J. C. Yan, `Performance tuning with AIMS---an automated instrumentation and monitoring system for multicomputers', Proc. 27th Hawaii International Conference on System Sciences, Wailea, Hawaii, January 1994.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....Software Technology Inc. Starkville, MS, USA. andrew mpi softtech.com Abstract This paper describes on going development of a performance analysis and tuning tool for parallel MPI [11] applications running on large clusters. Several parallel performance debugging tools such as VaMPIr [5] AIMS [3], and ParaVer [7] exist. Most of the existing tools provide post mortem analysis and rely extensively on program visualization techniques to aid the user with performance bottleneck detection. Post mortem analysis has limited utility for parallel programs that execute for several hours or days, ....

.... analysis of parallel applications using other programming paradigms such as PVM [9] and OpenMP [12] Several tools aiding performance visualization and analysis of parallel applications are available [1] These include SvPablo [8] Paradyn [6] VaMPIr [5] ParaVer [7] Upshot [4] AIMS [3] and other tools. Each of these tools focuses on a different aspect of performance analysis. For instance, Paradyn is a powerful general purpose tool that can be used even in parallel environments but with limited relevance to parallel programs per se. AIMS, Nupshot and VaMPIr are tailored for ....

AIMS. C. Yan. Performance Tuning with AIMS---An Automated Instrumentation and Monitoring System for Multicomputers. Proceedings of the 27th Hawaii international Conference on Systems Sciences, ACM, Jan 1994.

....monitoring system during the execution of the target application. Users can enable, disable, and reconfigure the monitoring system to suit their exploration of the application [19, 17, 9, 6] These interactive monitoring systems complement other forms of monitoring that use post mortem analysis [25, 18, 16, 2, 3], but they share with such systems the need to analyze and display monitoring data [20, 11, 12, 4] The interactive monitoring systems # NASA financially supported Vetter with a Graduate Student Researchers Program Fellowship while he was a Ph.D. candidate at Georgia Tech. This work was also ....

J. Yan, E. Hesham, and B. Shriver. Performance tuning with AIMS-an automated instrumentation and monitoring system for multicomputers. In Proc. Twenty-Seventh Hawaii Int'l Conf. System Sciences. Vol.II: Software Technology, pages 625--33, 1994.

....level meaningful to the user. Our approach differs significantly from previous parallel and distributed debuggers. Most of the prior work falls into two main categories: trace based (i.e. post mortem) debugging systems, some of which have addressed portability, such as Paragraph [3] and AIMS [15], and . the parallel and distributed analogs of conventional state based debuggers, such as ipd [5] or Prism [14] which have not tried to be portable. The only exceptions to this categorization that we could find are distributed debuggers that layer the user interface on top of complete vendor ....

J. C. Yan. Performance Tuning with AIMS---An Automated Instrumentation and Monitoring System for Multicomputers. Proceedings of the 27th Hawaii International Conference on Systems Sciences, ACM, Jan. 1994.

....general purpose systems software. Many distributed ISs have been developed over the past decade (e.g. 3, 8, 5] usually as components of larger software toolkits for analysis of parallel distributed systems. Only a few have been ported to multiple platforms and made available to broader usage [13, 9, 7, 4, 6]. Most of these contain subsystems that whose services could be provided by a common distributed IS kernel. In this paper we present a basic, reference implementation of a portable and flexible IS called BRISK (Baseline Reduced Instrumentation System Kernel) that we have developed as a part of a ....

J. Yan. Performance tuning with AIMS---an automated instrumentation and monitoring system for multicomputers. Proceedings of the Twenty-Seventh Hawaii International Conference on System Sciences, January 1994. 5

....in the traced system to extract useful information for visualization. There are some intrusive overheads of the tracing operation. There are several visualization systems available such as Distributed Object Visualization Environment(DOVE) 3] Automated Instrumentation and Monitoring System(AIMS)[4], Pablo[5] and Falcon [6] These seek to provide avenues for understanding, development, and debugging via visualization. Since they are specialized to specific systems, they cannot support a distributed object system like CORBA or Java. They communicate with other processors using an explicit ....

Jerry Yan, Performance tuning with AIMS-an automated instrumentation, and monitoring system for multicomputers , Proceedings of the Twenty-Seventh Hawaii International Conference on System Sciences, January 1994.

....suggestions to improve the performance of their parallel programs. 1. Introduction The design of performance tuning tools for parallel programming has historically been focused on analysing the data collected in the execution of a parallel program. A number of tools, such as Paragraph [1] AIMS [2], and JEWEL [3] have been developed, which collect and report the performance data. They rely on the user to examine the collected data or visualised performance and compare them with expected values to identify performance problems. This type of tools face a challenge: identifying a performance ....

