马上注册,结交更多好友,享用更多功能,让你轻松玩转社区。
您需要 登录 才可以下载或查看,没有账号?注册
x
I'm glad that nowadays SystemVerilog and UVM are most popular things engineers are concerned. But does everyone really think things clearly before adopt them? I think maybe not.
Generally speaking, there is not the best ones. All the verification methodologies are developed by simulator suppliers and verification languages and are commercial. They just compete with each other to earn more marketplaces. But the fierce compete leads to progress of the whole EDA industry, there do have step a big forward. Say, it much like weapons, are you sure the more complicated ones can beat all enemies?the answer maybe no. The more complicated the more difficult to master, you have to spend much time to learn the language and methodology, also you must gain enough experience before you can use it powerfully. In my opinion, one should choose the easiest way to handle problems, the easiest is the best, not the most complicated ones. If your project is small like AVS decoder, you may not need to apply the more advanced methodologies, verilog is just enough to deal all things, maybe some other verification languages also will do,such as E,Vera.If your project really very big, such as smart phones SOC etc. verilog is inefficient to take charge in, then, you may think of more advanced methodologies, such as OVM, VMM, UVM. they will make verification easier. the other choice when one face to SOC era, one may use the powerful SystemC as a system model and verification tools. the more abstract the more one can improve their productivity and efficiency.
Welcome everyone to participate in discussing this topic. |