An instruction prefetch content selection method for optimizing real-time task wcet

Abstract

The invention provides an instruction prefectching content selecting method for optimizing WCET (worst-case execution time) of a real-time task. A hardware structure of a basic-block information table (BBIT) is added in high-speed caching hardware of an existing instruction. The instruction prefectching content selecting method includes: initializing hardware configuration, collecting a static routine base block set via a compiler or a static routine character analyzing tool, removing base blocks only covering one storage block, collecting execution frequentness of every program base block under worst-case execution condition, sequencing the program base blocks according to the execution frequentness, selecting information of an appointed amount of program base blocks according to the sequence, and storing the information in the BBIT. During a high-speed cache accessing process, high-speed instruction cache and the BBIT are searched at the same time, high-speed instruction cache accessing loss under the worst-case execution condition can be effectively avoided, requirements on the real-time task are met, and meanwhile, the instruction prefectching content selecting method can effectively reduce hardware design cost, and is simple and direct to realize.

Description

technical field [0001] The invention relates to a method for selecting instruction prefetch content, in particular to a method for selecting instruction prefetch content oriented to real-time applications in real-time systems and optimizing worst-case execution time (WCET: Worst-CaseExecutionTime) analysis accuracy, belonging to computer System architecture and real-time embedded systems field. Background technique [0002] Unlike general-purpose computer systems, which mainly focus on performance under average execution conditions, real-time systems pay more attention to system performance and task execution time under worst-case execution conditions. Missing the execution deadline of the task usually leads to the failure of the real-time system, especially the hard real-time system, causing adverse consequences or even catastrophic accidents, such as: car crashes, nuclear leaks, etc. Accurately predicting the execution time of the system in the worst case is of great sign...

AI Technical Summary

Problems solved by technology

Although this type of method may be effective for a certain type of application, it is not universal. Once the type of application changes, the effectiveness of the method cannot be guaranteed.

Images