M. T. Heath, "Recent developments and case studies in performance visualization using paragraph," Performance Measurement and Visualization of Parallel Systems, pp. 175--200, 1993.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....parameters. However, the most commonly output facility is the creation of predictive trace files [6] These are analogous to traditional traces obtained at run time but contain predicted time information. The user can employ any monitoring tool to visualise the various performance aspects (e.g. [4,11]) An example predictive trace output is shown in Figure 6 which details a segment of the predictive behaviour of IPKERNEL on seven workstations. 5 subtask application Application layer Parallel template layer Hardware layer thresh blobcolour async mastcomm async swap async side async ....

M.Heath, Recent Developments and Case Studies in Performance Visualization Using Paragraph, In: Performance Measurement and Visualization of Parallel Systems, Elsevier Science Publishers, 1993.

....overview about monitoring and profiling systems. The hardware counters are accessed by using the PCL library. 1] The PerfAPI project [9] is a standardization effort for accessing of the hardware performance counters. Trace based profiling and analysis systems are described in the references [2] [3], 5] 10] 11] HP[6] has developed a local, user callable MPI counter profiling. Li and Zhang[7] combine counter profiling and a virtual clock approach to minimize the intrusiveness of the instrumentation. The new syslog based MPIcounter profiling closes a gap between the existing ....

M. T. Heath, Recent Developments and Case Studies in Performance Visualization using ParaGraph, Proceedings of the Workshop Performance Measurement and Visualization of Parallel Systems, G. Haring and G. Kotsis (ed.), Moravany, Czechoslovakia, Oct. 1992, pp 175--200.

....user MPI profiling tools. The system accounting gives an overview over the whole system, mainly about the usage of CPU and I O resources, but does not report any details about specific MPI routines. The existing MPI instrumentation libraries and profiling tools are trace based, like ParaGraph [2] [3], Upshot [4] and VAMPIR [6] 7] or counter based, as described e.g. in [5] In all cases they must be enabled by the user and they report only to the user. The key issue for syslog based MPI counter profiling is, that the overhead of the instrumentation must be reduced to a minimum in contrast ....

M. T. Heath, Recent Developments and Case Studies in Performance Visualization using ParaGraph, Proceedings of the Workshop Performance Measurement and Visualization of Parallel Systems, G. Haring and G. Kotsis (ed.), Moravany, Czechoslovakia, Oct. 1992, p. 175-200.

....a weekly mail. 5 Related and Future Work [1] describes the hardware counter library PCL. 5] is a standardization effort for accessing of the hardware performance counters. 10] is a comprehensive overview about monitoring and profiling systems. Trace based profiling and analysis is described in [2, 3, 6, 7]. 4] describes a local, user callable MPI counter profiling. 8] focuses on the global view of the counter based MPI profiling and includes the statistical results of half a year of profiling nearly all MPI applications running on a CRAY T3E 900 512, but without instrumenting the applications ....

M. T. Heath, Recent Developments and Case Studies in Performance Visualization using ParaGraph, Proceedings of the Workshop Performance Measurement and Visualization of Parallel Systems, G. Haring and G. Kotsis (ed.), Moravany, Czechoslovakia, Oct. 1992, p. 175-200.

....25 5.6.1 Upshot and Nupshot One tool that we use is called upshot, which is a derivative of Upshot [13] written in Tcl Tk. A screen dump of Upshot in use is shown in Figure 2. It shows parallel time lines Figure 2: A screendump from upshot with process states, like one of the paraGraph [12]. The view can be zoomed in or out, horizontally or vertically, centered on any point in the display chosen with the mouse. In Figure 2, the middle window has resulted from zooming in on the upper window at a chosen point to show more detail. The window at the bottom of the screen show a histogram ....

M. T. Heath. Recent developments and case studies in performance visualization using ParaGraph. In G. Haring and G. Kotsis, editors, Performance Measurement and Visualization of Parallel Systems, pages 175--200. Elsevier Science Publishers, 1993.

....results were reported in [4] for a traditional parallelization of branch and bound algorithms. 5 ParVis: A Tool for Visualizing Adaptive Techniques There are many well known systems for visualizing and debugging distributed parallel program performance among the best known are ParaGraph [23,24], the TAU environment [32] and Ariane [29,39] Programmer oriented visualizations have been incorporated into several parallel programming systems such as XPVM [28] and AIMS [40] Following the approach of interactive visual exploration [22] we are developing a visualization system (see also ....

