Catch up on stories from the past week (and beyond) at the Slashdot story archive

 



Forgot your password?
typodupeerror
×
IBM Technology (Apple) Technology

New info on IBM's Power5 chip (G5's) 28

phreemind writes "There's some news out today on IBM's Power5, which should make an appearance next year. Interestingly, from the sound of this article, they've put a lot of work into power consumption on this processor so that it can go in anything from blades to big iron. This may preclude the need for a specialized low-heat/power version, such as the 970, for anything other than laptops. Oh, yeah, and they hope to use it to wipe Itanium off the map. Check out the article at InfoWorld."
This discussion has been archived. No new comments can be posted.

New info on IBM's Power5 chip (G5's)

Comments Filter:
  • by Andy_R ( 114137 ) on Thursday February 06, 2003 @06:49AM (#5239401) Homepage Journal
    The G5 is the long overdue Motorola successor to the G4.
    The Power5 is the successor to IBM's Power4.

    While the architectures are closely related, the Power4 and G4 are distictly different chips, and the Power5 and G5 will be too.
    • by OppressiveGiant ( 558743 ) on Thursday February 06, 2003 @07:53AM (#5239589) Homepage Journal
      Actually, I believe IBM does produce a large amount of the G4s. They were making the copper G4s. They also produced a large quantity of the G3s. So they have the manufacturer correct, sort of. However they person who wrote the title may be indeed a little confused. I suppose this could mean that the G5 isn't fair away though.
      • by cheezfreek ( 517446 ) on Thursday February 06, 2003 @08:15AM (#5239660)
        Actually, I believe IBM does produce a large amount of the G4s. They were making the copper G4s. They also produced a large quantity of the G3s. So they have the manufacturer correct, sort of.

        Yes, IBM manufactured some of the G3s and G4s under contract, but they didn't design them. For the Power series of chips, it's all IBM, from the design to the manufacturing, so the G5 and the Power5 will be largely unrelated (aside from the basic underlying PPC architecture).

        I suppose this could mean that the G5 isn't fair away though.

        Just because IBM is making the Power5 soon doesn't mean that Motorola is making the G5 soon. They're two different companies with very different design methodologies. If Motorola releases the G5 soon, I'll be surprised. After all, who will their customers for the G5 be? Not Apple, if we believe the rumours.

      • by Matthias Wiesmann ( 221411 ) on Thursday February 06, 2003 @08:21AM (#5239704) Homepage Journal
        I suppose this could mean that the G5 isn't fair away though.
        Well I don't know if the G5 is not far away, but I'ma sure a G5 is not far away. A Gx is not really a product name or a processor model, it refers to a generation number.
        • The G1 was the 601, it was basically a scaled down POWER chip with still some POWER instructions.
        • The G2 were the 6xx series: 603, 604, 604e.
        • The G3 are the 7xx series: 740, 745, 750. This would also, I suppose, include the Flipper processor of Gamecubes.
        • The G4 are the 7xxx series: 7410, 7441, 7445, 7451, 7455.
        • The G5 looks like it will be the 9xx series: 970.
        As to what makes a generation change, its not very clear. From G1 to G2, there was a change in the MMU and caching structure and the removal from POWER instructions. From G2 to G3, I think the wiring changed. The only difference between G3 and G4 is Altivec. The difference between G4 and G5 will probably be numerous: 64 bits (albeit the PPC 620 already was 64 bits) and different interconnect. (Somebody correct me if I'm wrong).

        • The G4 are the 7xxx series: 7410, 7441, 7445, 7451, 7455.

          What's curious is that the PowerPC G4 (7400 series) processors have part numbers eerily similar to the names of discrete logic parts [uiowa.edu]: 7410 is a triple 3-input nand gate; 7441 and 7445 are 4-bit BCD to 7-segment LCD signal converters; 7451 is a dual AND-OR-INVERT gate; 7455 is a 2 wide 4-input AND-OR-INVERT gate.

    • by BigBir3d ( 454486 ) on Thursday February 06, 2003 @09:25AM (#5240133) Journal
      I have to agree. To infer that the Power5 is the new G5 is a huge dis-service to the P5.

      The G-series chips have been nothing special, whereas the Power-series of chips are more than impressive. Look at all the Mac faithfull drooling over the idea of the 970, which is essentially a dumbed down version of the Power-4. That is the rumored "G5" chip. Although, most just call it the 970 now instead of a G-series. Thank god.
      • by dbrutus ( 71639 ) on Thursday February 06, 2003 @09:32AM (#5240201) Homepage
        I'm guessing the marketing droids over at Apple will call the 970 something like a G4-64 (rolls off the tongue, doesn't it) and sometime in 2004-2005 roll out Power5 systems under the G5 moniker. Hopefully, at some point, Motorola will get its act together and get back into AIM with both feet but I'm guessing it'll take an Apple marketshare in the 6-10% range.
        • That could very well be.

