event-icon
Description
This session presents latest research that explores the use of approximate algorithms in embedded computing to solve issues of constrained resources. In particular, these papers (1) propose solutions to optimize power consumption in many-core processors and their effects on quality of service; (2) explore the tradeoffs of approximate kernels in error resilient applications in heterogeneous multicores; (3) introduce a tree-based classification algorithm as an approximation to replace heavy computation; and (4) use imprecise computation to improve task schedulability.
Sub-Sessions
Tags