Follow Slashdot blog updates by subscribing to our blog RSS feed

 



Forgot your password?
typodupeerror
×
Technology (Apple) Businesses Portables (Apple) Technology Apple Hardware

Freescale Debuts Faster, Cooler G4 37

acsinc writes "The Register reports that Freescale (Motorola's chip division) has launched the 90nm G4, and is planning a dual core version for next year. The chip is faster -- over 1.5GHz -- and cooler than the old chip, but it is also pin compatible. This ought to help provide a speed bump for PowerBooks, which still don't have G5s."
This discussion has been archived. No new comments can be posted.

Freescale Debuts Faster, Cooler G4

Comments Filter:
  • role on duel core..
  • by lhaeh ( 463179 )
    How can you have compatable pinout on a dual core cpu? Are all the motherboards dual cpu compatable? I simply cannot see how something like that would be possable unless it requires new hardware, but then its not exactily a G4 anymore, at least in my mind.

    • Re:Huh? (Score:5, Informative)

      by jhoffoss ( 73895 ) on Tuesday September 28, 2004 @01:00PM (#10375727) Journal
      Read it again. The one just released is compatible; the planned dual-core CPU probably will not be (my assumption).
    • Re:Huh? (Score:5, Informative)

      by Chucker23N ( 661210 ) <chucker23n+slashdot&gmail,com> on Tuesday September 28, 2004 @01:00PM (#10375731) Homepage Journal
      "unless it requires new hardware, but then its not exactily a G4 anymore, at least in my mind."

      What? G4 merely defines the PowerPC generation 4, which means existance of AltiVec, multi-processing capability, etc. It doesn't have to be on the same hardware at all. The G4 exists for everything from embedded devices over laptops to servers.
    • Re:Huh? (Score:5, Informative)

      by adamjaskie ( 310474 ) on Tuesday September 28, 2004 @01:02PM (#10375750) Homepage

      It is NOT dual core. The article is talking about a 90nm G4 processor, and they are planning a dual core for later. The single core is what is pin-compatible.

      The new chip is Freescale's first 90nm G4 and is based on the company's e600 core, the foundation for Freescale's upcoming line of dual-core chips. The 7448, however, contains just one core, clocked to beyond 1.5GHz. It contains 32KB of L1 cache and 1MB of L2, double the 512KB of L2 found in the MPC7447A currently driving Apple's PowerBook G4 and iBook G4 notebooks.
    • Re:Huh? (Score:2, Interesting)

      probably a sneaky little trick like Intel use with the miraculous HT technology i suppose it doesn't really matter if its pin compatible or not, Apple will just have to whack out a new motherboard. Its not like you can just go and build yourself a laptop from bits you can buy off the shelf (granted if you had the will you could but why would you unless you have access to necessary fabs, to make your contraption a laptop), an I assume Apple stick a fair markup on the powerbooks any way, so they could proba
    • Unless you insist on having full access to all features of the two cores, why should the dual core chip have a different pinout than the single core one? You put the shared memory bus plus arbiter logic from the mobo into the chip, same with the interupt logic. Now this doesn't help with the low bus-bandwidth of the previous G4s, but would be entirely possible.
    • Re:Huh? (Score:3, Interesting)

      How can you have compatable pinout on a dual core cpu?

      As others have pointed out, the newly released version is not dual core. But more to the point, what's to prevent the dual core version from having the same, or substantially same, pinout? No matter how many processing units you have on the chip, you still only need one address bus, one data bus, one clock, etc. Maybe multiple cores would require a few extra signals -- I'm certainly not a chip designer -- but it seems unlikely that the pinouts would ha
    • Re:Huh? (Score:3, Informative)

      by obirt ( 713598 )
      Not all, but most of the G4 lineup is. I have an AGP graphics G4 that was a 400 MHz G4 and can be upgraded to a dual cpu daughtercard from GigaDesign, sonnet, or a similar company.

      The dual cpu G4's can have two cpu's on a single daughtercard even though there are two separate daughtercard sockets. You just leave one of them empty. As long as you have a UniNorth rev 7 or later you can run multiprocessor. I do believe that two cpu's is the limit for some reason. I don't think you can have two dual cpu daughte

  • Cunningly...? (Score:5, Interesting)

    by funkdid ( 780888 ) on Tuesday September 28, 2004 @01:39PM (#10376100)
    Cunningly, the 8641D can not only appear to the host OS as two processors, but is capable of running a separate operating systems on each core.

    My that is rather cunning...!?

    So when the dual core G4s do come out I could run 2 separate operating systems simultaneously? That's odd, how (and which ones) is that pulled off. Getting a dual boot mac is enough of a pain in the ass, to have it multi boot is, yeah I don't see that happening.

    • Re:Cunningly...? (Score:5, Informative)

      by Guy Harris ( 3803 ) <guy@alum.mit.edu> on Tuesday September 28, 2004 @05:06PM (#10378346)
      Cunningly, the 8641D can not only appear to the host OS as two processors, but is capable of running a separate operating systems on each core.

      So when the dual core G4s do come out I could run 2 separate operating systems simultaneously?

