Announcement

Collapse
No announcement yet.

Apple Announces Its New M2 Processor

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

  • Originally posted by smitty3268 View Post
    They've got co-processors running an entire custom RTOS, sending XML data around, etc.
    Honest question: where did you get this info? It's the first time I read of it. But that said, the real question is why they chose XML for that. Or for launchd configs for what matters.

    Originally posted by smitty3268 View Post
    They're referring to something Hector Martin (one of the main Asahi linux devs) hyped up quite a bit. Probably because he's tired of people constantly telling him Apple is just going to screw him over at some point and wanted to put out a counter-narrative, but still. He's massively overhyping a minor issue. Some low-level engineer just spent 30 minutes one day making a fairly obvious change, which may indeed have been for Asahi's benefit, but isn't a sign of any major level of support. I guarantee nobody in a suit at Apple knew anything about it one way or another.
    Eh, it's enough that they gave the engineer the free will to do that IMO. Most corporate settings don't.

    Comment


    • Originally posted by Anux View Post
      You are pulling stuff outa your ass, thats highly unlikely. I mean Nvidia opensourced their driver so I guess nothing is impossible anymore ...
      man just check the forum i did write about Nvidia will opensource their drivers 5-8 years before they did it.

      back in the time people said something like this: "You are pulling stuff outa your ass"
      but in the end they did it and it is not happening.

      "You are pulling stuff outa your ass, thats highly unlikely."

      so maybe true or maybe not true but people read what i write and think about it and in the end logic wins.

      with logic wins means: they do it if it makes sense for them. and most of the time what i say make sense.

      so it is maybe true that i write about nvidia will opensource their driver long long time before nvidia really did it.
      but thats not the point the point is in the end they did it.

      believe it or not but apple with M1/M2 is allied with linux agaist the microsoft/Intel-ISA monopole.
      and apple needs linux to make the M1/M2 a success. and success means if a company like Valve sell the next steam deck with M1/M2 SOC hardware for example to max out the battery play time.

      why they do need linux to make their hardware a success it is very simple compared to the linux market they are "small" and apple alone is to small in marketshare to really hurt microsoft+intel.

      but together apple+linux can break the Wintel monopole.

      lets face it the linux community is good in software but if you see projects like libreSOC the linux community is not a Forcefull movement in the hardware market.

      Lets face it the linux people will never get a steam deck with good battery time with AMD hardware.
      on Desktop or server this has zero relevance but on a mobile device this does matter.

      steam deck users report they have 3-4 hours battery time at the same time apple M2 users report they have 24hours of battery time...

      steam deck is the biggest success for linux in the end-users market the year of the linux desktop is near...

      but on battery time apple M2 is better.

      Phantom circuit Sequence Reducer Dyslexia

      Comment


      • Originally posted by qarium View Post
        man just check the forum i did write about Nvidia will opensource their drivers 5-8 years before they did it.

        back in the time people said something like this: "You are pulling stuff outa your ass"
        but in the end they did it and it is not happening.

        "You are pulling stuff outa your ass, thats highly unlikely."

        so maybe true or maybe not true but people read what i write and think about it and in the end logic wins.

        with logic wins means: they do it if it makes sense for them. and most of the time what i say make sense.

        so it is maybe true that i write about nvidia will opensource their driver long long time before nvidia really did it.
        but thats not the point the point is in the end they did it.

        believe it or not but apple with M1/M2 is allied with linux agaist the microsoft/Intel-ISA monopole.
        and apple needs linux to make the M1/M2 a success. and success means if a company like Valve sell the next steam deck with M1/M2 SOC hardware for example to max out the battery play time.

        why they do need linux to make their hardware a success it is very simple compared to the linux market they are "small" and apple alone is to small in marketshare to really hurt microsoft+intel.

        but together apple+linux can break the Wintel monopole.

        lets face it the linux community is good in software but if you see projects like libreSOC the linux community is not a Forcefull movement in the hardware market.

