The Danger of Using Patents

Summary judgment uses recent Supreme Court ruling in fight between Mentor and Synopsys over emulation space.

popularity

As I have written about recently, emulation is a hot topic for EDA and the number and length of lawsuits related to the technology is almost overwhelming. The latest phase has just concluded with a summary judgment against Synopsys on Jan. 20. It all started in late 2012 when Synopsys, which had just acquired EVE (Sept. 27, 2012), filed a patent infringement suit against Mentor Graphics. The patents in question are old Synopsys patents related to logic synthesis, something most certainly required to put a design into an emulator, which can only deal with the equivalent of gates. These are the same patents that protected Synopsys from encroaching synthesis companies. I am not aware of them ever having filed suit against anyone, but fear would have been enough to keep most people away.

The patents in question  originally were claimed by Brent Gregory, an early Synopsys employee. The Gregory patents claim a way of performing synthesis, described as “[a] method and system . . . for generating a logic network using a hardware independent description means.” The Gregory patents describe a method for synthesizing a complex logic circuit from a “user description specifying only signals and the circumstances under which the signals are produced, i.e., without requiring the designer to specify the hardware components or connections needed to implement them.”

Now, for a patent to be valid (I am not a lawyer, so don’t rely on anything I say) it has to be non-obvious. This was the first problem for the patents. People had been designing using the same kinds of methods using paper and pencil for a long time. In the past, when I was a logic designer, I used truth tables, Karnaugh maps and many other techniques to take an abstract design down to an implementation. This means that in order for the claim to be patentable, Synopsys must have added something inventive and non-obvious that engineers had not been doing in the past.

In a deposition, Gregory said, “All of [the claims’] concepts and ideas are what Russ [Russell B. Segal] and I came up with in order to automate what the humans were doing to convert it into such a method that a computer could run.”

The court decided that not following exactly the same thought process does not change the thought process.

The second issue, and one I wrote about recently, was the result of Alice vs. CLS Bank and a ruling about software patents by the U.S. Supreme Court. If something is not patentable on its own, then doing it with a computer does not make it patentable. Thus software implementing an abstract process is not patentable.

The court said, “The claimed methods here at issue do not entail anything physical. Rather, the asserted claims are directed to the process of inference, which is fundamental to IC design and can be performed mentally.”

The fact that Synopsys defined “concrete steps in a computerized process for creating a netlist of hardware elements” was not enough and thus could be invalidated. Furthermore, adding that those steps are done by a computer is a pure software artifact which has recently been deemed to be non-patentable.

The result is that some of the fundamental patents that gave Synopsys the dominance to become the No. 1 EDA company have been found to be invalid and that competitors could have entered the market. However, few had the money to fight and at that time, the Supreme Court ruling on software patents did not exist. Any lawsuit filed back in those early days may well have led to a very different outcome. What is clear from this is that the Supreme Court ruling does indeed impact EDA.

The case is Synopsys, Inc. v. Mentor Graphics Corporation, case number 4:2012cv06467, in the U.S. District Court for the Northern District of California. The judgment can be located here.

The claims in question are claims 1, 2, 8, and 9 of Patent No. 5,748,488, claim 1 of Patent No. 5,530,841, and claims 32, 35, and 36 of Patent No. 5,680,318.



  • Brian – really interesting post. Could you elaborate further on what SCOTUS is saying? Are they saying any patent that merely describes an automated method for an unpatentable manual process is not patentable? Or is it going further than that?

    Is it saying I can’t patent a computer-mechanism for synthesizing an OR gate from NAND gates – which on it’s own would not be patentable if I described it as a manual process.

    Lots to digest related to EDA patents with this ruling.

    Warm Regards,

    Sam

  • Brian Bailey

    Hi Sam – I did write about that in a little more detail in this article: http://semiengineering.com/programming-is-not-patentable/

    You have the general idea. If an algorithm exists that would enable me to synthesize an OR gate from a NAND gate then that is not patentable because that is an abstract notion. But implementing it on a computer is not patentable either. The program can be copyrighted, but not patented. The issue at question here would appear to be if the techniques that Synopsys attempted to patent are different than a manual process and that is way beyond my knowledge level of the details of the case.