Announcement

Collapse
No announcement yet.

AMDGPU-PRO 17.10 vs. Linux 4.12 + Mesa 17.2-dev Git

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

  • #21
    @Michael, a comparison using GTX 1050ti and GTX 1060 with the same set of tests would be really interesting.

    Comment


    • #22
      Originally posted by Qaridarium
      the best AMD could do to end this Nvidia-Monopoly strategy of broken software by design fixed by the closed source driver is to simple END support for any kind of broken software. This would increase the pressure on the software companies to fix there bugs. and this in the end will break the nvidia strategy of broken software by design fixed by closed source drivers.
      No, this would result in a monopoly, which I'm sure is not what you want.
      Test signature

      Comment


      • #23
        Originally posted by Qaridarium

        AMD(bridgman) do not even understand that there is a Software-war they really think there is still a future for closed source drivers...
        but they misunderstand the market only one company benefit if software is broken by design and fixed by closed source drivers and this is Nvidia. AMD(bridgman) really thinks that there is a market for amd by copying nvidia strategy of broken software by design fixed by closed source drivers. But in the end this kind of customers will always land in the Nvidia land. It is a illusion that amd could even try to compete in this game.

        the best AMD could do to end this Nvidia-Monopoly strategy of broken software by design fixed by the closed source driver is to simple END support for any kind of broken software. This would increase the pressure on the software companies to fix there bugs. and this in the end will break the nvidia strategy of broken software by design fixed by closed source drivers.
        Blah!

        Perhaps AMDGPU-PRO (if that's the spelling) will serve another purpose, of running professional CAD software, 3D software (like Maya, I don't know what others are available) and such. Perhaps this will be in part for users of Radeon Pro (formerly Fire Pro or Fire GL) which might be a small market, but a market that does exist. There will always be "evil" $100,000 software whether we like it or not, because if you spend well over $10 million developing industry-specific crap for which there are a few thousands customers that's how it goes.

        For that matter, all hardware companies are "evil" including AMD, because they all play the patents game ("treasure chest", licensing all over the place). Maybe we should all ignore all patents because they're all barriers to entry and mafia extortion racket ; a handful countries might do that, but then they'd be living under US naval blockade, CIA "color revolutions" and such.
        So, let's advocate for abolishing patents why not, it will a require a few major nations to break free from alliance with the US though (or for the US itself to abolish patents)

        Anyway those "AMDGPU+RadeonSI" results are good and it's nice it's tested on a non-bleeding edge OS (but with bleeding edge kernel and driver)
        Good that the needed hardware gets cheaper too (e.g., Radeon RX 550 + cheapest DP to VGA adapter) although it's still that small tragedy of needing to buy hardware to get a good linux 3D driver.

        Comment


        • #24
          Originally posted by grok View Post
          Perhaps AMDGPU-PRO (if that's the spelling) will serve another purpose, of running professional CAD software, 3D software (like Maya, I don't know what others are available) and such. Perhaps this will be in part for users of Radeon Pro (formerly Fire Pro or Fire GL) which might be a small market, but a market that does exist.
          That was always its primary purpose, although over the years it and fglrx have also been useful for consumer/desktop customers while we finish off the all-open stack.
          Last edited by bridgman; 21 May 2017, 08:40 PM.
          Test signature

          Comment


          • #25
            Congrats on your 666. message after 10K

            Originally posted by bridgman View Post

            That was always its primary purpose, although over the years it and fglrx have also been useful for consumer/desktop customers while we finish off the all-open stack.
            There is no such a thing as all-open other than in marketing, in reality these are mostly-open and mostly-closed

            I understand if marketing said that once that is considered OK, but to keep calling it like that is disgusting
            Last edited by dungeon; 22 May 2017, 02:47 AM.

