The ESL Edge

Over the past couple of years, I have sensed a growing dissatisfaction with constrained random stimulus generation techniques. I have written about this in several articles, including “Constrained random test struggles to live up to promises” in SCDSource. While this remains a workhorse for large parts of the verification problem, it is both an inefficient use of computational resources and becoming increasingly ineffective at getting to verification closure. The industry seems to be answering this by trying to create more effective stimulus with what are generally being called intelligent testbenches. While it is not clear that these are gaining lots of traction, there is another horse that has slid past the pack, one that has been in our midst all along. That horse is formal property verification. Formal verification has been getting more and more capable over the past several years, and is not just for the Ph.D crowd anymore. Why do I say that it is winning now? Well I have collected a couple of pieces of unofficial data over the past several days. The first is that one of the major EDA companies is struggling to make money with their simulation/test generation products and is experiencing declines in sales. The second is that EDAC Marketing Statistics Survey reported very strong sales growth in the formal property verification segment. I have also confirmed that Jasper Design Automation is growing much faster than the rest of the companies in that segment (more than doubling their market share in the past 12 months) and has won several keystone accounts. Putting all this together tells us two things:

1. Jasper is winning technologically. This was also made very clear in some of the recent posting on Deep Chip, with John Cooley expressing surprise that a small company could be out-selling the big guys.
2. Formal is winning market share over its biggest competitor – namely simulation based techniques. This is in some ways the more notable of the two conclusions because it puts to rest the sentiment expressed by some that formal is a bit player.

Perhaps the reason why Jasper is pulling ahead of the pack is that all of the other vendors of formal are also tied to companies with simulation products. For them – it is a zero sum gain, and given that simulation has been their bread and butter for so long, they are loathed to spend enough to make formal a real competitor. Why destroy a market in which you have a nice cash cow? Well, it may be that they will have to wake up soon and recognize what is happening in the industry. Formal’s day has come.

(Disclosure: I sit on the technical advisory board of Jasper Design Automation.)