Abstract
While energy efficiency is essential to extend the battery life of embedded devices, performance cannot be ignored. High performance superscalar embedded processors are more energy efficient than low performance scalar processors, however, they consume more power which is very limited in battery o
Highlights
The challenge in designing energy efficient embedded devices is how to increase the device computational performance while using the least amount of power
The average increase in IPC for a dual issue processor is around 40%
The processor delivers an average 40% higher IPC than a single issue processor while consuming only 20% more power
Summary
The challenge in designing energy efficient embedded devices is how to increase the device computational performance while using the least amount of power. The Cortex-M4 scalar processor consumes much lower power than the Cortex-A7, it can only issue a single instruction per cycle which limits its performance potential. The reason superscalar processors consume more power than scalar processors is because they use aggressive and complex hardware techniques, such as out-of-order issue, speculative execution, etc., to issue more than one instruction per cycle. The IPC performance of these benchmarks is well above average; it is close to the max issue capacity of a dual issue processor These results confirm that there is an abundance of ILP in embedded programs that can be extracted using minimal effort. We propose a dual issue embedded processor that uses simple hardware techniques to extract available ILP from applications.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