J. C. Yan, "Performance Tuning with AIMS -- an Automated Instrumentation and Monitoring System for Multicomputers", Proceedings of the Twenty--Seventh Hawaii International Conference on System Science Vol II: Software Technology, Wailea, HI, USA, Jan. 1994, pp. 625--633.

....model results in a perfectly synchronized pipeline operation in which no processor ever waits for data once the first message is received, leading to pipeline fill times and total completion times that are linear functions of the processor number p. We use the performance monitoring tool AIMS [7] (Automated Instrumentation and Monitoring System) to illustrate qualitatively the pipeline behavior of a problem where p = 4 and N = 100, implemented on an Intel iPSC 860 hypercube computer. In Figure 1, horizontal striped bars indicate processor status. Dark sections signify that a processor is ....

J.C. Yan, Performance tuning with AIMS---An automated instrumentation and monitoring system for multicomputers, Proceedings of the 27 th Hawaii International Conference on System Sciences, Vol. II, pp. 625-633, Wailea, HI, January 4-7, 1994.

....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 typically insert instrumentation code to count and or time the execution of fragments of ....

.... 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 with profiling is that detailed ....

J. C. Yan. Performance Tuning with AIMS -- An Automated Instrumentation and Monitoring System for Multicomputers. In Proceedings of the 27th Hawaii International Conference on System Sciences, pages 625--633, Wailea, Hawaii, January 1994.

....0.7 0 1 2 3 4 5 6 7 8 9 Performance Number of Processors Ideal Manual Finesse(Version 6) Figure 11. Final performance curves. of these tools, but, restricting attention to tools which support FORTRAN or C, significant research efforts in this area include Paragraph [5] Pablo [16] AIMS [26], Paradyn [9] PMA [23] and XPVM [4] Commercial systems include Vampir [11] MPP Apprentice [24] Nupshot [25] and VT [2] Tools for shared memory systems have received less attention, possibly due to the lack of a widely accepted standard for the associated programmingparadigm, and because of ....

J.C. Yan, Performance Tuning with AIMS- An Automated Instrumentation and Monitoring System for Multicomputers, Proc. 27th Hawaii Intl. Conf. on System Sciences II, IEEE Comp. Soc. Press, 625--633, Jan. 1994.

....have been used with several environments. PICL [5] is a portable library of efficient communication functions that also supports instrumentation. We have modeled and evaluated two performance data management policies supported by PICL in [35] AIMS (Automated Instrumentation and Monitoring System [39]) is a toolkit consisting of an instrumentation library and a set of off line performance analysis and visualization tools. Its IS support is almost identical to that of PICL. A user can specify different sizes of buffers or usage of flushing functions in a configuration file as part of a static ....

Yan, Jerry, "Performance Tuning with AIMS---An Automated Instrumentation and Monitoring System for Multicomputers," Proc. of the Twenty-Seventh Hawaii Int. Conf. on System Sciences, Hawaii, January 1994.

....a set of actions by the development methodology and the core prediction and analysis tools are integrated, Section 5. The form of the user interface needed to be considered. A number of display types were influential: ffl The space time diagram is common to a number of event trace systems [ 11 ] Its advantage may stem from the persistence of information displayed, allowing the mind to build up a pattern of activity. ffl The diagram based display, such as SPV for the Paragon [ 12 ] is a way of showing process meters indicating parameters such as the instantaneous arrival and ....

J. C. Yan. Performance tuning with AIMS---an automated instrumentation and monitoring system for multicomputers. In 27 th Annual Hawaii International Conference on System Sciences, 1994.

....inserted. This technique is used in tools such as Quantify [7] and qpt [5] Quantify and qpt operate on static binaries whereas Dynamic Instrumentation modifies binaries on the fly. Several trace based tools have been designed for performance tuning parallel applications, such as Pablo [8] AIMS [11], and IPS 2 [6] These systems require an application to be re compiled, then run while trace data is gathered for the entire application, then analyzed post mortem using the trace data. Such trace based approaches may have problems with the amount of trace data needed to analyze a single large, ....

J. C. Yan, "Performance Tuning with AIMS -- An Automated Instrumentation and Monitoring System for Multicomputers", 27th Hawaii International Conference on System Sciences, Jan 4-7 1994, pp. Vol II, pp.

