HTR: On-Chip Hardware Task Relocation for Partially Reconfigurable FPGAs

Partial reconfiguration (PR) enables shared FPGA systems to nonintrusively time multiplex hardware tasks in partially reconfigurable regions (PRRs). To fully exploit PR, higher priority tasks should preempt lower priority tasks and preempted tasks should resume execution in any PRR. This preemption/ resumption requires saving/restoring the preempted task's execution context and relocating the task to another PRR, however, prior works only provide partial solutions and impose limitations and/or overheads. We propose on-chip hardware task relocation (HTR) software, which enables a task's execution state to be saved, relocated to, and restored in any PRR with sufficient resources. The HTR software executes on a soft-core processor in the FPGA's static region, and is thus portable across any system/application. Experimental results evaluate HTR execution times, enabling designers to tradeoff task/PRR granularity and HTR execution times based on application requirements.

[1]  Alan D. George,et al.  Bitstream relocation with local clock domains for partially reconfigurable FPGAs , 2009, 2009 Design, Automation & Test in Europe Conference & Exhibition.

[2]  Marco D. Santambrogio,et al.  Internal and External Bitstream Relocation for Partial Dynamic Reconfiguration , 2009, IEEE Transactions on Very Large Scale Integration (VLSI) Systems.

[3]  Marco Platzner,et al.  Field Programmable Logic and Application , 2004, Lecture Notes in Computer Science.

[4]  Philip James-Roxby,et al.  A Self-reconfiguring Platform , 2003, FPL.

[5]  Christian Haubelt,et al.  Efficient hardware checkpointing: concepts, overhead analysis, and implementation , 2007, FPGA '07.

[6]  Ulrich Rückert,et al.  REPLICA: a bitstream manipulation filter for module relocation in partial reconfigurable systems , 2005, 19th IEEE International Parallel and Distributed Processing Symposium.

[7]  Wayne Luk,et al.  Enhancing Relocatability of Partial Bitstreams for Run-Time Reconfiguration , 2007, 15th Annual IEEE Symposium on Field-Programmable Custom Computing Machines (FCCM 2007).

[8]  Reinhard Männer,et al.  Multitasking on FPGA Coprocessors , 2000, FPL.

[9]  B. Harrison.  Las Vegas, Nevada , 2002 .

[10]  Heiko Kalte,et al.  Context saving and restoring for multitasking in reconfigurable systems , 2005, International Conference on Field Programmable Logic and Applications, 2005..

[11]  Heiko Kalte,et al.  REPLICA2Pro: task relocation by bitstream manipulation in virtex-II/Pro FPGAs , 2006, CF '06.

[12]  Hiroaki Takada,et al.  A Novel Mechanism for Effective Hardware Task Preemption in Dynamically Reconfigurable Systems , 2010, 2010 International Conference on Field Programmable Logic and Applications.

[13]  Marco D. Santambrogio,et al.  An Enhanced Relocation Manager to Speedup Core Allocation in FPGA-based Reconfigurable Systems , 2012, 2012 IEEE 26th International Parallel and Distributed Processing Symposium Workshops & PhD Forum.

[14]  Heiko Kalte,et al.  Relocation and Defragmentation for Heterogeneous Reconfigurable Systems , 2006, ERSA.

[15]  Alessandro Forin,et al.  Relocation and Automatic Floor-planning of FPGA Partial Configuration Bit-Streams , 2008 .

[16]  Marek Gorgon,et al.  PixelStreams-based implementation of videodetector , 2007 .

[17]  John W. Lockwood,et al.  PARBIT: A Tool to Transform Bitfiles to Implement Partial Reconfiguration of Field Programmable Gate Arrays (FPGAs) , 2001 .

[18]  Wayne Luk,et al.  Automated placement of reconfigurable regions for relocatable modules , 2010, Proceedings of 2010 IEEE International Symposium on Circuits and Systems.

[19]  Aravind Dasu,et al.  PRR-PRR Dynamic Relocation , 2009, IEEE Computer Architecture Letters.

[20]  Brad L. Hutchings,et al.  Multitasking Hardware on the SLAAC1-V Reconfigurable Computing System , 2002, FPL.

[21]  Camel Tanougast,et al.  A Hardware Preemptive Multitasking Mechanism Based on Scan-path Register Structure for FPGA-based Reconfigurable Systems , 2007, Second NASA/ESA Conference on Adaptive Hardware and Systems (AHS 2007).