TY - JOUR T1 - Task Transition Scheduling for Data-Adaptable Systems JF - ACM Transactions on Embedded Computing Systems (TECS) Y1 - 2017 A1 - Nathan Sandoval A1 - Casey Mackin A1 - Sean Whitsitt A1 - Gopinath, Vijay Shankar A1 - Sachidanand Mahadevan A1 - Milakovich, Andrew A1 - Merry, Kyle A1 - Jonathan Sprinkle A1 - Roman Lysecky KW - Data adaptability KW - hardware/software codesign KW - model-based design KW - runtime transition scheduling AB -

Data-adaptable embedded systems operate on a variety of data streams, which requires a large degree of configurability and adaptability to support runtime changes in data stream inputs. Data-adaptable reconfigurable embedded systems, when decomposed into a series of tasks, enable a flexible runtime implementation in which a system can transition the execution of certain tasks between hardware and software while simultaneously continuing to process data during the transition. Efficient runtime scheduling of task transitions is needed to optimize system throughput and latency of the reconfiguration and transition periods. In this article, we provide an overview of a runtime framework enabling the efficient transition of tasks between software and hardware in response to changes in system inputs. We further present and analyze several runtime transition scheduling algorithms and highlight the latency and throughput tradeoffs for two data-adaptable systems. To evaluate the task transition selection algorithms, a case study was performed on an adaptable JPEG2000 implementation as well as three other synchronous dataflow systems characterized by transition latency and communication load.

VL - 16 UR - http://doi.acm.org/10.1145/3047498 ER -