama: A Portable Extensible Parallel Debugger", Proceedings of the ACM/ONR Workshop on Parallel and Distributed Debugging, San Diego, May 1993, pp 96-106.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....since the introduction of parallel computing architectures. Most of these debuggers look and behave like debuggers for conventional architectures, but are usually enhanced to support the management and control of multiple processes. Many of these debuggers employ a client server mechanism [9] 12][16][21] Some, such as DETOP [8] have been developed to support both task and data parallel codes. A relative debugger must also provide the functionality of a conventional parallel debugger, at least internally, because in order to carry out comparisons the debugger must ensure that processes can ....

J. May, F. Berman, "Panorama: A Portable Extensible Parallel Debugger", Proceedings of the ACM/ONR Workshop on Parallel and Distributed Debugging, San Diego, May 1993, pp 96-106.

....functionality of the system is manipulation of large distributed data objects to, for example, investigate convergence and stability of algorithms used in scientific simulations. Typically, a client server architecture is used to implement a portable debugger for distributed systems (c.f. [12 17]) Our approach is unique in that we use a three tier architecture. Therefore we can easily integrate our source level debugger with the HPF interpreter and a visualization tool which together comprise a powerful application development environment. 6. Integrated Environment for HPF Compiler and ....

J. May & F. Berman, "Panorama: A portable, Extensible Parallel Debugger," Proceedings of ACM/ONR Workshop on Parallel and Distributed Debugging, May 1993, pp. 96-106

....GUARD. It then describes the application under consideration which is known as the Shallow Water Equations [16] This is followed by a discussion of how GUARD facilitates data comparison across distributed memory machines. 2 Relative Debugging Unlike other conventional parallel debuggers [6] 12][13] a relative debugger provides the ability to dynamically compare data between two executing programs regardless of their location and configuration. In our implementation of a relative debugger, called GUARD, data comparisons can be performed either using an imperative scheme or a declarative ....

May, J., Berman, F., "Panorama: A Portable Extensible Parallel Debugger", Proceedings of the ACM/ONR Workshop on Parallel and Distributed Debugging, pp. 96 -- 106, San Diego, May 1993.

....all traced events. Visualization level has to deal with the presentation of large volumes of resulting data that may be produced by massively parallel applications. The trace collection issue is not a recent one. A number of user tools are heavily based on traces, including correctness debuggers [5, 17, 26, 16, 22, 15], performance debuggers[9, 10, 19, 18, 8] trace driven simulators [3, 28, 6, 24, 4] etc. A broad classification [25] of tracing techniques distinguishes four basic classes: hardware based methods which use a hardware monitor to record all requests on the address bus of a processor, ....

J. May and F. Berman. "Panorama: A Portable, Extensible Parallel Debugger". ACM SIGPLAN, 28(12), December 1993.

....functionality of the system is manipulation of large distributed data objects to for example, investigate the convergence and stability of algorithms used in scientific simulations. Typically, a client server architecture is used to implement a portable debugger for distributed systems (c.f. [12 17]) Our approach is unique in that we use a three tier architecture. Therefore, we can easily integrate our source level debugger with the HPF interpreter and a visualization tool which together comprise a powerful application development environment. 6. Integrated Environment for HPF Compiler and ....

J. May & F. Berman, "Panorama: A portable, Extensible Parallel Debugger," Proceedings of ACM/ONR Workshop on Parallel and Distributed Debugging, May 1993, pp. 96-106

....traced events. The visualization level has to deal with the presentation of large volumes of result data that may be produced by massively parallel applications. The trace collection issue is not a recent one. A number of user tools are heavily based on traces, including correctness debuggers [4, 15, 24, 14, 20, 13], performance debuggers[8, 9, 17, 16, 7] trace driven simulators [2, 26, 5, 22, 3] etc. A broad classification [23] of tracing techniques distinguishes four basic classes: hardware based methods which use a hardware monitor to record all requests on the address bus of a processor, ....

