Reducing impact of cache miss stalls in embedded systems by extracting guaranteed independent instructions

摘要

Today, embedded processors are expected to be able to run algorithmically complex, memory-intensive applications that were originally designed and coded for general-purpose processors. As such, the impact of memory latencies on the execution time increasingly becomes evident. All the while, it is also expected that embedded processors be power-conscientious as well as of minimal area impact, as they are often used in mobile devices such as wireless smartphones and portable MP3 players. As a result, traditional methods for addressing performance and memory latencies, such as multiple issue, out-of-order execution and large, associative caches, are not aptly suited for the mobile embedded domain due to the significant area and power overhead. This paper explores a novel approach to mitigating execution delays caused by memory latencies that would otherwise not be possible in a regular in-order, single-issue embedded processor without large, power-hungry constructs like a Reorder Buffer (ROB). The concept relies on efficiently leveraging both compile-time and run-time information to safely allow non-data-dependent instructions to continue executing in the event of a memory stall. The simulation results show significant improvement in overall execution throughput of approximately 11%, while having a minimal impact on area overhead and power.