Reducing RAM footprint of the generated tests for AUTOSAR RTE

With the advancement of the automotive industry, a need for developing and maintaining in-car embedded software has risen. In-car embedded systems are limited with the number of available resources including random access memory. In this paper, the goal was to reduce the RAM consumption of automatically generated tests for the embedded computer systems based on the Infineon Tricore AURIX microcontroller. The program that was optimized was written in C, and the code itself was generated using the proprietary Python tools. There are various techniques used for reducing the RAM footprint. In this paper we used loop inversion, loop unrolling, data type size reduction, memory alignment and heuristic methods. Also, compiler optimization options were used but not tested on the board. We achieved up to 10.97% lower RAM footprint with the optimized program code successfully passing all the initial tests.