....message passing. 1.5 PVM Parallel Virtual Machine (PVM) 19] is a message passing library designed to support parallel processing on a network of heterogeneous Unix workstations. Although PVM is the most widely used library of its kind [39] there are few performance debugging tools available [19, 31, 38], and none capable of performance prediction. 1.6 ATM Asynchronous Transfer Mode (ATM) is the transmission system standardized by the CCITT and the ATM Forum for the high speed digital network that will replace current telephony networks [11, 32] Small ATM switches are also being used in ....

J. Yan. Performance tuning with AIMS: An automated instrumentation and monitoring system for multicomputers. In Proceedings of the 27th Hawaii International Conference on System Sciences, January 1994. Describes the AIMS performance debugging tool, which runs on several platforms, including PVM.

....overview of the Unify distributed shared memory system which Xunify is designed to evaluate. Section 4 describes the design of the Xunify performance tool, its novel features and the views it supports. 2. Related Work Several tools have been developed to monitor parallel and distributed systems [21, 1, 8, 19, 20, 9] with the majority of distributed monitering systems targeted towards message passing systems. Each system relies on a monitoring component to gather run time information, a facility to route this trace information to a central location and an analysis and visualization system to display the ....

Jerry Yan. Performance Tuning with AIMS--An Automated Instrumentation and Monitoring System for Multicomputers. In proceedings of the 27th HICS, Wailea, Hawaii, January 1994.

....algorithm [2] based on the deadlock avoidance algorithm called null message [3] developed by Chandy and Misra. 2.2. Scalability Performance Metric Predicate Library (PMPL) 5] attempts to solve the problem of performance data size. PMPL can be used with existing trace monitors such as AIMS [18, 19] and PICL [6] It uses user defined predicates (PMPs) to control the output of event trace data during the application s execution. These predicates test the state of the program performance; if the performance does not fall into a specified tolerance, the monitor will output the performance ....

....trace data. With tracing, one can capture components interactions, dynamic behavior, transients, as well as load imbalances due to message waiting. Since the performance information is retrieved from trace files, this technique is usually used in post mortem performance tuning tools, such as AIMS [18, 19], PICL [6] and Pablo [15] These tools present information from the perspective of time, and try to give the user an idea of what was happening in every processor in the system at any given moment. Event traces have great potential to help the user to understand the program behaviour, but also ....

J. Yan, Performance Tuning with AIMS - an Automated Instrumentation and Monitoring System for Multicomputers, In Hesham El-Rewini and B. D. Shriver (eds.), Proceedings of the Twenty-Seventh Hawaii International Conference on System Sciences, Vol. II: Software Technology, IEEE Computer Society Press, USA, Jan. 1994, pp. 625-633.

....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 ....

J. C. Yan. Performance Tuning with AIMS -- An Automated Instrumentation and Monitoring System for Multicomputers. 27th Hawaii International Conference on System Sciences, Wailea, Hawaii, pages 625--633, January 1994. This article was processed using the L A T E X macro package with LLNCS style

....of those encountered in a production environment. ffl Analysis and visualization. After post processing, performance data is visualized and analyzed to identify application program performance bottlenecks (e.g. using text based profiling tools or visualization systems like IPS 2 [8] AIMS [15], or Pablo [9] ffl Application optimization. Based on measurement and analysis, either the program is modified to alleviate the perceived bottlenecks or runtime system policies are adjusted to better match program resource requests. Although effective for application codes with repeatable ....

J. C. Yan. Performance Tuning with AIMS -- An Automated Instrumentation and Monitoring System for Multicomputers. In Proceeedings of the 27th Hawaii International Conference on System Sciences, pages 625--633, Jan. 1994.

.... process periodically determine the state of an running program, constructing an approximate profile based on these samples [GKM82] Alternatively, special recording instructions can be explicitly added to an executable at appropriate locations (some methods are described in [ROA 91, BBG92, Yan94, LB92] and others are discussed later) Such techniques can also be used to obtain precise profiles of optimized programs [NEC92] In both the manual and automatic cases, such profiling can provide increasingly detailed execution information, but at a cost. The additional information collection ....

Jerry C. Yan. Performance tuning with AIMS --- an automated instrumentation and monitoring system for multicomputers. In Hesham El-Rewini and Bruce D. Shriver, editors, Proceedings of the 27th Hawaii International Conference on System Sciences (HICSS27, Volume II), pages 625--633. IEEE Computer Society Press, 1994.

