The following blog entry was written by Jim Foley, R&D Director at Real Intent
Lately I’ve been thinking a lot about RTL lint. This might be helpful to you, because if you’re tasked with solving multi-corner timing closure or characterizing activity for realistic power estimates, you know you don’t have much time left to think about lint. Nor should you have to. Granted, there are very many things that lint can check for – more about this in a bit — and there are as many different opinions about what constructs, conventions, and potential problems ought to be checked for and at which stage in the design process as there are individual design engineers. I tend to put high value on lint rules that report on constructs that are legal but have suspicious modeling, that in Ascent Lint we categorize as “dubious”. There’s also a lot of value to checking coding conventions to make code clear and readable – by whatever criteria meets that objective for you or your design group, as well as catching things that may simulate fine but may cost you time and trouble downstream in the design flow.