            Comment


            • #26
              I understand the "mostly closed" (what we call hybrid) but no idea what you mean by "mostly open". Can you elucidate?
              Test signature

              Comment


              • #27
                Originally posted by bridgman View Post
                I understand the "mostly closed" (what we call hybrid) but no idea what you mean by "mostly open". Can you elucidate?
                He's the guy who keeps trying to bring up the firmware flamewars.

                Comment


                • #28
                  Originally posted by bridgman View Post
                  I understand the "mostly closed" (what we call hybrid) but no idea what you mean by "mostly open". Can you elucidate?
                  That was elucidated even when i we first commented on amdgpu plans back in 2014. All-Open was fine to put on slide back then together with another side as a presentation point and to sort of describe or to clearify difference between these... but to use all-open non stop even out of presentation is bullshit

                  Really i have feeling while reading your comments which include these that i reading from sort of bilboard instead of speaking with human being

                  Bridgman, an AMD's billboardman Hybrid is also all-open, as that also consist of various parts which is what hybrid word means - you see, out of presentation you can't clearfy anything with using these names alone.

                  Just use driver names
                  Last edited by dungeon; 23 May 2017, 02:31 AM.

                  Comment


                  • #29
                    Originally posted by dungeon View Post

                    That was elucidated even when i we first commented on amdgpu plans back in 2014. All-Open was fine to put on slide back then together with another side as a presentation point and to sort of describe or to clearify difference between these... but to use all-open non stop even out of presentation is bullshit

                    Really i have feeling while reading your comments which include these that i reading from sort of bilboard instead of speaking with human being

                    Bridgman, an AMD's billboardman Hybrid is also all-open, as that also consist of various parts which is what hybrid word means - you see, out of presentation you can't clearfy anything with using these names alone.

                    Just use driver names
                    We call it the all-open stack internally as well, ie that is the driver name. That's why it appears in both presentations and my responses.

                    What would you suggest we call it instead, and could you provide a bit more detail re: why you are calling it "mostly open" than telling me you answered some time back in 2014 ?
                    Test signature

                    Comment


                    • #30
                      Originally posted by bridgman View Post
                      We call it the all-open stack internally as well, ie that is the driver name. That's why it appears in both presentations and my responses.
                      Well i understand internally could be AO, but your users does not live internally but externally.

                      What would you suggest we call it instead, and could you provide a bit more detail re: why you are calling it "mostly open" than telling me you answered some time back in 2014 ?
                      I call it mostly open becuase you call it all open, as that is bullshit word outside and means nothing to the user. If you didn't noticed i also complain on word hybrid which also describe nothing outside. I already suggested how to call it - just use driver names for the God's sake.

                      You makes driver, driver has name and what else? Do you wanna know how i call mine underwear internally?

                      Is mesa+llvm hybrid already? Of course it is, but you call that all-open You see how that goes . If it is not as simple describable as that it is nothing, without any potentional on raising further unneeded questions. As if you say OK that is hybrid also but that is not all of a all-open stack it is just "UMD part of all-open", then we go further into questions "What is UMD?", "OK, but what are other parts?" and so on Forget about other parts lets explain just one of these, hybryid combined internally/externally on all-open we use one of these as you see as all-open, but internally with hybrid and not on hybrid which we also use externally but even not on all-open but also on non-hybrid... you see how you just can't explain nothing and it goes nowhere

                      "Wait a minute" - user thinks. "But hybrid is entirely something else, what in the hell is that!!!" That is what it is, UMD is AMD which beggins with U instead of A and is mostly-open, marketing hybrid wording under knees you know

                      You had just fglrx and radeon, before drivers wear their names. Suddenly and out of nowhere Popeye with a bit of spinach and likely some mushrooms became hybrid and Olive decided to be all-open... in reality nothing happened, hybrid actually means that guys just join forces internally and else is just usual giant pile of marketing mostly-open bullshit
                      Last edited by dungeon; 23 May 2017, 10:30 PM.

                      Comment

                      Working...
                      X