Is Apple G5 = Motorolla neccessary?

Posted:
in Future Apple Hardware edited January 2014
We're all talking about Motorolla 85xx series of chips as being the G5 - but could the G5 be something completely different from the 85xx architecture? If Apple designed the G5 PowerPC it surely doesn't need to have anything in common with any other chip family but the PowerPC instruction set and maybe AltiVec. Or not even that.



Are there any infos on a G4 successor chip other than the 85xx mobile series that Moto is making (which has nothing to do with desktop chips anyway)?
«1

Comments

  • Reply 1 of 26
    thttht Posts: 3,063member
    We're pretty much guaranteed that the hypothetical Power Mac G5 will not use a chip based on Moto's 85xx line. The current MPC 8450, due to ship in 2H 02, isn't much faster, if at all, than the current 1 GHz MPC 7455 shipping in current Power Macs. That's assuming Moto will ship a 1 GHz 8450 in 2002.



    I suppose Moto could come out with a multi-core e500 based chip, which would be interesting I suppose. Cheaper than, but not much faster than, the current dual G4 machines.



    The Register came up with some believable speculation last month. The 7470, nee 7460, is a 0.13 micron 7455 with a DDR MPX bus and support for 4 MB backside cache, shipping in Summer 02. The 7500 is an 11 stage pipeline G4 with RapidIO and speculatively an on-die memory controller due in 1H 03. One would hope this 7500 would have an additional FPU unit in it. This is quite reasonable, but it's speculation at best right now.
  • Reply 2 of 26
    brussellbrussell Posts: 9,812member
    Motorola's roadmap does call the 74xx the G4, and the 85xx the G5.
  • Reply 3 of 26
    matsumatsu Posts: 6,558member
    I've been thinking about this for a little while and I think it holds some merit.



    AMD and Intel weren't that far ahead untill about 18 months ago. When K7 and P4 came onto the scene, clock rates exploded, and performance has actually followed suit in latter iterations. Maybe they're hot and hungry, but P4 2.2 Northwood, is by most accounts, the fastest desktop chip money can buy.



    What's interesting is the time before this explosion in clock rate and (to a lesser degree -- but still significantly) overall speed. Updates were incremental, but with more regularity than on the PPC side.



    It's as if the CPU designers were just ironing out some key concepts and once they were ready/experienced they could push their silicon a lot further than anyone had expected. A plateau followed by an explosion -- the first P4s were scarcely better than the top P3's of the time, but the latest P4s are actually quite good.



    Now look at G4 (and PPC) in general. The updates haven't been quite so regular, but there seems to be a bit of a plateau. Given the behavior of other CPU roll-outs, I might think that the PPC side has similar boosts in the pipe. And, as all the technology is ironed out, we might expect a similar explosion in clock rates once a susbstatially revised design makes its debut.



    Wishful thinking? I dunno. AMD and Intel both kinda levelled off for a while before they got to covering a whole lot Hz in a very short time.



    [ 03-09-2002: Message edited by: Matsu ]</p>
  • Reply 4 of 26
    dhagan4755dhagan4755 Posts: 2,148member
    This line of thinking may be applicable to a future processor beyond the G5. It's only a matter of time, in my opinion, before Apple takes over the processor design arena and begins to farm out the production to IBM. This will happen in years to come, and by that time, Apple will be already selling G5 processor-equipped Power Macs.
  • Reply 5 of 26
    powerdocpowerdoc Posts: 8,123member
    If the G5 is the brand new chip it is supposed to be, he will have a entirely new design (like the athlon and the P4)

    - he will have an altivec unit (first or second generation aka 256 bits)

    - he will be more superscalar than he is now and thus far much transistors : more than 40 millions (expect more than one fpu unit)

    - he will have 512 kb of L2 cache, rapid I/O architecture and on die memory controller.

    - he will not scale has much in mhz as a P4 , but will have more raw power like an athlon, at equal mhz he will outperform easily the G4.



    I'll bet this chip appear in 2003 for MWSF with a new mobo with firewire 2, Serial ATA and DDR ram.



    [ 03-10-2002: Message edited by: powerdoc ]</p>
  • Reply 6 of 26
    marcukmarcuk Posts: 4,442member
    REF: intel/AMD competition,



    suppose AMD didn't exist. A purely single race between PPC and X86 supplied by Moto/Intel. Where would x86 be now?.



    I'd think that probably it'd be 1GHZ G4 v 1.2GHZ P3 with P4 just about to launch at 1.4GHZ
  • Reply 7 of 26
    sdw2001sdw2001 Posts: 16,929member
    [quote]Originally posted by Matsu:

    <strong>I've been thinking about this for a little while and I think it holds some merit.



    AMD and Intel weren't that far ahead untill about 18 months ago. When K7 and P4 came onto the scene, clock rates exploded, and performance has actually followed suit in latter iterations. Maybe they're hot and hungry, but P4 2.2 Northwood, is by most accounts, the fastest desktop chip money can buy.



    What's interesting is the time before this explosion in clock rate and (to a lesser degree -- but still significantly) overall speed. Updates were incremental, but with more regularity than on the PPC side.



