scispace - formally typeset
Search or ask a question
Proceedings ArticleDOI

Probe acquisition for the MSPARC hybrid monitor

12 Apr 1992-pp 323-326
TL;DR: The authors give an overview of the monitor system and design considerations regarding the probe software, which provides initialization, probe processing, communications, and debug capabilities for a monitor node.
Abstract: The MSPARC multicomputer uses processor-based monitoring hardware and probe acquisition software for instrumentation purposes. The authors give an overview of the monitor system and design considerations regarding the probe software. Each node of the mesh connected MSPARC consists of a numerical element and a communications board used to pass messages via custom routers located on a backplane. In addition to the message-passing circuitry, the communications board also contains the instrumentation and monitoring hardware. The monitor and its separate interconnection network are used to assist in application performance monitoring, program design, hardware and software debugging, and hardware analysis. Performance data are transferred from an application program to the monitor system in the form of software probes. The software provides initialization, probe processing, communications, and debug capabilities for a monitor node. >
Citations
More filters
Proceedings ArticleDOI
J. Harden1, D. Reese1, F. To1, D. Linder1, C.B. Borchert1, G. Jones1 
12 Apr 1992
TL;DR: A hybrid performance monitor developed for MSPARC, a mesh-connected, message-passing multicomputer, is described, where data selection and presentation techniques played an important role in the visualization of the parallel system's execution.
Abstract: A hybrid performance monitor developed for MSPARC, a mesh-connected, message-passing multicomputer, is described. The development of the hybrid performance monitor was a cross-disciplinary enterprise requiring custom hardware and a range of software support including monitor code, driver interfaces, probe history acquisition and processing, graphical display, and application probe injection. Programmable hardware was designed to unobtrusively collect events on each node and maintain their accurate chronological order. This distributed collection system was coupled by its independent network to a central monitor where data selection and presentation techniques played an important role in the visualization of the parallel system's execution. >

13 citations


Cites background or methods from "Probe acquisition for the MSPARC hy..."

  • ...The monitor directly uses a probe pointer value to identify the event processing routine [6]....

    [...]

  • ...To balance overhead and cost, a hybrid scheme has gained acceptance in a number of recent designs: the test and measurement processor (TMP hardware) for the INCAS system [2], MCC’s ES–Kit performance monitor [8], Intel iPSC/2 [7] and MSU’s MSPARC multicomputer performance monitor [3][6][13]....

    [...]

  • ...idle time start and end will automatically place software probes [6] that simply write a few words to “magic” memory locations....

    [...]

Proceedings ArticleDOI
12 Apr 1992
TL;DR: Developing of a real-time monitoring environment to provide distributed programmers with valuable information about an executing program as events occur and design and implementation of analysis software which accepts user requests, queries and receives event histories from the instrumentation system are discussed.
Abstract: The author discusses development of a real-time monitoring environment to provide distributed programmers with valuable information about an executing program as events occur. Of particular importance is design and implementation of analysis software which accepts user requests, queries and receives event histories from the instrumentation system, and provides sufficient program abstractions to support graphical displays and metric evaluations of execution progress. The analysis software has been named DECIPHER. The DECIPHER distributed debugging system is described. The primary obstacle involves organizing and managing voluminous event histories in real-time. Both efficient organization and data management are crucial to effectiveness and accuracy. >

8 citations

References
More filters
Journal ArticleDOI
TL;DR: A status report on the architecture and programming of a family of concurrent computers that are organized as ensembles of small programmable computers called nodes, connected by a message-passing network, each with its own private memory is provided in this article.
Abstract: A status report is provided on the architecture and programming of a family of concurrent computers that are organized as ensembles of small programmable computers called nodes, connected by a message-passing network, each with its own private memory. The architecture of the multicomputer is described and contrasted with that of the shared-memory multiprocessor, and the concept of grain size (which depends on the size of the individual memories) is explained. Medium-grain and fine-grain multicomputers, with nodes containing megabytes and tens of kilobytes of memory, respectively, are examined, and their programming is discussed. >

532 citations

Proceedings ArticleDOI
J. Harden1, D. Reese1, F. To1, D. Linder1, C.B. Borchert1, G. Jones1 
12 Apr 1992
TL;DR: A hybrid performance monitor developed for MSPARC, a mesh-connected, message-passing multicomputer, is described, where data selection and presentation techniques played an important role in the visualization of the parallel system's execution.
Abstract: A hybrid performance monitor developed for MSPARC, a mesh-connected, message-passing multicomputer, is described. The development of the hybrid performance monitor was a cross-disciplinary enterprise requiring custom hardware and a range of software support including monitor code, driver interfaces, probe history acquisition and processing, graphical display, and application probe injection. Programmable hardware was designed to unobtrusively collect events on each node and maintain their accurate chronological order. This distributed collection system was coupled by its independent network to a central monitor where data selection and presentation techniques played an important role in the visualization of the parallel system's execution. >

13 citations

Proceedings ArticleDOI
02 Dec 1991
TL;DR: This paper discusses the object oriented Fortran language and support routines developed at Mississippi State in support of parallelizing complex field simulations and is implemented on top of the Unix operating system for portability.
Abstract: Parallel programming has to date remained inaccessible to the average scientific programmer. Parallel programming languages are generally foreign to most scientific applications programmers who only speak Fortran. Automatic parallelization techniques have so far proved unsuccessful in extracting large amounts of parallelism from sequential codes and do not encourage development of new, inherently parallel algorithms. In addition, there is a lack of consistency of programmer interface across architectures which requires programmers to invest a lot of effort in porting code from one parallel machine to another. This paper discusses the object oriented Fortran language and support routines developed at Mississippi State in support of parallelizing complex field simulations. This interface is based on Fortran to ease its acceptance by scientific programmers and is implemented on top of the Unix operating system for portability. >

8 citations