Analysis of multi-core architecture for automotive applications

Abstract

Transportation is the artery of India's economy and also the back bone of cultural exchange, national unity and integrity. Nearly 60% of the world's population has vehicles. This vital increase in vehicles is due to the rapid development of electronics. In a car nearly 30% of the weight goes to electronics which symbolizes the major role of electronics. These advancements led to the increase in the number of electronic control units (ECUs) in modern vehicles. Higher safety concern and good comfort in turn puts higher computational demand on ECUs, which further increases the design complexity of automotive control systems. Multi-core processors, whose usage in the general purpose machines have now extended its hand to support the additional computation needs of the automotive industry. This paper illustrates the multicore processor based ECU system and states its advantages compared to the current single core based ECU. In this paper, the timing analysis is made by comparing both the systems. The timing analysis is based on the number of instructions capable of running in the processors and time taken to execute a number of tasks. This timing analysis is made using RTaW — ECU and SNIPER. The former is scheduling software and the latter is the multicore based simulator of tasks. Also in this paper, the power consumed in both the systems is manipulated using SNIPER power estimation tool. This proves that the number of ECU's get reduced which in turn reduces the cost and power consumed thus paving way to the modern transportation system.