The G5 is an AMD Clawhammer

2

Comments

  • Reply 21 of 46
    razzfazzrazzfazz Posts: 728member
    [quote]Originally posted by FotNS:

    <strong>1. The x86 kernel already exists. It is available in Darwin. GNU Darwin 2.5 even incorporates newer ATA drivers and can be installed on Athlon systems.</strong><hr></blockquote>



    It can be installed on some non-Intel-systems now, but it's still extremely picky.

    It also supports exactly one ethernet chipset, and there seem to be quite some problems getting more than a blank screen from many recent graphics cards (Radeon 8500, GF4) too.

    All in all, Darwin/x86 is still far from being on par with the PPC version, and even further from being a mature, production-quality OS.





    [quote]<strong>

    All that is needed is Quartz and then Aqua for a full desktop environment.

    </strong><hr></blockquote>



    Some people might also want to have printing and sound architectures, the Carbon and Cocoa frameworks, and then some.

    <img src="graemlins/smokin.gif" border="0" alt="[Chilling]" />



    [quote]<strong>

    2. Apple would have the full source code of Classic (obviously), and it would be relatively straightforward for them to make a classic environment in x86 that would run at reasonable speed, ie at least half the megahertz of the x86 for PPC code.

    </strong><hr></blockquote>



    Easy?

    Can you tell me why, in your opinion, all the x86 Mac emulators only emulate a 68k processor, then?

    :confused:



    Bye,

    RazzFazz



    [ 04-07-2002: Message edited by: RazzFazz ]</p>
  • Reply 22 of 46
    I agree with Programmer and RazzFazz that porting OS X to x86 (as opposed to porting Darwin, which is just a kind of minimal core) is a HUGE job. They won't do it unless they have no better option.



    But they do have a better option: design a PPC decoder stage to tack onto the Hammer core. No software ports required, everything runs fully native.
  • Reply 23 of 46
    airslufairsluf Posts: 1,861member
  • Reply 23 of 46
    timortistimortis Posts: 149member
    [quote]Originally posted by Junkyard Dawg:

    <strong>



    Oh, and this is a report from the where? The Register? The same source that promised inside sources revealed that we'd get Powermac G5s at MWSF, running at 1.2, 1.4, and 1.6 GHz? Yeah, that's what I thought.

    </strong><hr></blockquote>



    I agree with you that it's just impossible to imagine Apple switching to X86 at a time like this. But the report is not from the Register, it's from the Inquirer, which has an excellent track record. They hardly ever deal with Mac rumors at all, but they have almost always been accurate with their rumors including the biggest news from last year, the sale of Alpha to Intel.



    [ 04-07-2002: Message edited by: timortis ]</p>
  • Reply 25 of 46
    razzfazzrazzfazz Posts: 728member
    [quote]Originally posted by AirSluf:

    <strong>Beefing up a 64-bit x86 Darwin with some basic video driver or graphics accelerator support could be a step towards a server/render-farm box that doesn't need to be a full fledged desktop OS provider.

    </strong><hr></blockquote>



    Why would you need a gfx accelerator in a renderfarm or server? Aren't those mostly used in headless configurations?

    Also, again, some more NIC drivers would probably be a prerequisite for that (GigE might be, too).





    [quote]<strong>

    Lotsa targeted horsepower without the resource sapping of Aqua where it's not needed or desired.

    </strong><hr></blockquote>



    I seem to remember that some Connectix developer claimed that the scheduler currently used in OS X and Darwin is sub-optimal for compute-bound stuff.



    Also, more importantly, why would anyone want to use Darwin/x86 for that kind of stuff rather than Linux or FreeBSD? What kind of advantage would Darwin bring here?



    Bye,

    RazzFazz
  • Reply 25 of 46
    xmogerxmoger Posts: 242member
    Porting to osX to x86 would take too long, and not enough developers would stick with it. Certainly this isn't going to happen in the near future.

