Atrenta says that their customers consider their SpyGlass tool to be the definitive answer on whether their RTL code is up to snuff. But that’s largely been on the basis of whether the RTL looks good or follows well-behaved, well-understood styles. It also looks at whether clocks are likely to behave well as they jump domains, whether the power is optimized, whether the thing is testable, whether the timing constraints are right, and whether the routing is too congested.
But they haven’t been able to tell you whether the actual RTL behavior is as you expect; they’ve remained outside the functional verification category. There are two main tools for checking behavior: simulation and the use of assertions. And the latter is reasonably congruent to the kinds of technology that Atrenta has – specifically, formal methods.
But assertions have taken some time to catch on. They’ve been described as hard to write for anything more than simple tests and finicky in that you have to do it just right or else you might end up chasing your tail. This is where assertion synthesis has tried to make life easier.
One company engaging in assertion synthesis is NextOp, and Atrenta recently announced that they’re acquiring the company in order to add functional verification to their portfolio. They claim that this will give them full coverage of front-end design activities.
You can find out more about the acquisition (except for the terms) in their release.