        Lets face it the linux people will never get a steam deck with good battery time with AMD hardware.
        on Desktop or server this has zero relevance but on a mobile device this does matter.

        steam deck users report they have 3-4 hours battery time at the same time apple M2 users report they have 24hours of battery time...

        steam deck is the biggest success for linux in the end-users market the year of the linux desktop is near...

        but on battery time apple M2 is better.
        I'm not sure they aim for the same market segments honestly. And, as it is, Apple makes much more revenue than Intel and Microsoft. Not as much, but in the same ballpark, as both combined, actually (a quick Google can confirm this). Besides, a part of the performance of their chips comes from being the sole customer of the smallest processes. I'm not sure there are enough fabs of that level to produce in similar volumes to what Intel/AMD sell in chips.
        Not only that, but if they get that to be the norm they'd lose the competitive advantage that gives them the ability to sell at a premium. They have their closed ecosystem and that's the model that works for them.
        Of course, everything is possible, and your POV is pretty much as valid and as speculative as mine, but I'm not sure they have a use for bare metal Linux on their chips. They don't seem to oppose, but I don't see them benefited enough to actively help it. Selling computers is actually more profitable than selling chips in the end, it's horizontal integration.

        Comment


        • Originally posted by sinepgib View Post
          I'm not sure they aim for the same market segments honestly. And, as it is, Apple makes much more revenue than Intel and Microsoft. Not as much, but in the same ballpark, as both combined, actually (a quick Google can confirm this). Besides, a part of the performance of their chips comes from being the sole customer of the smallest processes. I'm not sure there are enough fabs of that level to produce in similar volumes to what Intel/AMD sell in chips.
          Not only that, but if they get that to be the norm they'd lose the competitive advantage that gives them the ability to sell at a premium. They have their closed ecosystem and that's the model that works for them.
          Of course, everything is possible, and your POV is pretty much as valid and as speculative as mine, but I'm not sure they have a use for bare metal Linux on their chips. They don't seem to oppose, but I don't see them benefited enough to actively help it. Selling computers is actually more profitable than selling chips in the end, it's horizontal integration.
          "competitive advantage that gives them the ability to sell at a premium"

          what premium ? last time i checked price you can go buy a M2 for 1200€ and a similar AMD (GPU+CPU) based laptop also costs you 1200€ and you can buy the M1 at 999€... of course again similar AMD notebooks you can also buy for that.

          the price is not the point you can say the AMD notebook allows you to add 64gb ram and the apple m2 is only 16gb/24gb ram... but on the other side if you care for battery time apple wins there.

          but dude back in 1994 my family bought a Apple Macintosh performa 5200 at a real premium price for very low performance... https://www.pinterest.de/pin/486459197227619207/

          but in 2022 apple sells at similar price point.. i can not see any "Premium"

          "hey have their closed ecosystem and that's the model that works for them."

          this is true for apple iphone ... but M1/M2 allow linux to install so it is not as closed ecosystem as you think.

          "but I'm not sure they have a use for bare metal Linux on their chips."

          they have it in the: server and workstation and embedded market.

          Phantom circuit Sequence Reducer Dyslexia

          Comment


          • Originally posted by sinepgib View Post

            Honest question: where did you get this info? It's the first time I read of it. But that said, the real question is why they chose XML for that. Or for launchd configs for what matters.
            Ok, honest answer. It's possible i'm mis-remembering some of that and I don't feel like digging around for the details. So take that for what it's worth.

            But I'm specifically referring to the DCP (display coprocessor) in the M1. It's responsible for doing all the display stuff, and AFAIK the only way to communicate with it is to send commands in XML. Which is kind of painful, for a device that needs to be managed from the linux kernel. I'm pretty sure I heard (and it would have been from Hector Martin) that it was running an entire mini-OS on that co-processor, and it was something that linux wouldn't be able to touch from the M1 CPU itself. I guess it would just be loaded by the firmware - which is 7MB. But I guess I don't know all the details of how all that works, so maybe it's less extreme than I got the impression of. The firmware is 7MB though, so it's doing a lot of something in all that code.

            Comment


            • Originally posted by smitty3268 View Post
              Ok, honest answer. It's possible i'm mis-remembering some of that and I don't feel like digging around for the details. So take that for what it's worth.

              But I'm specifically referring to the DCP (display coprocessor) in the M1. It's responsible for doing all the display stuff, and AFAIK the only way to communicate with it is to send commands in XML. Which is kind of painful, for a device that needs to be managed from the linux kernel. I'm pretty sure I heard (and it would have been from Hector Martin) that it was running an entire mini-OS on that co-processor, and it was something that linux wouldn't be able to touch from the M1 CPU itself. I guess it would just be loaded by the firmware - which is 7MB. But I guess I don't know all the details of how all that works, so maybe it's less extreme than I got the impression of. The firmware is 7MB though, so it's doing a lot of something in all that code.
              I'm not too surprised about the mini OS, it seems to be quite common in the ARM graphics space for some reason. It's more the choice of XML for the protocols I find odd (and somewhat disgusting really).

              Comment

              Working...
              X