| 基于伪随机的微处理器验证方法及改进 |
| |
| 引用本文： | 张振军,毛志刚.基于伪随机的微处理器验证方法及改进[J].信息技术,2008,32(10). |
| |
| 作者姓名： | 张振军 毛志刚 |
| |
| 作者单位： | 上海交通大学微电子学院,上海,200240 |
| |
| 摘 要： | 处理器验证是验证工作中最重要也是时间开销最大的部分.文中提出了一种基于仿真的微处理器验证方法,以伪随机算法为主自动生成测试向量,实现了验证的自动化和验证环境的可重用性.并在此基础上引入神经网络算法,提高了整个验证过程的自动化程度和效率.通过一款DSP处理器的验证结果表明,该方法确实提高了微处理器功能验证的有效性和完备性.
|
| 关 键 词： | 指令集验证 伪随机 覆盖率 人工神经网络 |
Pseudo-random functional verification method of microprocessor and its improvement |
| |
| ZHANG Zhen-jun,MAO Zhi-gang.Pseudo-random functional verification method of microprocessor and its improvement[J].Information Technology,2008,32(10). |
| |
| Authors: | ZHANG Zhen-jun MAO Zhi-gang |
| |
| Affiliation: | ZHANG Zhen-jun,MAO Zhi-gang(School of Microelectronics,Shanghai Jiaotong University,Shanghai 200240,China) |
| |
| Abstract: | Verification of microprocessor is the most important part of IC verification and becomes the bottleneck.In this paper,pseudo-random generation is used as the main method to generate test benches,and the verification environment is highly automatic and reusable.In addition,neural network is introduced to improve the verification efficiency and automation.By the verification results of a 16-bit fixed point DSP the validity and the completeness of this method has been proven. |
| |
| Keywords: | instruction set verification pseudo-random coverage ratio artificial neural network |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
