Applying autonomic principles for workload management in multi-core systems on chip

This paper explores various possibilities for autonomic enhancements to a multi-core network processor system on chip. Based on the autonomic system on chip paradigm, it is shown how monitors can be added to quantify the operating state of a typical processor core, whereupon a learning classifier system evaluator can determine appropriate actions to be performed in order to optimize the frequency and task distribution across the system. A hardware prototype is used to demonstrate the feasibility of autonomic concepts for dynamic component parameterization and task management at run time.

[1]  Walter Stechele,et al.  AutoVision - Reconfigurable Hardware Acceleration for Video-Based Driver Assistance , 2010, Dynamically Reconfigurable Systems.

[2]  Wolfgang Rosenstiel,et al.  An Architecture for Runtime Evaluation of SoC Reliability , 2006, GI Jahrestagung.

[3]  Stewart W. Wilson.  ZCS: A Zeroth Level Classifier System , 1994, Evolutionary Computation.

[4]  Walter Stechele,et al.  Learning Classifier Tables for Autonomic Systems on Chip , 2008, GI Jahrestagung.

[5]  Walter Stechele,et al.  Concept and Design of an SNR-adaptive DRM+/FM Receiver using Dynamic Partial Reconfiguration (DPR) of FPGAs , 2010 .

[6]  Jörg Henkel,et al.  Closing the SoC Design Gap , 2003, Computer.

[7]  Christian Haubelt,et al.  Dynamic task binding for hardware/software reconfigurable networks , 2006, SBCCI '06.

[8]  Michael F. P. O'Boyle,et al.  Mapping parallelism to multi-cores: a machine learning based approach , 2009, PPoPP '09.

[9]  Wolfgang Rosenstiel,et al.  Organic Computing at the System on Chip Level , 2006, 2006 IFIP International Conference on Very Large Scale Integration.

[10]  Sally A. McKee,et al.  Machine learning based online performance prediction for runtime parallelization and task scheduling , 2009, 2009 IEEE International Symposium on Performance Analysis of Systems and Software.

[11]  Wolfgang Rosenstiel,et al.  Combining Software and Hardware LCS for Lightweight On-chip Learning , 2011, Organic Computing.

[12]  Stewart W. Wilson.  Classifier Fitness Based on Accuracy , 1995, Evolutionary Computation.