Task clustering approach to optimize the scheduling on a partially dynamically reconfigurable FPGAs for image processing algorithms

E.M. Abdali (1) , Maxime Pelcat (2) , F. Berry (1) , J.-P. Diguet (3) , D. Heller (3)

1 LASMEA - Laboratoire des sciences et matériaux pour l'électronique et d'automatique
2 IP - Institut Pascal
3 Lab-STICC - Laboratoire des sciences et techniques de l'information, de la communication et de la connaissance

Field-programmable gate array (FPGAs) are classified as high efficient computational execution platform. However their limited density makes them not suitable for highly demanding algorithms. The Partial Dynamic Reconfiguration (PDR) concept overcomes this problem by rising the FPGA reuse from one task to another. Nonetheless, the scheduling on partially dynamically reconfigurable architectures involves several degrees of freedom and hardware constraints to be managed, which makes the PDR more challenging. In this paper we propose a task clustering approach to optimize the scheduling on PDR FPGAs. By clustering tasks, the approach moves the optimizing overhead from hardware side into the software side, which is largely simple. © 2016 ACM.

Keywords: Cameras; Computer hardware; Degrees of freedom (mechanics); Dynamic models; Field programmable gate arrays (FPGA); Hardware; Image processing; Reconfigurable architectures; Scheduling; Tracking (position); Dynamically reconfigurable architecture; Execution platforms; Hardware constraints; High efficient; Image processing algorithm; Partial dynamic reconfiguration; Reconfigurable; Task clustering; Reconfigurable hardware

Domains: Engineering Sciences [physics]; Electronics; Computer Science [cs]; Networking and Internet Architecture [cs.NI]

E.M. Abdali, Maxime Pelcat, F. Berry, J.-P. Diguet, D. Heller. Task clustering approach to optimize the scheduling on a partially dynamically reconfigurable FPGAs for image processing algorithms. 10th International Conference on Distributed Smart Cameras, ICDSC 2016, Sep 2016, Paris, France. pp.230--231, ⟨10.1145/2967413.2974042⟩. ⟨hal-01380635⟩