Cache-aware SPM allocation algorithms for hybrid SPM-cache architectures

In this paper, we explore 6 SPM allocation algorithms to improve performance or energy efficiency for Hybrid SPM-Cache (HSC) architectures, including 4 heuristic-based algorithms and 2 optimal algorithms based on model checking. Our experimental results show that the Greedy Stack Distance based Allocation for Energy (GSDA-E) algorithm can efficiently minimize the energy consumption and achieve near optimal performance results.

[1]  David A. Padua,et al.  Estimating cache misses and locality using stack distances , 2003, ICS '03.

[2]  Mordechai Ben-Ari,et al.  Principles of the spin model checker , 2008 .

[3]  Stephen A. Cook,et al.  The complexity of theorem-proving procedures , 1971, STOC.

[4]  Kristof Beyls,et al.  Reuse Distance as a Metric for Cache Behavior. , 2001 .

[5]  Gerard J. M. Smit,et al.  A mathematical approach towards hardware design , 2010, Dynamically Reconfigurable Architectures.

[6]  Wei Zhang,et al.  Hybrid SPM-cache architectures to achieve high time predictability and performance , 2013, 2013 IEEE 24th International Conference on Application-Specific Systems, Architectures and Processors.

[7]  Jason Cong,et al.  An energy-efficient adaptive hybrid cache , 2011, IEEE/ACM International Symposium on Low Power Electronics and Design.

[8]  Jan Gustafsson,et al.  The Mälardalen WCET Benchmarks: Past, Present And Future , 2010, WCET.

[9]  Alexander G. Dean,et al.  Leveraging both Data Cache and Scratchpad Memory through Synergetic Data Allocation , 2012, 2012 IEEE 18th Real Time and Embedded Technology and Applications Symposium.