          6-10% will not happen for some time though... Last I heard was they were at something near 3-1/4% And that was with the help of OS X and the iPod.
        • I'm guessing the marketing droids over at Apple will call the 970 something like a G4-64 (rolls off the tongue, doesn't it) and sometime in 2004-2005 roll out Power5 systems under the G5 moniker.

          This is _way_ off. The PowerPC architecture is jointly owned by Motorola, IBM, and Apple. Each company has the right to make processors based on the initial PowerPC architecture. IBM more or less stayed away from the consumer market and focused on PPC server chips (although IBM did produce the G3 series of processors). The Gn naming convention is entirely Apple's marketing spin. The G4 is produced by Motorola.

          Now, all the spin about the IBM PowerPC 970 is because the 970 (IBM's name for the processor) is based on the Power4 core which IBM originally produced for it's huge Rigatta processor line. These chips are really awesome because they are just as fast and use a _lot_ less power. Just in power savings alone, mainframe owners save hundreds of thousands of dollars by using Power4 derivatives.

          Needless to say, watering down this processor into a consumer chip is really exciting especially for Apple since lower-power == lower-heat == cuter case designs.

          Now, Apple has announced that they are a "customer" of the 970 but that does not guarentee that it will end up in the next round of iMacs nor that it will become part of the G-brand.

          All in all, if I was looking to buy an Apple computer, I'd wait a few months and see what happens with 970...
    • Motorola's G5 is overdue. It is looking more and more like the 68060, a chip that came years too late, provided adequate performance when it reached market, but failed to penetrate many of the classic 680x0 platforms (Sun 3, HP 9000/300, Macintosh 68k, SGI), mostly limiting itself to embedded VME.

      If any time was appropriate for Apple to move away from the PowerPC platform, now would be it. Either they will move over in the next two years to x86-64, or IBM will assume Motorolas position as the lead PowerPC designer for Apple (I realize that they have done lots of stuff, like coming up with the Power platform, in the past, but Apple seems to lean towards Motorola for actual chips).

      • Sun never even released a 68040 based computer, having long committed to SPARC. I doubt 68060 was even considered, delays or not.

        I found some info on a website, a Sun 3/60 with a 20MHz 68020 was codenamed "Ferrari". Seems pathetic now in the day of 3GHz desktops.
        • Sun released 68000-68030 based machines, usually VME based. I know Sun didn't release a 68040 machine, but they built their name on the Sun 1/2/3 architechtures, and SunOS. None of that was SPARC based, it was Motorola 68k based. It is only with the Sun 4 SPARC designs that Sun moved away from that.
    • You may be mistaken because the article title clearly refers to "G5's" and not "G5s".

      The possessive apostrophe indicates the poster is talking about an entity named "G5" which owns or produces the new Power5 chip. Perhaps "G5" is a previously unknown department of IBM responsible for desktop CPUs.

  • by bill_mcgonigle ( 4333 ) on Thursday February 06, 2003 @08:22AM (#5239708) Homepage Journal
    The [IBM] xSeries team has an Itanium box, and we are out to make sure Itanium doesn't survive ... the pSeries team hopes to relegate Itanium to a niche in high performance computing or better yet exterminate the processor altogether.

    Wow, it's great to see some real red-blooded competitive engineers again. This is how good stuff gets done.

    A couple years ago they would have been squashed by the PR deparment for casting ill light on a potential relationship with a potential business partner.

    Maybe there is hope for IBM afterall.
    • by Anonymous Coward
      I don't think itanium will be dead. No, I'm sure once IBM does away with their xSeries division, itanium will conveniently find its way into Lexmark printers.
  • So is there any chance that they will ever move the instruction set away from its 16 bitness? To load a 64 bit constant, you either have to have a relative memory address or use 5 instructions with the 64 bit PPC modules. Jumps are great too as long as they are within something like 21 bits. While you can deal with massive amounts of memory, it takes hacks that make the segment offsets in the 8086 look like an elegent solution.
  • Is there any reason to think apple would use Power5 for anything? Even though they should, so that they can make a nice fast desktop computer...

    There has been speculation about apple using the PPC970, though. I guess it depends on whether Power5 is at all suitable for a desktop machine (i.e. power and bus requirements).

In the long run, every program becomes rococco, and then rubble. -- Alan Perlis

Working...