Please see all COVID-19 updates here as some shipments may be delayed due to CDC safety and staffing guidelines. If you have an order or shipping question please refer to our Customer Support page. For technical questions please check out our Forums. Thank you for your continued support.

Member #1687046

Member Since: July 2, 2021

Country: United States

  • Usual disclaimers, I am not a lawyer, etc, but I do deal with patents occasionally for work. This is hugely simplified, of course.

    As this is a post-1995 US patent, it has a term of 20 years, which appears to be expired. Depending on age and jurisdiction, patents can last up to 28 years, so it's always worth checking.

    Anyway, you can skip 92 of the 93 pages on this one. To determine infringement, the ONLY thing that matters are the claims (on the last page here, but can be elsewhere). This one only has one claim, so it's easy. Your product needs to do EVERYTHING in that claim to infringe. If there are multiple claims, they'll be arranged in a hierarchy ("the apparatus according to claim X, in which..."), and you need to do everything in one of the independent claims (root nodes, in CS terms) to infringe. Dependent claims are important in the application process, but not too significant once the patent is granted.

    While all the stuff about malloc is amusing, they aren't actually trying to patent that. The detailed description (and this is an impressively long one) counts as prior art if you're trying to invalidate a later patent or apply for your own, but is irrelevant for determining infringement. It can be used to clarify terms used in the claims, usually when defending against it. Prior art can invalidate a patent if it describes an entire independent claim, but prior art for something in the description doesn't matter.

    Anyway, the actual claim here is for a multiprocessor system with some kind of shared memory, with certain specific modules of each processor able to run concurrently. It might apply to the system you used if it has more than one core ("plurality" in patentese is explicitly more than one), but probably not. If they haven't sued at least Intel, AMD and nVidia, and won or reached a big settlement, it's probably too narrow to apply to any processor that wasn't designed by Cognigine (the original owners, who do appear to be a real company with products) themselves. The architecture they describe looks a lot more like a modern graphics card than a CPU or microcontroller, but given Cognigine's business, it was probably designed for networking.

    As for whether it was ever valid? 1998 was way before modern multi-core CPUs, and about the time early consumer 3D graphics cards came out, so it was probably novel and non-obvious back then.

    Incidentally, the fact that there is only one claim is also a hint that this is a really narrow patent. Patents are often drafted with a big hierarchy of claims, mostly non-novel, in the hope of getting broad coverage. As part of the application process, claims get combined until each independent claim is novel in itself (this can take years). If the entire tree got collapsed into one claim, then there's probably only one tiny detail in that claim that is actually novel. Or rather, was actually novel in 1998.

No public wish lists :(