Declustering: a new multiprocessor scheduling technique

Sih, G.C.; Lee, Edward A.

doi:10.1109/71.242160

Cited by 60 publications

(26 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…An indivisible load (or job) is a load that cannot be divided so that all of the load must be processed by one processor. There has been intensive work on indivisible load theory by many parallel and distributed system researchers [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13]. Only more recently has there been interest in multiprocessor scheduling with loads that need to be assigned to more than one processor [14], [15], [16].…”

Section: Introductionmentioning

confidence: 99%

Optimizing computing costs using divisible load analysis

Sohn

Robertazzi

Luryi

1998

IEEE Trans. Parallel Distrib. Syst.

View full text Add to dashboard Cite

Abstract-A bus oriented network where there is a charge for the amount of divisible load processed on each processor is investigated. A cost optimal processor sequencing result is found which involves assigning load to processors in nondecreasing order of the cost per load characteristic of each processor. More generally, one can trade cost against solution time. Algorithms are presented to minimize computing cost with an upper bound on solution time and to minimize solution time with an upper bound on cost. As an example of the use of this type of analysis, the effect of replacing one fast but expensive processor with a number of cheap but slow processors is also discussed. The type of questions investigated here are important for future computer utilities that perform distributed computation for some charge.

show abstract

Section: Introductionmentioning

confidence: 99%

Optimizing computing costs using divisible load analysis

Sohn

Robertazzi

Luryi

1998

IEEE Trans. Parallel Distrib. Syst.

View full text Add to dashboard Cite

show abstract

“…The classic approach is based on the so-called list scheduling technique Adam et al (1974);Coffman et al (1976). More recent approaches are the UNC (Unbounded Number of Clusters) Scheduling Gerasoulis et al (1992); Sarkar (1989), the BNP (Bound Number of Processors) Scheduling Adam et al (1974); Kruatrachue et al (1987); Sih et al (1993), the TDB (Task Duplication Based) Scheduling Colin et al (1991); Kruatrachue et al (1988), the APN (Arbitrary Processor Network) Scheduling Rewini et al (1990), and the genetic Hou et al (1994); Shahid et al (1994). Our problem differs from the multiprocessor scheduling precedence-constrained task graph problem in many factors.…”

Section: Related Workmentioning

confidence: 99%

w-TG: A Combined Algorithm to Optimize the Runtime of the Grid-Based Workflow Within an SLA Context

Quân¹,

Altmann²,

Yang³

2012

Grid Computing - Technology and Applications, Widespread Coverage and New Horizons

View full text Add to dashboard Cite

“…A wide variety of signal processing applications have been developed using Ptolemy, Ptolemy is capable of automatically partitioning, scheduling, and generating DSP assembly language for a multi-processor DSP system [29] [30]. Currently, it is required that the problem fit the synchronous dataflow model; a future version will exploit heterogeneity to remove that restriction by mixing Schedulers that operate at code-generation time with Schedulers that synthesize run-time code [9].…”

Section: Signal Processingmentioning

confidence: 99%

Ptolemy: A Framework for Simulating and Prototyping Heterogeneous Systems

Buck

Lee

et al. 2002

Readings in Hardware/Software Co-Design

489

328

View full text Add to dashboard Cite

Ptolemy is an environment for simulation and prototyping of heterogeneous systems. It uses modern object-oriented software technology (C++) to model each subsystem in a natural and efficient manner, and to integrate these subsystems into a whole. Ptolemy encompasses practically all aspects of designing signal processing and communications systems, ranging from algorithms and communication strategies, simulation, hardware and software design, parallel computing, and generating real-time prototypes. To accommodate this breadth, Ptolemy must support a plethora of widely-differing design styles. The core of Ptolemy is a set of object-oriented class definitions that makes few assumptions about the system to be modeled; rather, standard interfaces are provided for generic objects and more specialized, application-specific objects are derived from these. A basic abstraction in Ptolemy is the Domain, which realizes a computational model appropriate for a particular type of subsystem. Current examples of domains include synchronous and dynamic dataflow, discrete-event, and others appropriate for control software and embedded microcontrollers. Domains can be mixed as appropriate to realize an overall system simulation. Some current applications of Ptolemy include networking and transport, call-processing and signaling software, embedded microcontrollers, signal processing (including implementation in realtime real-time), scheduling of parallel digital signal processors, board-level hardware timing simulation, and combinations of these.Key words: Heterogeneity, mixed-mode, simulation, prototyping, object-oriented programming INTRODUCTIONPtolemy is a simulation and rapid prototyping framework for heterogeneous systems. It is ideal for applications in which heterogeneity is a key characteristic. such as:• Design of multimedia networks;• Real-time embedded software;• Hardware/Software codesign;• Control and call-processing in telecommunications networks;• Rapid prototyping of new telecommunications services;• Mixed-mode hardware simulation;• Mapping applications onto heterogeneous multiprocessor systems; and• Mixed signal processing and real-time control.For example, in the design of multimedia networks, we are interested in studying the interaction between transport, compression or compositing signal processing, and control software in voice or video services over cell-relay networks. In telecommunication systems we are interested in studying the interaction between call-processing software and hardware switch elements. To develop new telecommunication services, we must jointly design control software, signal processing, transport, and hardware elements.In hardware design we are interested in modeling components with varying detail, such as behavioral, logic, timing, and circuit. In system-level design, we may wish to jointly design the communications infrastructure and the processing elements. We may also wish to synthesize (in combination) microcode for specialized processors, C code for generic processors, routi...

show abstract

Declustering: a new multiprocessor scheduling technique

Cited by 60 publications

References 10 publications

Optimizing computing costs using divisible load analysis

Optimizing computing costs using divisible load analysis

w-TG: A Combined Algorithm to Optimize the Runtime of the Grid-Based Workflow Within an SLA Context

Ptolemy: A Framework for Simulating and Prototyping Heterogeneous Systems

Contact Info

Product

Resources

About