Title
Hardware Scheduling for Dynamic Adaptability using External Profiling and Hardware Threading
Abstract
While performance, area, and power constraints have been the driving force in designing current communication-enabled embedded systems, post-fabrication and run-time adaptability is now required. Two dominant configurable hardware platforms are processors and FPGAs. However, for compute-intensive applications, neither platform delivers the needed performance at the desired low power. The need thus arises for custom, application-specific configurable (ASC) hardware. This paper addresses the optimization of ASC hardware. Our target application areas are multimedia and communication where an incoming packet (task) is processed independently of other packets. We innovatively utilize two concepts: external profiling and hardware threading. We utilize an M/M/c queueing model to profile task arrival patterns and show how profiling guides design decisions. We introduce the novel concept of hardware threading which allows on-the-fly borrowing of unutilized hardware, thus maximizing task-level parallelism, to either boost performance or to lower power consumption. We present a scheduling algorithm that synthesizes a hardware-threaded architecture, and discuss experimental results that illustrate adaptability to different workloads, and performance/power trade-offs.
Year
DOI
Venue
2003
10.1109/ICCAD.2003.79
Proceedings of the 2003 IEEE/ACM international conference on Computer-aided design
Keywords
Field
DocType
field programmable gate arrays,scheduling algorithm,embedded systems,multi threading,embedded system
Multithreading,Scheduling (computing),Computer science,Profiling (computer programming),Real-time computing,Electronic engineering,Queueing theory,Computer hardware,Fair-share scheduling,Hardware compatibility list,Network packet,Field-programmable gate array,Embedded system
Conference
ISBN
Citations 
PageRank 
1-58113-762-1
2
0.36
References 
Authors
9
2
Name
Order
Citations
PageRank
Brian Swahn1444.59
Soha Hassoun2535241.27