The Impact of Different Multi-Threading Interfaces on Embedded Systems

Abstract

The popularity of multicore embedded systems brings new challenges to the development of parallel applications: at the same time it is necessary to exploit the availability of multiple cores, it is also mandatory to consume less energy. To speed up the development process and make it as more transparent as possible to the programmer, parallelism is exploited through the use of Application Programming Interfaces (API). Each one of these API implements different ways to exchange data. However, data exchange occurs between the threads in shared memory regions, which have higher energy consumption. Therefore, each API will present different energy costs to communicate. In this paper, we present the first step to show that different APIs have different impacts on energy consumption, through the analysis of the communication mechanism that each one employs, the number of memory accesses necessary for the communication, and the number of executed instructions according to the API used. Our results show that OpenMP presents a higher communication overhead, with more memory accesses and instructions executed, when compared to Pthreads.