M. T. Heath, Recent developments and case studies in performance visualization using PARAGRAPH, in Performance Measurement and Visualization of Parallel Systems, G. Haring and G. Kotsis, eds., Elsevier Science Publishers, 1990, pp. 175-200.

....as the program runs. 5.3 upshot One tool that we use is called upshot, which is a derivative of Upshot [6] written in Tcl Tk. A screen dump of Upshot in use is shown in Figure 1. It shows parallel time lines with Figure 1: A screendump from upshot process states, like one of the paraGraph [5]. The view can be zoomed in or out, horizontally or vertically, centered on any point in the display chosen with the mouse. In Figure 1, the middle window has resulted from zooming in on the upper window at a chosen point to show more detail. The window at the bottom of the screen show a histogram ....

M. T. Heath. Recent developments and case studies in performance visualization using ParaGraph. In G. Haring and G. Kotsis, editors, Performance Measurement and Visualization of Parallel Systems, pages 175--200, Amsterdam, The Netherlands, 1993. Elsevier Science Publishers.

....runs. 5.5 upshot and nupshot One tool that we use is called upshot, which is a derivative of Upshot [13] written in Tcl Tk. A screen dump of Upshot in use is shown in Figure 2. It shows parallel time lines Figure 2: A screendump from upshot with process states, like one of the paraGraph [12]. The view can be zoomed in or out, horizontally or vertically, centered on any point in the display chosen with the mouse. In Figure 2, the middle window has resulted from zooming in on the upper window at a chosen point to show more detail. The window at the bottom of the screen show a histogram ....

M. T. Heath. Recent developments and case studies in performance visualization using ParaGraph. In G. Haring and G. Kotsis, editors, Performance Measurement and Visualization of Parallel Systems, pages 175-200. Elsevier Science Publishers, 1993.

....during the observed application. Tape Pvm also includes an intrusion compensation tool, based on perturbation models, which attempts to remove the additional delays due to tracing from the resulting traces. Traces generated by Tape PVM were converted into the PICL format accepted by ParaGraph [7] to visualize the execution of parallel programs (see figure 6) 6 Mapping Tool Simulation Tool integration (UAB) In the simulation tool developed at the UAB there exists the possibility of using an automatic mapping tool also developed at the UAB. The mapping tool is invoked from the mapping ....

M.T. Heath. Recent Developments and Case Studies in Performance Visualization using ParaGraph. In Workshop on Performance Measurement and Visualization of Parallel Systems, Moravany, Czecho-Slovakia, October 1992.

....five years, visualization technologies has been applied to manage the vast amount of performance data collected in multiprocessors. Many tools supporting performance visualization are available today, either in the public domain (e.g. AIMS [YSM95] Pablo [RO 91] Paradyn [MC 95] and ParaGraph [Hea93]) or as part of a multiprocessor s system software (e.g. CXTRACE on Convex s SPP 1, ParAid on Intel s Paragon, MPP Apprentice on CRI s T3D, PRISM on TMC s CM5, and PV on IBM s SP2) Typically, performance data are displayed postmortem. These displays provide valuable feed back about users ....

M. T. Heath, "Recent Developments and Case Studies in Performance Visualization using ParaGraph," in Performance Measurement and Visualization of Parallel Systems, ed. G. Haring and G. Kotsis, Elsevier Science, 1993, pp. 175-200.

....the advances that have been made in visualizing scientific data, techniques for visualizing performance of parallel programs remain ad hoc and performance visualization becomes more difficult as the parallel system becomes more complex. Popular views like space time Feynman diagram (ParaGraph [9, 10]) or event timelines (Express Etool [16] often provide some insight about program behaviour. Unfortunately, for these tools to be truly useful in the domain of large scale parallel machines they must include abstract high level views [20] This situation motivates the use of AVS (Application ....

Michael Heath. Recent Developments and Case Studies in Performance Visualization using ParaGraph. In Workshop on Performance Measurement and Visualization of Parallel Systems, Moravany, Czecho-Slovakia, October 1992.