    It's as if the CPU designers were just ironing out some key concepts and once they were ready/experienced they could push their silicon a lot further than anyone had expected. A plateau followed by an explosion -- the first P4s were scarcely better than the top P3's of the time, but the latest P4s are actually quite good.



    Now look at G4 (and PPC) in general. The updates haven't been quite so regular, but there seems to be a bit of a plateau. Given the behavior of other CPU roll-outs, I might think that the PPC side has similar boosts in the pipe. And, as all the technology is ironed out, we might expect a similar explosion in clock rates once a susbstatially revised design makes its debut.



    Wishful thinking? I dunno. AMD and Intel both kinda levelled off for a while before they got to covering a whole lot Hz in a very short time.



    [ 03-09-2002: Message edited by: Matsu ]</strong><hr></blockquote>





    I don't agree really. X86 speeds have in fact been incremental. It just seems like they are not. Remember, they hit 1 GHZ is late 99/early 2000......TWO YEARS ago. 18 months ago was mid-2000, with x86 at about 1.2/1.3 GHZ and PPC at 500MHZ. And the P4 is still not as efficient as the P3, even the latest ones.



    I actually think the explosion is clock rates you speak of is happening in a way now on the PPC side, though. Actually, it is probably just a perception because we had zero speed increase for so long. I think we'll either see a G5 or G4 running up to 1.4 or 1.6GHZ by July, but by then Intel will be at 2.6GHZ or so.



    Apple seems pretty firmly routed in the thought that "MHZ doesn't matter". I don't think we'll ever see the clockspeed we are seeing out of Intel/AMD. Though, I do think when things go 64 bit for both platforms you'll see clockpseeds DROP on both sides.



    What we need now is DDRAM, faster HD contollers, etc. That would put Apple ahead I'm sure.
  • Reply 8 of 26
    matsumatsu Posts: 6,558member
    Yeah maybe it is just a perception. Not really a true plateau and then take-off, but the steps seemed to get bigger after the debut of the P4. They'd been dealing in 66Mhz increments during the life of P3. Then P4 brought a 300Mhz jump (if not an equivalent jump in real performance) 500Mhz if you count the three models that debuted the P4 line.



    If we could get a 1.6Ghz (die shrunk) G4 on a modernized motherboard, I think a lot of people would be happy. If it happened by early summer. The longer the wait the faster the replacement will have to be.
  • Reply 9 of 26
    [quote]AMD and Intel both kinda levelled off for a while before they got to covering a whole lot Hz in a very short time.<hr></blockquote>



    The MHz race began after this event June 23, 1999. The AMD Athlon, available in speed grades of 600, 550, and 500 MHz.



    Competition is good, capitalism is cool!

    <a href="http://www.amd.com/us-en/Corporate/VirtualPressRoom/0,,51_104_543_553~778,00.html"; target="_blank">AMD Homepage</a>



    [ 03-10-2002: Message edited by: [email protected] ]</p>
  • Reply 10 of 26
    big macbig mac Posts: 480member
    [quote]Originally posted by [email protected]:

    <strong>

    [SNIP]

    Competition is good, capitalism is cool!



    </strong><hr></blockquote>



    Truly capitalism is the greatest.
  • Reply 11 of 26
    haraldharald Posts: 2,152member
    "Capitalism is cool"



    I vote we keep platitudinous, empty and incomplete comment on this board restricted to the subject of Future Hardware. I benefit from the capitalist system, but don't forget, infant mortality is higher in Washington DC then Havana.
  • Reply 12 of 26
    slackerslacker Posts: 127member
    [quote]Originally posted by Harald:

    <strong>"Capitalism is cool"



    I vote we keep platitudinous, empty and incomplete comment on this board restricted to the subject of Future Hardware. I benefit from the capitalist system, but don't forget, infant mortality is higher in Washington DC then Havana.</strong><hr></blockquote>



    [quote]From <a href="http://www.importanceofphilosophy.com/Index.html?http://www.importanceofphilosophy.com/Politics_Capitalism.html"; target="_blank">http://www.importanceofphilosophy.com/Index.html?http://www.importanceofphilosophy.com/Politics_Capitalism.html</a>;

    <strong>"Capitalism" is conventionally defined along economic terms such as the following:



    An economic system in which the means of production and distribution are privately or corporately owned and development is proportionate to the accumulation and reinvestment of profits gained in a free market.

    Source: Dictionary.com



    This is an example of a definition by non-essentials. An essential definition of capitalism is a political definition:

    Capitalism is a social system based on the principle of individual rights.

    Source: Capitalism.org



    In order to have an economic system in which "production and distribution are privately or corporately owned", you must have individual rights and specifically property rights. The only way to have an economic system fitting the first definition is to have a political system fitting the second definition. The first is an implication of the second. Because the second, political, definition is fundamental and the cause of the first, it is the more useful definition and is preferable.



    Because people often use the term "Capitalism" loosely, "Laissez Faire Capitalism" is sometimes used to describe a true Capitalist system. But this phrase is redundant.



