An Adaptive Dataflow Scheduler for an N-Processor Architecture
Abstract
The problem of scheduling software tasks for execution in a Multiple Instruction, Multiple Data (MIMD) computer system architecture, such as that used in aircraft flight simulation, is addressed. Of the standard scheduling approaches discussed, the method employing a combination of synchronous (repetitive) and asynchronous (event-driven) techniques is the most widely used, but requires the system designer to manually determine the execution order of the tasks. When this manual process is applied to a system containing several processors (as found in an MIMD architecture) and hundreds of tasks (as found in aircraft simulation) the system designer is faced with a large number of possible combinations of processors and tasks. This makes it difficult to effectively arrange the task execution sequence. As an alternative to the manual scheduling process, a dataflow scheduler is presented. The dataflow scheduler automatically maximizes processor utilization across the MIMD architecture, and ensures that the set of tasks in the system is executed in the correct order given by the flow of data from task to task. The scheduler correctly handles synchronous and asynchronous tasks, giving priority to synchronous tasks, while fairly executing asynchronous tasks. The asynchronous task fairness allows new data to enter the simulation even if the current set of synchronous tasks occupy the entire simulation frame.
Notes
This item is only available in print in the UCF Libraries. If this is your thesis or dissertation, you can help us make it available online for use by researchers around the world by STARS for more information.
Graduation Date
1989
Semester
Spring
Advisor
Linton, Darrell G.
Degree
Master of Science (M.S.)
College
College of Engineering
Department
Computer Engineering
Format
Pages
188 p.
Language
English
Length of Campus-only Access
None
Access Status
Masters Thesis (Open Access)
Identifier
DP0027044
Subjects
Dissertations, Academic -- Engineering; Engineering -- Dissertations, Academic
STARS Citation
Pierce, Reuben Dennis, "An Adaptive Dataflow Scheduler for an N-Processor Architecture" (1989). Retrospective Theses and Dissertations. 4202.
https://stars.library.ucf.edu/rtd/4202
Accessibility Status
Searchable text