    I doubt AMD is going to spend resources to get osX running on a ppc hammer. AMD has a reputation of not giving OEM's sweetheart deals on an individual basis. So I don't see why they would they would pay apple for the privilege of making a cpu that 1 company will use and increase their sales maybe 10%? Would there even be more than a handful of 64bit server and scientific apps if they did? They could garner far more customers by having excellent chipsets and optimized compilers available at launch. It's most likely that ms will have hammer support but it's not critical to have it on launch day. Hammer will be the fastest 32bit proc for the forseeable future, so with steady marketshare, 64bit support will follow.
  • Reply 27 of 46
    programmerprogrammer Posts: 3,455member
    [quote]Originally posted by morzipan:

    <strong>But they do have a better option: design a PPC decoder stage to tack onto the Hammer core. No software ports required, everything runs fully native.</strong><hr></blockquote>



    This wasn't the content of the rumour in question though.



    The difficulty in adding a PPC decoder to the front of the Hammer's core depends heavily on how specific to x86 that core is. If its anything like the Intel P3 or P4 cores, then doing this would be very difficult due to some basic differences in the instruction sets. The Hammer is a new core, however, so if a PPC decoder was an original design goal then it would be much easier. Heck, maybe the Hammer is a PowerPC with an x86 decoder on the front! :eek:
  • Reply 28 of 46
    programmerprogrammer Posts: 3,455member
    !!!



    It just occured to me where TheInquirer's rumour may have originated. Sometimes things get completely misrepresented, distorted, and come out rather twisted, right? Well consider this then: Apple / ATI / nVidia have been working on a few extensions to OpenGL which are being proposed as official ARB extensions. If this happens then those extensions will need to be ported to all the platforms. Since version 1 of this code may very well be on the Mac (as Apple is doing much of the work), it would then make sense that "ATI and nVidia are looking at porting the MacOS X graphics drivers to the x86".



    Boring, but a whole lot more plausible that the wild, rampant (but entertaining) speculation that this thread contains.





    <img src="graemlins/bugeye.gif" border="0" alt="[Skeptical]" />
  • Reply 29 of 46
    If at first you dont succeed... clone, clone again!
  • Reply 30 of 46
    junkyard dawgjunkyard dawg Posts: 2,801member
    ding ding ding ding!



    And Programmer solves the riddle!



    Thanks Programmer, you made sense out of a totally absurd rumor.
  • Reply 31 of 46
    airslufairsluf Posts: 1,861member
  • Reply 32 of 46
    bodhibodhi Posts: 1,424member
    At least he could get dates rights when making stories up. MWNY in June?
  • Reply 33 of 46
    razzfazzrazzfazz Posts: 728member
    [quote]Originally posted by AirSluf:

    <strong>There are many ways to get output from a renderfarm. Using a GPU to accelerate the scene rendering does not mean the machine HAS to stream those bits to a monitor, and this is implemented on some current hardware.

    </strong><hr></blockquote>



    It is? Didn't know that. I was under the impression that gfx cards were used for visualisation in the design phase primarily (bringing models to the screen fast so you can work on them), and that the real rendering was later done on the clusters' CPUs (in non-reatltime and at a higher quality level than gfx cards would allow for).

    Then again, maybe this isn't the case anymore nowadays.





    [quote]<strong>The Connectix blokes later recanted and admitted they did a very poor job of writing their whitepaper.</strong><hr></blockquote>



    Oh, ok, never mind then





    [quote]<strong>

    As for why Darwin? Today, you're right, none. But with Apples apparent forays into 3D or at least cinema/video production/compositing, a fast server running an Apple software product would make a nice if not required addition to a total solution package. That whole vertical integration angle.

    </strong><hr></blockquote>



    Hm, got a point there admittedly. I was more looking at it from the end-user perspective rather than from Apple's.

    Not sure how far along clustering support is in Darwin (x86 or not), though.



    Bye,

    RazzFazz
  • Reply 34 of 46
    g::mastag::masta Posts: 121member
    So let's speculate behind closed doors and see what could be. AMD is introducing their Hammer core built on x86-64bit technology. It is well known that the Unix scene is excited about this and *ntel is scared. AMD does not have any US OEMs willing to go against *ntel and feature AMD even though their cores far outperforms *ntel clock for clock for quite some time now, and the Hammer will only severely widen that margin. So Jobs and Sanders have a Mocha Java and a marriage is formed.



    operative word being speculate!!



    IMHO i don't think Apple would do that. It would be an exercise in futility and go against everything Apple stand for. read further on in the article and you realise this is pure guesswork on their part.

    OTOH i could be totally wrong and be eating humble pie very soon.. lol
  • Reply 35 of 46
    zozo Posts: 3,117member
    Programmer has a good suggestion to the riddle.