....message passing. 1.5 PVM Parallel Virtual Machine (PVM) 19] is a message passing library designed to support parallel processing on a network of heterogeneous Unix workstations. Although PVM is the most widely used library of its kind [39] there are few performance debugging tools available [19, 31, 38], and none capable of performance prediction. 1.6 ATM Asynchronous Transfer Mode (ATM) is the transmission system standardized by the CCITT and the ATM Forum for the high speed digital network that will replace current telephony networks [11, 32] Small ATM switches are also being used in ....

J. Yan. Performance tuning with AIMS: An automated instrumentation and monitoring system for multicomputers. In Proceedings of the 27th Hawaii International Conference on System Sciences, January 1994. Describes the AIMS performance debugging tool, which runs on several platforms, including PVM.

....performance tools [52] it can be used with any visualization system that accepts SDDF files. The Paradyn [42] performance tools, which are designed for runtime performance monitoring and bottleneck detection, have been ported to the PVM platform. Other debuggers are available, including Aims [61], Xab [5] and Xmdb [19] Some graphical programming environments are available for PVM, including HeNCE [6] and VPE [44] In addition to visual programming facilities, these environments also provide some debugging capabilities. Some performance problems are more easily solved with performance ....

J. Yan. Performance Tuning with AIMS:An Automated Instrumentation and Monitoring System for Multicomputers. In Proceedings of the 27th Hawaii International Conference on System Sciences, January 1994.

....configurations, and retrieve data. In x3 x5, we will examine all of these issues in greater detail. 3 Flexible Performance Sensors Historically, performance instrumentation required users, parsers, or compilers to insert instrumentation library calls in application source code. IPS 2 [15] AIMS [26], Pablo [17] and other popular performance analysis toolkits all operated in this way. After instrumentation insertion, the application code was recompiled and performance data was collected during subsequent execution. While this approach has proven successful for off line performance analysis, ....

Jerry C. Yan. Performance Tuning with AIMS -- An Automated Instrumentation and Monitoring System for Multicomputers. In Proceeedings of the 27th Hawaii International Conference on System Sciences, pages 625--633, January 1994.

No context found.

J. C. Yan, `Performance tuning with AIMS---an automated instrumentation and monitoring system for multicomputers', Proc. 27th Hawaii International Conference on System Sciences, Wailea, Hawaii, January 1994.

....partitions of a multiprocessor . attempting to compensate for clock skews across different nodes . supporting visualization of multi partition computation . offering inter partition and intra partition communication analysis In this paper we address all the above issues in extending AIMS [8] to support the performance monitoring of multidisciplinary applications. In the next section, we will briefly describe the intercube communication library that supports the development of multidisciplinary applications on the iPSC 860. In Section 3, the general framework of the AIMS toolkit is ....

....partition communications. 3: AIMS An Automated Instrumentation and Monitoring System The Automated Instrumentation and Monitoring System (AIMS) was originally designed to facilitate performance evaluation of parallel applications on multicomputers by capturing and visualizing execution data [8]. AIMS has four major components: a source code instrumentor (xinstrument) a monitoring library (monitor) a trace pre processor (pc) and a performance analysis and visualization toolkit (SK, VK) Currently, AIMS accepts FORTRAN and C parallel programs written for three message passing ....

J. C. Yan. "Performance Tuning with AIMS --- An Automated Instrumentation and Monitoring System for Multicomputers". In Proceedings of the 27 th Hawaii International Conference on System Sciences, Wailea, Hawaii, January 4 - 7, 1994. Vol. II. pages 625-633

....ParAide for Intel s Paragon, MPP Apprentice for CRI s T 3D, PRISM for TMC s CM 5, PV for IBM s SP2 and CXTrace for Convex s SPP1. Many public domain tools have also emerged from various national laboratories and universities including UIUC s Pablo [RO 91] ORNL s XPVM [Kohl95] and NASA s AIMS [Yan94, YSM94]. A few observations can be made: Tools provided by the vendors only run on their own machines (as expected) Compounded with unreliable system software, a lot of these sophisticated performance tools are glanced over by many users who prefer gprof (which is supported in one form or another by ....

J. C. Yan. "Performance Tuning with AIMS --- An Automated Instrumentation and Monitoring System for Multicomputers". In Proceedings of the 27th Hawaii International Conference on System Sciences, Wailea, Hawaii, January 4 - 7, 1994. Vol. II. pages 625-633.

No context found.

J. Yan, Performance Tuning with AIMS--An Automated Instrumentation and Monitoring System for Multicomputers, in proceedings of the 27th HICSS, Wailea, Hawaii, January 1994.

First 50 documents

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