....the time taken to solve a particular problem, end users often have less confidence in a parallel solution than the corresponding uniprocessor solution. There are many well known systems for visualizing and debugging distributed parallel program performance among the best known are ParaGraph [4, 5], the TAU environment [8] and Ariane [7, 10] Programmer oriented visualizations have been incorporated into several parallel programming systems such as XPVM [6] and AIMS [11] Visualizations geared to programmers do not necessarily help end users understand computations or gain confidence in ....

HEATH, M. T. Recent developments and case studies in performance visualization using PARAGRAPH. In Performance Measurement and Visualization of Parallel Systems, G. Haring and G. Kotsis, Eds. Elsevier Science Publishers, 1990, pp. 175--200.

....the behaviour of a parallel program. In recent years, a wide variety of performance visualization tools have been developed for this purpose using different visualization techniques to understand complex performance processes (e.g. Express Tools [1] SIMPLE [2] PA Tools [3] ParaGraph [4] [5], TOPSYS PATOP [6] and PABLO [7] A more detailed survey of Performance Tools is contained in reference [8] Performance visualization is the use of graphical display techniques to present an analysis of performance data for an improved understanding of complex performance phenomena [9] ....

....advances that have been made in visualizing scientific data, techniques for visualizing performance of parallel programs remain ad hoc and performance visualization becomes more difficult as the parallel system becomes more complex. Popular views like space time Feynman diagram (ParaGraph [4] [5]) or event timelines (Express Etool [1] often provide some insight about program behaviour. Unfortunately, for these tools to be truly useful in the domain of large scale parallel machines they must include abstract high level views [10] Massively parallel processors can generate a huge amount ....

[Article contains additional citation context not shown here]

Michael Heath, "Recent Developments and Case Studies in Performance Visualization using ParaGraph", in Workshop on Performance Measurement and Visualization of Parallel Systems, Moravany, Czecho-Slovakia, October 1992.

.... the advances that have been made in visualizing scientific data, techniques for visualizing performance of parallel programs remain ad hoc and performance visualization becomes more difficult as the parallel system becomes more complex [8] Popular views like space time Feynman diagram (ParaGraph [9, 10]) or event timelines (Express Etool [11] often provide some insight into program behaviour. Unfortunately, for these tools to be truly useful in the domain of large scale parallel machines they must be extended to include abstract high level views [12] Massively parallel processors can generate ....

....[ 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 of ParaGraph. 2.3.3.1 General description The principal design objectives of the original version of ParaGraph were ease of understanding, ease of use and ....

[Article contains additional citation context not shown here]

Michael Heath, "Recent Developments and Case Studies in Performance Visualization using ParaGraph", in Workshop on Performance Measurement and Visualization of Parallel Systems, Moravany, Czecho-Slovakia, October 1992.

....Libraries and MISS PVM 11 Visualizing the Output Using ParaGraph. ParaGraph takes the full name of the output le (including the extension .trf) as parameter. Visualization takes place in ParaGraph via various animation windows, which are, to a great extent, self explanatory (see Heath [12], Tomas, Ueberhuber [16] 2.3 Communication Libraries and MISS PVM Once the high level library routines (based on PVM) have been recompiled, no further modi cations are needed. The user program is linked with the new library. The result is a program that performs the same task as before, except ....

M. T. Heath, Recent Developments and Case Studies in Performance Visualization using ParaGraph, Performance Measurement and Visualization of Parallel Systems, (G. Haring, G. Kotsis, Eds.), Elsevier Science Publishers, Amsterdam, 1993, pp. 175-200. 39

....You can analyze the logfile produced at the end with a variety of tools. One that we use is called Upshot, which is a derivative of Upshot [5] written in Tcl Tk. A screen dump of Upshot in use is shown in Figure 1. It shows parallel time lines with process states, like one of the paraGraph [4]. The view can be zoomed in or out, horizontally or vertically, centered on any point in the display chosen with the mouse. In Figure 1, the middle window has resulted from zooming in on the upper window at a chosen point to show more detail. The window at the bottom of the screen show a histogram ....

M. T. Heath. Recent developments and case studies in performance visualization using ParaGraph. In G. Haring and G. Kotsis, editors, Performance Measurement and Visualization of Parallel Systems, pages 175--200, Amsterdam, The Netherlands, 1993. Elsevier Science Publishers.