    But it WOULD be great for Apple to have the possibility to choose between two manufacturers of PPC chips.



    My friend thatt works for Moto, and as we well know all to well, the semi-conducter business at moto is NOT doing well. Rumors of its spin off, sell off are always around the corner.



    The main reason Intel is at 2.4Ghz and AMD is at 1.733Ghz is because of competition.



    Moto has none so they got to be the lazy asses they are. We are still at 1Ghz (a year and a half later than everyone else) and the "Reality Distortion Field" is wearing thinner and thinner every day.



    If AMD could either become a competitor or take over the production of PPC chips for Apple, I guess it would be agreat thing.
  • Reply 36 of 46
    bilogicbilogic Posts: 1member
    [QUOTE]Originally posted by powerdoc:

    [QB]This pure craps . This has been discussed here so many times.

    Athlon and Modern Intel chips are basically RISC processor with a X86 decoder in the front end.



    DOn?t forget that the underlying system kernel is MACH. This give not only the possibility of multiple processors, but als the possibility of a mixed processor environment. Let?s see what Apple is up to.
  • Reply 37 of 46
    Thing is, the mantra that AMD is The One as far as chip development goes is somewhat flawed; it's more the case that their Athlons turned out to be faster on that particular processor revision (processor revisions being what, one every three years now?), and that Intel had to play catch-up until the next one; before that, AMD chips weren't as fast as Intel.



    My point being, we won't know which is going to be faster until we see them, whatever "inadvertantly" leaks out of the labs.



    P.S. the rumour that Xbox2 will use an AMD and the rumour that new Windows won't run on AMD pretty much cancel each other out, huh?
  • Reply 38 of 46
    blackcatblackcat Posts: 697member
    [quote]Originally posted by -@-:

    <strong>according to Mad Mike @ The Inquirer



    <a href="http://www.theinquirer.net/07040202.htm"; target="_blank">http://www.theinquirer.net/07040202.htm</a></strong><hr></blockquote>;



    I can't believe you all read so much into this! It's by Mike McGee formally of The Reg. The Inquirer is a Reg style semi-serious IT web tabloid.



    Besides, it was by Eva Glass, not Mike.



    The article is based on an older article which was about Apple mocking rumour sites.



    Oh the irony...



    [ 04-08-2002: Message edited by: Blackcat ]</p>
  • Reply 39 of 46
    [quote]Originally posted by timortis:

    <strong>



    I agree with you that it's just impossible to imagine Apple switching to X86 at a time like this. But the report is not from the Register, it's from the Inquirer, which has an excellent track record. They hardly ever deal with Mac rumors at all, but they have almost always been accurate with their rumors including the biggest news from last year, the sale of Alpha to Intel.



    [ 04-07-2002: Message edited by: timortis ]</strong><hr></blockquote>



    I think a lot of people are ignoring the fact that Jobs & Co. have been there before. NeXTSTEP and OpenStep ran fine on several architectures, from the Motorola 68K to x86, PA RISC and SPARC. The upper level stuff is definitely portable (Carbon being the biggest hurdle, I suspect.) Already, Apple has Quicktime and Darwin on x86, at least partially.



    My point is that Jobs has no reservation about moving to different hardware when it becomes necessary...



    And I sure as hell would like to have a decent OS to put on the super-quick custom-built AMD box I'm getting with my new wife.



    L.
  • Reply 40 of 46
    Hmm... I have thought about this rumor and the only reason why ATI and NVidia would be looking at AMD Clawhammers is if they were designing a new integrated chipset for PPC Clawhammers. I find it silly to entertain the notion Apple will switch to x86. An AMD chip with PPC instructions is not that far fetched though. NVIDIA and ATI are both trying to get into the chipset market. Apple would likely let the two companies bid on such a project. NVIDIA already makes PC chipsets (for AMD chips no less) and ATI announced that it plans to follow suit. I think this is a very likely scenario.



    [quote] We also understand that both ATI and Nvidia are pushing neck and neck to produce integrated graphics drivers for such a platform. <hr></blockquote>



    I think the Inquirer's source meant integrated graphics chipsets not drivers. What the heck would an "integrated graphics driver" be anyhow?



    [ 04-11-2002: Message edited by: gafferino ]</p>
Sign In or Register to comment.