{"title":"硬件设计功能验证的高性能机器学习模型","authors":"Khaled A. Ismail, M. A. E. Ghany","doi":"10.1109/NILES53778.2021.9600502","DOIUrl":null,"url":null,"abstract":"Fast and accurate Machine Learning (ML) models for predicting input stimulus in verification testbenches are proposed in this paper. Multiple (ML) models: Artificial Neural Network (ANN), Deep Neural Network (DNN), Support Vector Regression (SVR) and Decision Trees (DT) are examined to constrain randomization of input values to hit the planned coverage metrics. Accuracy of the models is measured using (ML) evaluation metrics such as: Mean Squared Error (MSE) and (R2 score). Training time required for each (ML) model is calculated and compared. Investigated (ML) models show an average improvement of 63.5% in the number of simulation cycles needed to reach full coverage closure compared to existing work. Comparative analysis between the models shows that (DT) is the most suitable (ML) model for a functional verification environment, due to its high accuracy and low training time required.","PeriodicalId":249153,"journal":{"name":"2021 3rd Novel Intelligent and Leading Emerging Sciences Conference (NILES)","volume":"61 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-10-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":"{\"title\":\"High Performance Machine Learning Models for Functional Verification of Hardware Designs\",\"authors\":\"Khaled A. Ismail, M. A. E. Ghany\",\"doi\":\"10.1109/NILES53778.2021.9600502\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Fast and accurate Machine Learning (ML) models for predicting input stimulus in verification testbenches are proposed in this paper. Multiple (ML) models: Artificial Neural Network (ANN), Deep Neural Network (DNN), Support Vector Regression (SVR) and Decision Trees (DT) are examined to constrain randomization of input values to hit the planned coverage metrics. Accuracy of the models is measured using (ML) evaluation metrics such as: Mean Squared Error (MSE) and (R2 score). Training time required for each (ML) model is calculated and compared. Investigated (ML) models show an average improvement of 63.5% in the number of simulation cycles needed to reach full coverage closure compared to existing work. Comparative analysis between the models shows that (DT) is the most suitable (ML) model for a functional verification environment, due to its high accuracy and low training time required.\",\"PeriodicalId\":249153,\"journal\":{\"name\":\"2021 3rd Novel Intelligent and Leading Emerging Sciences Conference (NILES)\",\"volume\":\"61 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2021-10-23\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"1\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2021 3rd Novel Intelligent and Leading Emerging Sciences Conference (NILES)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/NILES53778.2021.9600502\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2021 3rd Novel Intelligent and Leading Emerging Sciences Conference (NILES)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/NILES53778.2021.9600502","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
High Performance Machine Learning Models for Functional Verification of Hardware Designs
Fast and accurate Machine Learning (ML) models for predicting input stimulus in verification testbenches are proposed in this paper. Multiple (ML) models: Artificial Neural Network (ANN), Deep Neural Network (DNN), Support Vector Regression (SVR) and Decision Trees (DT) are examined to constrain randomization of input values to hit the planned coverage metrics. Accuracy of the models is measured using (ML) evaluation metrics such as: Mean Squared Error (MSE) and (R2 score). Training time required for each (ML) model is calculated and compared. Investigated (ML) models show an average improvement of 63.5% in the number of simulation cycles needed to reach full coverage closure compared to existing work. Comparative analysis between the models shows that (DT) is the most suitable (ML) model for a functional verification environment, due to its high accuracy and low training time required.