....to this name. Visualizing the output using ParaGraph. ParaGraph takes the full name of the output le (with the extension .trf) as parameter. Visualization takes place in ParaGraph via various animation windows, which are, to a great extent, self explanatory (see Tomas, Ueberhuber [15] Heath [8]) 2.3 Communication Libraries and MISS PVM Once the high level library routines (based on PVM) have been recompiled, no additional modi cations are necessary. The user program is linked with the new library. The result is a program that performs the same as before the library has been ....

Heath M.T., Recent Developments and Case Studies in Performance Visualization using ParaGraph, in Performance Measurement and Visualization of Parallel Systems (G. Haring and G. Kotsis, eds.), Elsevier Science Publishers, Amsterdam, The Netherlands, 1993, pp. 175-200.

....tools which are based on extensive studies of existent systems and try to provide report summaries about potential sources of bottlenecks. HPAM Sim takes advantage of these tools by generating a log file in PICL format [28] This log file can be then analyzed and graphically displayed by ParaGraph [16]. There are instances where the generated log file requires a large amount of disk storage and causes high I O overhead. For these instances, HPAM Sim allows the user to select a concise summary of the simulation results rather than the regular log file. 3.1 Augmentation Module The augmentation ....

M. Heath. Recent Developments and Case Studies in Performance Visualization using ParaGraph. Performance Measurementand Visualization of Parallel Systems, Elsevier Science Publishers, pages 175--200, 1993.

....SparcStations, and SGIs connected by a LAN) a trace post processor that compensates for data collection overhead, and a trace animation facility that supports simultaneous visualization of computation and communication patterns as well as analysis of data movements. 4.6. 2 ParaGraph ParaGraph [56] is a graphical display system developed at Oak Ridge Laboratory for visualizing the behavior and performance of parallel programs on message passing parallel computers. It takes as input execution trace data provided by PICL or PVM, which optionally produce an execution trace during an actual run ....

M. Heath. Recent developments and case studies in performance visualization using ParaGraph. In Performance Measurement and Visualization of Parallel Systems, pages 175-- 200. Elsevier Science Publishers, 1993.

....end can be analyzed by a variety of tools. One that we use is called Teeshot, which is a derivative of Upshot [3] written in Tcl Tk. A screen dump of Teeshot in use is shown in Figure 2. It shows parallel time lines with process states, like one of Fig. 2. A screendump from Teeshot the paraGraph [2]. The view can be zoomed in or out, horizontally or vertically, centered on any point in the display chosen with the mouse. In Figure 2, the middle window has resulted from zooming in on the upper window at a chosen point to show more detail. The window at the bottom of the screen show a histogram ....

M. T. Heath, Recent developments and case studies in performance visualization using ParaGraph, in Performance Measurement and Visualization of Parallel Systems, G. Haring and G. Kotsis, eds., Amsterdam, The Netherlands, 1993, Elsevier Science Publishers, pp. 175-- 200.

No context found.

M. T. Heath, "Recent developments and case studies in performance visualization using paragraph," Performance Measurement and Visualization of Parallel Systems, pp. 175--200, 1993.

No context found.

M. T. Heath, Recent Developments and Case Studies in Performance Visualization using ParaGraph, Performance Measurement and Visualization of Parallel Systems, (G. Haring, G. Kotsis, Eds.), Elsevier Science Publishers, Amsterdam, 1993, pp. 175-200. 39

No context found.

M. T. Heath, Recent Developments and Case Studies in Performance Visualization using ParaGraph, Proceedings of the Workshop Performance Measurement and Visualization of Parallel Systems, G. Haring and G. Kotsis (ed.), Moravany, Czechoslovakia, Oct. 1992, pp 175--200.

No context found.

M. T. Heath, Recent Developments and Case Studies in Performance Visualization using ParaGraph, in Performance Measurement and Visualization of Parallel Systems, ed. G. Haring and G. Kotsis, Elsevier Science, 1993, pp. 175-200.

No context found.

M. T. Heath, Recent Developments and Case Studies in Performance Visualization using ParaGraph, Proceedings of the Workshop Performance Measurement and Visualization of Parallel Systems, G. Haring and G. Kotsis (ed.), Moravany, Czechoslovakia, Oct. 1992, p. 175-200.

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