Dynamic Memory Management for Embedded Systems

A memory footprint optimization framework for Python applications targeting edge devices

“…Authors of [13] and [14] propose the Dynamic Data Type Refinement (DDTR) methodology, which enables the systematic customization and refinement of dynamic data structures for embedded applications, in terms of performance, memory footprint, and energy consumption [11]. A complementary methodology is the Dynamic Memory Management (DMM), which proposes customized dynamic memory allocators, to meet the application requirements and embedded system constraints, in terms of performance and memory footprint [15]. However, these approaches have been developed for C/C++ embedded applications, while Python and other modern high-level languages are not supported.…”