J. May and F. Berman. "Panorama: A Portable, Extensible Parallel Debugger". ACM SIGPLAN, 28(12), December 1993.

....a full recompilation process. The modifications generally include minor changes to source code and creation of fast conditional breakpoints and watch points. Future work on Mantis includes the addition of such features along with performance tuning tools. 4. 4 Other Parallel Debuggers Panorama [5] was developed to integrate the myriad techniques developed for parallel debugging in an easily portable and extensible manner. It uses Tcl Tk to enhance portability and extensibility, translating between its own portable debugging interface and that used by the machine vendor s debugger via a ....

J. May, F. Berman, "Panorama: a portable, extensible parallel debugger," SIGPLAN Notices, pp. 96-106, December 1993.

....algorithm uses O(n 4:38 ) processors while the Helmbold and Mayer s algorithm uses O(n 10 ) processors. Several other researchers have considered restricting the precedence constraints and allowing the number of procesors to vary. If the precedence constraints are restricted to forests [2,8], then optimal schedules can be found either sequentially or in parallel. If the precedence constraints are restricted to interval orders, then a sequential algorithm has been proved by Papadimitriou and Yannakakis [10 ] Formally, all graphs considered in what follows are finite, undirected, ....

....disjoint set union, J. Assoc. Comput. Mach. 30 (1985) 209 221. 6] M. Garey and D. Johnson, Computers and Intractability, A guide to the theory of NPcompleteness, Freeman, San Francisco, CA, 1979. 7] D. Helmbold and E. Mayer, Two processor scheduling is in NC, SIAM J. Computing 16 (1987) 47 759. [8] T. Hu, Parallel sequencing and assembly line problems, Operations Research 9 (1961) 84184. 9] L. Lovasz and M. Plummer, Matching Theory, North Holland Mathemtics Studies, 1986. 10] C. Papadimitriou and M. Yannakakis, Scheduling interval ordered tasks, SIAM J. Computing 8 (1979) 405 409. 11] ....

[Article contains additional citation context not shown here]

John May, Francine Berman, "Panorama: A Portable, Extensible Parallel Debugger", ACM/ONR Workshop on Parallel and Distributed Debugging (in ACM SIGPLAN Vo.28, No.12, December 1993), pp. 96-106.

....have automated much of this task. We have avoided defining the internal structure of a program view so that users will have maximum flexibility in gathering, analyzing, and presenting information. Programmers can use our tool to create displays for the Panorama parallel debugger. An earlier paper [4] described Panorama s structure and operation. The system offers extensible displays, and it is designed to work with a variety of message passing multicomputers. We describe its retargeting features more fully in another paper [5] The present paper focuses on the features of Panorama that allow ....

J. May and F. Berman, "Panorama: A portable, extensible parallel debugger," in Proceedings of the ACM/ONRWorkshop on Parallel and DistributedDebugging, pp. 96--106, May 1993.

....the vendors tools and third party tools, but it requires the active cooperation of the vendors, and so far it is in the early stages of development. To deal with the variety of interfaces and capabilities in the vendor debuggers that are available now, we have created a system called Panorama [2]. Panorama runs on a user s workstation and uses the base debugger supplied by the vendor to control the target program running on the parallel machine. This paper describes how Panorama can be adapted to work with new machines using a minimum of platform dependent code. 1 The strategy we ....

....of these low level functions. Panorama is written in a combination of C and Tcl [4] Tcl is a language and an interpreter developed to make programs easier to extend and customize, and it also makes certain features of our debugging strategy particularly easy to implement. An earlier paper [2] describes Panorama s user interface and other features in more detail. The remainder of this paper focuses on the strategies we use to make Panorama retargetable. 3 Templates and regular expressions In some cases, the commands for doing the same job on different base debuggers differ only in ....

J. May and F. Berman, "Panorama: A portable, extensible parallel debugger," in Proceedings of the ACM/ONR Workshop on Parallel and Distributed Debugging, pp. 96--106, May 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