Announcement

Collapse
No announcement yet.

Microsoft Granted A Patent For GPU Video Encoding

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #11
    Well, great! Then in court, when Microsoft will have to prove
    non-obviousness again, insiders like you can testify for the
    defense - nvidia or whoever can possibly get sued.
    However, if what you're saying is true, the patent is junk and
    nobody will ever get sued.
    Many times it is the case that patent office will grant a patent if
    it's not sure, because during the trial the whole patent procedure
    (like proving the first inventor and non-obviousness) has to repeated,
    and all potential mistakes of the patent office wouldn't matter.
    It creates these junk patents, that make the news in non-legal,
    opinionated websites.

    Originally posted by movieman View Post
    I was working on GPUs in 2004 and I can guarantee you that using them for video encoding was pretty damn obvious. If I remember correctly we were even talking with
    a third-party about using GPUs for video encoding before then.

    Comment


    • #12
      This is absolutely amazing.

      A patent on running a program on a processor. :/

      Comment


      • #13
        Originally posted by aznabaal View Post
        Many times it is the case that patent office will grant a patent if
        it's not sure, because during the trial the whole patent procedure
        (like proving the first inventor and non-obviousness) has to repeated,
        and all potential mistakes of the patent office wouldn't matter.
        It creates these junk patents, that make the news in non-legal,
        opinionated websites.
        This is the real problem in the system, because for anyone OUTSIDE the legal profession it creates a nasty quagmire of confusion which can only be resolved by having expensive lawyers to protect you... Which if you aren't a major corporation, you can't afford.

        And you say this doesn't matter?

        Comment


        • #14
          I agree. In general lawyer's services in the US are too expensive and it applies to patent law but also to many other issues
          like medical or labor law suits. It is the characteristic of US legal system not a patent system per se.

          However, in many cases it does not matter.
          There are situations in which small business
          is working on a technology that can violate one of these junk patents.
          If the implementation is inferior to the incumbent's implementation
          then the incumbent will usually not sue. If it's better the incumbent
          will acquire it and incorporate the invention.
          Usually we see lawsuits only in case someone big is infringing the patent.

          Originally posted by RobbieAB View Post
          This is the real problem in the system, because for anyone OUTSIDE the legal profession it creates a nasty quagmire of confusion which can only be resolved by having expensive lawyers to protect you... Which if you aren't a major corporation, you can't afford.

          And you say this doesn't matter?

          Comment


          • #15
            Originally posted by aznabaal View Post
            It is not possible since it does not meet the non-obviousness patentability requirement.
            Then why would this patent be any different? All it is patenting is doing vector processing on a GPU, which, by definition, at the date the application was filed, was pretty obvious (libSH was produced in 2003...which was one of the first GPGPU type frameworks allowing for a lot more than rendering shaders to be developed- and it was only ONE of several around at the time.)

            Comment


            • #16
              I am not sure because I don't know this particular patent.
              But, it must be protecting something more specific than vector operations,
              because I am sure that patent office was aware of existing implementations.
              We would have to dissect the text of the patent the private thread if you'd like to.

              Originally posted by Svartalf View Post
              Then why would this patent be any different? All it is patenting is doing vector processing on a GPU, which, by definition, at the date the application was filed, was pretty obvious (libSH was produced in 2003...which was one of the first GPGPU type frameworks allowing for a lot more than rendering shaders to be developed- and it was only ONE of several around at the time.)

              Comment


              • #17
                I''m not an american, so I couldn't do it.
                But if I were, I should start patenting every crazy idea that I had, so I would probably cover many areas in computing.
                It should work as a protest against software patents.
                As I can see, US is becoming a place where only the monopolists giants can say what can be done and what can't.
                Patenting in software is a way to barrier software competition and you, americans, should fight against it.
                Patents are being confounded with copyrights.
                Everyone should be able to implement whatever the wish, as long as they don't copy.

                Comment


                • #18
                  Originally posted by aznabaal View Post
                  I am not sure because I don't know this particular patent.
                  But, it must be protecting something more specific than vector operations,
                  because I am sure that patent office was aware of existing implementations.
                  You presume too much of the USPTO. Seriously.

                  I have made a patent application to them in the past. The application was rejected on some of the most ludicrous grounds. If I had not ran out of funds, it'd probably would have been granted eventually.

                  The application in question was a hybrid hardware/software solution and I was patenting the specific system configuration so as to not be overbroad and to NOT have an idiot software patent (Though nail down a pretty large range of implementations of a fairly secure SCADA solution...). The examiner HAD to have taken a bong hit before doing the work on my patent as they rejected it outright, claiming that several utterly irrelevant patents to mine that ALSO mentioned RPC as a component in the early claims anticipated my patent. My attorney, after a thorough explanation of what was wrong, had come to the conclusion they needed to lay the crack pipe down.

                  I don't give them any benefit of the doubt on this stuff- they almost never get it right. Never.

                  We would have to dissect the text of the patent the private thread if you'd like to.
                  Indeed one would probably have to dissect the patent. However, based on the title of the application, the odds are pretty good it's an overbroad patent that should have never been given the time of day.

                  However, I intend on doing a dissection of it here shortly to verify that statement as being accurate or not.

                  Comment


                  • #19
                    Was this article pulled?

                    Comment


                    • #20
                      This patent bullshit is not a surprise to me. What else would you all expect from the country where about 80% of all wealth is owned by a few?

                      Comment

                      Working...
                      X