    It is important to define "Capitalism" correctly because a proper definition is a prerequisite to a proper defense. Capitalism is the only moral political system because it is the only system dedicated to the protection of rights, which is a requirement for human survival and flourishing. This is the only proper role of a government. Capitalism should be defended vigorously on a moral basis, not an economic or utilitarian basis. </strong><hr></blockquote>



    So you are saying because there are a lot of alcohol and drug addicted women in DC who raise the infant mortality rate that Cuba is better??? I'll take freedom any day and the consequences that it brings, rather than live in a country where my best option is to hop on a raft and pray I make it to America.



    Just remember, Capitalism doesn't create the problems, corrupt politicians and leaders do. Non effective law enforcement and prosecution cause many of America's problems. But we still live free and can say what we want without fear of retribution. If you don't like it, move to Cuba they need to make up for the people leaving.



    (By the way the infant mortality rate for Cuba is 6.2 and for the entire US it's 7, Washington DC is approx 13)
  • Reply 13 of 26
    brussellbrussell Posts: 9,812member
    I heard the G5 has an SIMD that would reduce infant mortality.
  • Reply 14 of 26
    matsumatsu Posts: 6,558member
    Have you been reading old Kormac threads?
  • Reply 15 of 26
    [quote] An economic system in which the means of production and distribution are privately or corporately owned and development is proportionate to the accumulation and reinvestment of profits gained in a free market.

    Source: Dictionary.com



    This is an example of a definition by non-essentials. An essential definition of capitalism is a political definition:

    Capitalism is a social system based on the principle of individual rights.

    Source: Capitalism.org

    <hr></blockquote>



    I think most socialists would argue that a social system based on the principle of individual rights is unworkable when the means of production and distribution are privately or corporately owned and development is proportionate to the accumulation and reinvestment of profits gained in a free market.
  • Reply 16 of 26
    tjmtjm Posts: 367member
    I guess I'll be boring and stay on-topic...



    Based on miscellaneous comments from a variety of sources (nothing explicit, just trying to put 1 and 1 together), it appears to me that Apple has indeed taken over the chip design, possibly with lots of help from Motorola. The actual G5 chip, though, I think will be fabbed by IBM.



    This would explain why we aren't seeing much from Moto about future desktop processors. When Moto tanked, it seems likely that Apple reacted to save their own butts - meaning keeping PPC development going while Moto got its stuff together again. I have read several comments from people who seem generally knowledgeable and credible (to me, anyway) who insist that Moto will NOT fab the G5 - that it would come from "another major semiconductor manufacturer." That most likely seems IBM. There has been a fair amount of discussion about AMD and Apple, but it seems a stretch for AMD to start fabbing PPC chips.



    Again, I have no "insider" info - I'm just trying to make some sense out of a lot of conflicting information. This looks like a reasonable scenario to me, anyway.
  • Reply 17 of 26
    daveleedavelee Posts: 245member
    [quote] Originally posted by TJM:

    Based on miscellaneous comments from a variety of sources (nothing explicit, just trying to put 1 and 1 together), it appears to me that Apple has indeed taken over the chip design, possibly with lots of help from Motorola. The actual G5 chip, though, I think will be fabbed by IBM. <hr></blockquote>



    I have been thinking about this too and was wondering if anybody had any idea of the financial logistics of this potential approach?

    If the desktop G5 has been developed by both Apple and Motorola (I assume this to be true as it would have to include Altivec) and then fabbed by IBM, what would be the financial implications for Apple?



    I understand that Altivec would have to be licenced by IBM, thus ensuring an income for Motorola, but the fact that they would not actually manufacture the chips would surely mean that the 'royalties' (if that is how it would work) are the only income from the chips (plus a bit of R+D). Would Apple purchase the G5 from IBM (lets say for $250 a pop) and then IBM pay Motorola for Alitivec licencing n each chip? How would this be divvied up?



    Any speculation is strongly encouraged, as it would be quite nice to think there is some possiblity of this alternative.
  • Reply 18 of 26
    outsideroutsider Posts: 6,008member
    I understand that Altivec would have to be licenced by IBM,



    Not true. If IBM's only role is to fab it and Apple and Motorola design it, then there is no need for an IBM license Altivec.
  • Reply 19 of 26
    airslufairsluf Posts: 1,861member
  • Reply 20 of 26
    whisperwhisper Posts: 735member
    [quote]Originally posted by AirSluf:

    <strong>



    Without Mot's explicit permission (licence), nobody can use Mot's proprietary implementation (HW design/layout) even though all AIM players have rights to the instruction set. IBM or anyone else can design their own HW implementation though. In the past IBM did not see a significant market for SIMD other than Apple so they didn't push making their own and Mot didn't give IBM a licence option IBM wanted to live with. IBMs PPC roadmap shows they are adding some sort of SIMD implementation to their future chips, what instructions it will support I haven't yet heard.</strong><hr></blockquote>



    I think you're confusing "using" with "building". AFAIK, Apple can have IBM make whatever Apple wants, but IBM can' do anything with it other than give it back to Apple. IBM hasn't used the chip, just built it. Follow?
Sign In or Register to comment.