An Evolutionary Approach to Dynamic Task Scheduling on FPGAs with Restricted Buffer
✍ Scribed by Martin Middendorf; Bernd Scheuermann; Hartmut Schmeck; Hossam ElGindy
- Publisher
- Elsevier Science
- Year
- 2002
- Tongue
- English
- Weight
- 247 KB
- Volume
- 62
- Category
- Article
- ISSN
- 0743-7315
No coin nor oath required. For personal study only.
✦ Synopsis
Dynamically and partially reconfigurable field-programmable gate arrays (FPGAs) allow to swap in and out tasks without interrupting the execution of other tasks. The FPGA controller can decide on-line where to place new tasks onto the FPGA. Rearranging a subset of the tasks executing on the FPGA may allow the next pending task to be processed sooner. When tasks are rearranged, the arriving input data have to be buffered while the execution is suspended. In this paper, we describe and evaluate an evolutionary approach to solve the problem of placing and rearranging tasks that are supplied by input streams which have constant data rates. We use two genetic algorithms (GAs): one for identifying feasible rearrangements and the other for scheduling a selected rearrangement so that the delay caused by this rearrangement is small and the limited input buffer size is respected. # 2002 Elsevier Science (USA)