gp-ul: Single or Dual?

mrmistermrmister
Posted:
in Future Apple Hardware edited January 2014
According to the Forbes article on the new GPUL processor IBM is developing will have only one central processing unit...I assume this is confirming that the chip is single-cored and not multi-cored.



if this is the case, and of course assuming Apple uses these puppies, will we see dual-Regattas in the high end?



My gut says no, but I wonder if all the coding for 2 processors and attendant hype about "2 being better than 1" means they'll want to keep pushing in that direction. I'm worried that it would be more expensive.



Subjectively, would people miss duals? I like the idea that one processor can be specified to a task, and then the other if free to work--are there other advantages that duals have that might be lost if we go to a more powerful single processor?



[ 10-13-2002: Message edited by: mrmister ]</p>
«1

Comments

  • Reply 1 of 37
    henriokhenriok Posts: 537member
    No URL?
  • Reply 2 of 37
    gardenofearthlydelightsgardenofearthlydelights Posts: 963member
    Maybe it's <a href="http://www.forbes.com/technology/newswire/2002/10/13/rtr749520.html"; target="_blank">this article.</a>
  • Reply 3 of 37
    telomartelomar Posts: 1,804member
    Needless post now.



    [ 10-13-2002: Message edited by: Telomar ]</p>
  • Reply 4 of 37
    juniorjunior Posts: 3member
    It's not called Regatta -- that's the name for the IBM big iron that currently runs on the Power4.



    Now, regarding your question, if we were to assume that this chip (the PowerPC 970) was tailor-made for Apple, then things point toward single-chip systems.



    Why? Because it wouldn't make much sense to spend time and money engineering a single-core variant of a dual-core cpu if the thing was bound to end up in a dual configuration anyway.



    But then again, what do I know...
  • Reply 4 of 37
    mrmistermrmister Posts: 1,095member
    At any rate, even if the damn thing isn't code named REGATTA, you know what I am asking about: the future of dual processors.



    [ Edited to remove REGATTA references. ]



    [ 10-13-2002: Message edited by: mrmister ]



    [ 10-13-2002: Message edited by: mrmister ]</p>
  • Reply 6 of 37
    snoopysnoopy Posts: 1,901member
    [quote]Originally posted by Junior:

    <strong>



    . . . Now, regarding your question, if we were to assume that this chip (the PowerPC 970) was tailor-made for Apple, then things point toward single-chip systems. . .



    </strong><hr></blockquote>



    Not necessarily! A single core just means Apple has the option to use the IBM 970 in single or dual configurations. That way, one chip is all they need for the lower and higher end PowerMacs. I would assume a single 970 is at least as powerful as a present dual G4. The top end would be that much more powerful, and probably a workstation class PowerMac.
  • Reply 7 of 37
    miami craigmiami craig Posts: 172member
    Not only is apple strongly dedicated to having the option of dual processors available, but they've painted themself into a corner where they basicaly HAVE to offer it. They've pushed for companies to write dual-processor-capable apps, and now that many companies have, it would be extremely bad business to go ahead and drop dualies from any machine theyre in.



    That being said, Apple doesn't do business normally and i wouldn't put it past them to make an odd (and ultimetly dumb) descision like that. -My $0.02

    <img src="graemlins/lol.gif" border="0" alt="[Laughing]" />
  • Reply 8 of 37
    grad studentgrad student Posts: 184member
    "Why? Because it wouldn't make much sense to spend time and money engineering a single-core variant of a dual-core cpu if the thing was bound to end up in a dual configuration anyway."



    some advantages of having a single core proc that could be used in a multi proc configuration are: each individual processor would be much cheaper due to higher yields, and a smaller die size (among other things). The processor would have a much larger target market, in that it could be used in less expensive or more power sensitive devices such as a single proc workstation or a powerbook, while at the high end, it could be used in a multi-processor configuration to deliver more scalable performance. Besides - the power4 needed modifications/redesign to be used for apple due to its lack of a G4 compatible vector unit, as well as the need for a different/more standard interface to the rest of the system.



    I am very interested to see the specs on this part more closely.
  • Reply 9 of 37
    ast3r3xast3r3x Posts: 5,012member
    [quote]Originally posted by grad student:

    <strong>"Why? Because it wouldn't make much sense to spend time and money engineering a single-core variant of a dual-core cpu if the thing was bound to end up in a dual configuration anyway."



    some advantages of having a single core proc that could be used in a multi proc configuration are: each individual processor would be much cheaper due to higher yields, and a smaller die size (among other things). The processor would have a much larger target market, in that it could be used in less expensive or more power sensitive devices such as a single proc workstation or a powerbook, while at the high end, it could be used in a multi-processor configuration to deliver more scalable performance. Besides - the power4 needed modifications/redesign to be used for apple due to its lack of a G4 compatible vector unit, as well as the need for a different/more standard interface to the rest of the system.



