Institutional Repository
Technical University of Crete
Technical University of Crete
EN
|
EL
| URI: | http://purl.tuc.gr/dl/dias/95C0A557-C048-4507-85B2-2E158B000F81 | ||
| Year | 2006 | ||
| Type of Item | Conference Full Paper | ||
| License |
|
||
| Bibliographic Citation | K. Papademetriou and A. Dollas, "A task graph approach for efficient exploitation of reconfiguration in dynamically reconfigurable systems," in IEEE International Symposium on Field-Programmable Custom Computing Machines, 2006, pp. 307-308. | ||
| Appears in Collections |
Partial reconfiguration suffers from the inherent high latencyand low throughput which is more considerable whenreconfiguration is performed on-demand. This work dealswith this overhead in processors combining a fixed processingunit (FPU), and a reconfigurable processing unit (RPU).Static and dynamic prefetching [1], and instruction forecasting[2] are targeting at reduction of the overhead throughpreloading of configurations. Banerjee et al. [3] transformthe task graph of an application and a heuristic algorithmevaluates the reduction in schedule length and selects themost promising configuration. Tasks are scheduled accordingto the physical resource constraints. In our work weaugment the prefetching model of [1] by taking into accountthe hardware area constraints of a partially reconfigurablesystem. Given the task graph of an application, tasks withlow probability to be executed are split and preloaded accordingto the hardware in order to be fully utilized. Thus,the time during which reconfiguration is overlapped withprocessor execution is increased.
Copyright © DIAS 2013
