Constant Complexity Management of 2D HW Multitasking in Run-Time Reconfigurable FPGAs
This paper presents a constant complexity and fast algorithm for the management of run-time reconfigurable resources by an operating system with extended hardware multitasking functionality. Our algorithm manages a two dimensional reconfigurable device by dividing the resource area into four partitions with different sizes. Each partition has an associated queue where the hardware manager places each arriving task depending on its size, shape, deadline requirements and the state of queues. It is possible to merge partitions for tasks not fitting any partition. The sizes of the partitions may be adapted to different circumstances.
Unable to display preview. Download preview PDF.
- 1.Compton, K., Hauck, S.: Reconfigurable Computing: A Survey of Systems and Software. ACM Computing Surveys, 171–210 (2002)Google Scholar
- 2.Budiu, M.: Application-Specific Hardware: Computing Without CPUs. In: Fourth CMU Symposium on Computer Systems, pp. 1–10 (2001)Google Scholar
- 3.Diessel, O.: Opportunities for Operating Systems Research in Reconfigurable Computing. Technical report ACRC-99-018 (1999)Google Scholar
- 4.Diessel, O.: On Dynamic Task Scheduling for FPGA-based Systems. International Journal of Foundations of Computer Science, IJFCS 2001, 645–669 (2001)Google Scholar
- 5.Handa, M., Vemuri, R.: An efficient algorithm for finding empty space for online FPGA placement. DAC, 960–965 (2004)Google Scholar
- 7.Walder, H., Platzner, M.: Online scheduling for block-partitioned reconfigurable devices. In: DATE 2003, pp. 10290–10295 (2003)Google Scholar
- 8.Román, S., Septién, J., Mecha, H., Mozos, D., Tabero, J.: Partition-based algorithm for efficient 2D HW multitasking. In: EUROMICRO WIP, pp. 26–27 (2003)Google Scholar