    I am very interested to see the specs on this part more closely.</strong><hr></blockquote>



    I hop apple woudllstay with two, because top of the line dual now would kinda make me think that will be as fast as one of these new chips. So one of these new chips in an imac, ibook, power book, and two in a powermac sounds nice to me! ...although two in a powerbook would catch my eye even more, haha
  • Reply 10 of 37
    outsideroutsider Posts: 6,008member
    I could have sworn this thing would be a dual core part. One core with one third the cache, scaled down L3 tags would make for a small die especially made on 130nm. If this information is solid then I can only conclude that this part was designed to be made to go into a portable upon release as well as full sized desktop machines. I shouldn't question IBM's wisdom when it comes to design. Thinking on die memory controller would be a speed boost i failed to consider it would severely limit the choice of memory for the systems. Similarly multicore chips, for now, will limit the application and versatility of the chip. A dual core chip, while great for a tower, would be less than ideal for a portable. Until the process shrinks more multicore may have to wait for the second generation of the 9XX family.
  • Reply 11 of 37
    jlljll Posts: 2,713member
    [quote]Originally posted by Outsider:

    <strong>Until the process shrinks more multicore may have to wait for the second generation of the 9XX family.</strong><hr></blockquote>



    Already talking about the next version
  • Reply 12 of 37
    catalystcatalyst Posts: 226member
    For anyone looking for a confirmation:



    <a href="http://www-3.ibm.com/chips/news/2002/1014_powerpc.html"; target="_blank">http://www-3.ibm.com/chips/news/2002/1014_powerpc.html</a>;



    [edit:confirmation of the chip's existance that is]



    [ 10-14-2002: Message edited by: catalyst ]</p>
  • Reply 13 of 37
    robsterrobster Posts: 256member
    [quote]Originally posted by snoopy:

    <strong>

    I would assume a single 970 is at least as powerful as a present dual G4. The top end would be that much more powerful, and probably a workstation class PowerMac.</strong><hr></blockquote>



    A single 970 is going to run about 4 times faster than an equivalent G4, then consider the speed Altivec (or similar SIMD) will go on top of that, then consider the speed of memory access on the new bus. This is going to likely be a quantum leap for the Macintosh.
  • Reply 14 of 37
    gfeiergfeier Posts: 127member
    IMHO, single vs. dual processor configurations will be determined by the performance level of these chips on 32-bit software vs. the competition. I don't think we'll see 64-bit software until mid-2004 at the earliest.



    [ 10-14-2002: Message edited by: gfeier ]</p>
  • Reply 15 of 37
    mrmistermrmister Posts: 1,095member
    This Wired article points to use of multiple processors, including quads and octos, as one of the main strengths of the new processor:



    <a href="http://www.wired.com/news/mac/0,2125,55722,00.html"; target="_blank">Wired PPC970 article</a>
  • Reply 16 of 37
    programmerprogrammer Posts: 3,458member
    [quote]Originally posted by gfeier:

    <strong>IMHO, single vs. dual processor configurations will be determined by the performance level of these chips on 32-bit software vs. the competition. I don't think we'll see 64-bit software until mid-2004 at the earliest.</strong><hr></blockquote>



    The 32-bit software performance will probably be 2-3% faster than equivalent 64-bit software. Only in situations where 64-bit integers are being manipulated or 64-bit addresses are required (both rare) will 64-bit software have the advantage because then 32-bit software will need to do a considerable amount of work to get equivalent functionality. The rest of the time the smaller size of pointers and integers will make the 32-bit software faster.



    Note that I'm talking about 32-bit and 64-bit software that has been optimized for this particular processor. There will undoubtably be benefit to recompiling using a compiler that understands the new processor and its pipelines, as opposed to one which understands the 7455 (or earlier) PowerPCs.
  • Reply 17 of 37
    bartobarto Posts: 2,246member
    [quote]Originally posted by Programmer:

    <strong>

    Only in situations where 64-bit integers are being manipulated or 64-bit addresses are required (both rare).</strong><hr></blockquote>



    "Rare" right now.



    When a 64-bit Photoshop arrives... bake-offs of Power Mac G5s (if that is what Apple will call it) vs 32-bit Pentium 4s could get interesting.



    Barto
  • Reply 18 of 37
    amorphamorph Posts: 7,112member
    Also, don't forget that Oracle has started to move to OS X Server in earnest. They can have a 64 bit 9i up and running not long after Apple ships a 64 bit OS X Server - its clustering capabilities are almost custom made for the XServe. Then, of course, Sybase and possibly DB2 are coming, and they'll tax the platform as thoroughly as it can be taxed.



    Lightwave and Maya and the other 3D apps will see an immediate benefit from the 64 bit FP support.



    I would say that there's no shortage of 64 bit apps waiting in the wings. But they're not apps in the Mac's traditional strongholds. 64 bit color is coming, and doubtless Adobe will be on top of that game, but there are more immediate uses for the extra bits. If nothing else, FCP/DVDSP users might appreciate the ability to slurp a &gt;4GB file into memory, and possibly even into &gt;4GB of physical RAM.
  • Reply 19 of 37
    powerdocpowerdoc Posts: 8,123member
    according from IBM the ppc 970 have 53 millions of transistors and he is derivated from the power 4 and have an altivec unit : it's a single core chip.



    Just cannot resist to quote my self :

    [quote] I prefer a single core architecture with a good altivec unit totally independant, than a dual core architecture with bad tricks in order to save transistors.



    A single core means less transistors and less heat. With 50 millions of transistors you can make a very sophisticated single core chip, that can be use in The I mac and powerbook line some months later.

    If the chip is a dual core, we will have to wait a long time before seing it in a I mac or a powerbooK . Apple sell more I mac than powermac, it's very important for him to have good processors for all his product line. <hr></blockquote>
  • Reply 20 of 37
    airslufairsluf Posts: 1,861member
Sign In or Register to comment.