Instruction vulnerability prediction method and system based on deep random forest

Abstract

The invention discloses a method and system for predicting instruction vulnerability based on deep random forest. The method includes: extracting instruction feature information related to each program instruction and instruction vulnerability, and generating instruction feature vectors representing instruction vulnerability; training program Perform fault injection to obtain the vulnerability value of each program instruction; combine the instruction feature vector and instruction vulnerability value to generate an instruction vulnerability sample data set; perform sliding sampling on the instruction vulnerability sample data set through a sliding window to generate an extended sample data set ; Construct and train an instruction vulnerability prediction model based on deep random forest; extract the instruction feature vector of the target program to be predicted, and combine the instruction vulnerability prediction model to realize the instruction vulnerability prediction of the target program to be predicted. The system is used to implement the above method. The invention has high prediction accuracy, low demand for sample sets and less manual adjustment work, and can be effectively applied to the prediction of instruction vulnerability after a program is affected by a transient fault.

Description

technical field [0001] The invention belongs to the field of software reinforcement and software reliability, and particularly relates to a method and system for predicting instruction vulnerability based on a deep random forest. Background technique [0002] With the rapid development of semiconductor manufacturing processes, the size of computer chips has continued to shrink, resulting in a significant increase in their sensitivity to space radiation. In the space radiation environment, one of the main reasons for the failure of computer systems is the single event flipping effect caused by the irradiation of high-tech integrated circuit chips by high-energy particles or electromagnetic interference. Single event upset (Single Event Upset, SEU) refers to a phenomenon in which a certain bit of a memory value is affected and the logic state is reversed, and this kind of phenomenon is usually called a soft error. [0003] Soft errors are usually divided into the following ca...

AI Technical Summary

Problems solved by technology

In recent years, researches have used methods such as support vector machines and neural networks to predict program fragile instructions, but such methods require a large number of prediction data sets and complex manual parameter adjustments in order to achieve high accuracy.

Images