      Not necessarily. Freescale's Web site [freescale.com] speaks of them as a vendor of embedded microprocessors; perhaps that feature is intended for use in embedded systems, where the OS can be tweaked as necessary to handle two processors sharing an I/O bus and peripherals, so that doesn't necessarily imply that the dual-core processor magically turns any system into which it's plugged into two separate systems that can boot separately and communicate separately with peripherals on the system.

      (BTW, when I tried Googling for "8641D freescale" and "MPC8641D freescale" to see if I could find anything giving details about that feature - nothing turned up - Google asked whether I wanted "firesale" instead of "freescale". Is Google expressing its opinion of Motorola's spinning off Freescale?)

      (No, don't bother explaining to me how that Google feature works; I'm already aware of it. "It's a joke, son....")

  • new powerbook cpu (Score:5, Insightful)

    by boomerny ( 670029 ) on Tuesday September 28, 2004 @01:44PM (#10376132)
    this cpu will probably be in the last generation of G4 Powerbooks before they move to G5. Personally I think I'd rather have a nice, cool running, fast, energy efficient, 3rd or 4th gen G4 than a 1st gen G5.
    • Re:new powerbook cpu (Score:5, Interesting)

      by Dr Reducto ( 665121 ) on Tuesday September 28, 2004 @02:09PM (#10376419) Journal
      And the other benefit is that people will start getting rid of G4's to try to get a nice shiny new G5. Everyone who i thinking about a Powerbook, but can't quite justify it is waiting for the G5.
      • Re:new powerbook cpu (Score:5, Interesting)

        by TheRaven64 ( 641858 ) on Tuesday September 28, 2004 @02:31PM (#10376666) Journal
        Personally, I'm more interested in a dual-core G4 powerbook. I do a few things that would possibly benefit from 64bit address space, but usually not on my laptop. On the other hand, I do a lot of highly parallel stuff while on the go, and so I would notice a lot more of an improvement from a dual core CPU. Not to mention the fact that a dual 1.5GHz machine will run a lot cooler than a 3GHz one (since power is roughly proportional to the square of the clock speed for ICs), not to mention the fact that the G4s have a history of being significantly cooler than the G5s. A dual 2GHz G4 would be really nice (not that my 1.5GHz PowerBook feels slow, but more speed is always nice)
    • not to mention the fact that a 1.8 Ghz G4 and a 1.8 Ghz G5 should be fairly comparable. Other than that damned bus speed. But there is a light at the end of the tunnel... the dual core should have a RadidIO bus finally ridding us of that horrible constraint!...

      this is going to be an interesting marketing conundrum to Apple when it happens because how are they going to sell us this new G4 which in all respects is as good as a G5 (sans 64Bit) and possibly better (cooler running, better FPU/SIMD performance
  • The one thing that is putting me off buying a G4 powerbook is the thought that in a year or two when the G5 power book comes out, the whole OS will probably go 64 bit native and the old powerbooks will no longer get software upgrades.
    • Re:G4's last gasp? (Score:4, Informative)

      by Ster ( 556540 ) on Tuesday September 28, 2004 @03:55PM (#10377588)
      That's not going to happen.

      Apple would be nothing short of certifiably insane if they stopped supporting 32-bit machines before probably five years after they stopped shipping them.

      Let's say the entire lineup goes 64-bit tomorrow. They're wouldn't require 64-bit machines for OS upgrades until 2009 or so. Individual *features* might require 64-bits, but none of the core functionality.

      For example, Panther still supports original iMacs, which are from '98 ('97? I really should know that...). Features like Quartz Extreme won't work, but the OS itself installs and runs fine. Panther updates install and work just fine.

      -Ster
    • Re:G4's last gasp? (Score:5, Insightful)

      by Graymalkin ( 13732 ) * on Tuesday September 28, 2004 @04:13PM (#10377765)
      Tiger isn't going to be released until sometime in the first half of 2005, even then it will contain fat binaries which are binaries containing multiple versions of the same executable in different ISAs or what have you. Tiger then will have 32-bit and 64-bit versions of all of its frameworks and the OS will pick which to use at runtime based on what processor the system is running on. This was a capability of OpenStep, fat binaries allowed you to compile for different architectures yet only distribute one executable file without JITC.

      After Tiger there isn't likely to be another OSX upgrade until the first half of 2007 or so. So between now and the release of Tiger's successor you've got plenty of time to use a Powerbook. Even when Tiger's successor (Ocelot?) is released it will very likely still support old 32-bit Macs considering there will still be millions of them in use. It was only with Panther's release the Apple dropped official support for the last of the Old World machines, Wallstreet Powerbook and Beige G3 PowerMac, though with a little hacking [opendarwin.org] they work just fine. Those systems were both five years old when Panther was released.
    • in a year or two when the G5 power book comes out, the whole OS will probably go 64 bit native and the old powerbooks will no longer get software upgrades.

      Apple has too much interest in selling OS upgrades to people with still-usable hardware for that to happen. I have a 1998 vintage iMac (G3) which has been a modest source of income for Apple via OS upgrades since then, leading to (so far) version 10.3. To say nothing of the G3 iBooks that Apple was still selling a year ago, which they aren't about to

  • Missing the Bus... (Score:5, Interesting)

    by valkraider ( 611225 ) on Tuesday September 28, 2004 @04:05PM (#10377657) Journal
    What some people don't realize, is that more than the extra Mhz that this processor provides, it does up the bus speed to 200Mhz (current 1.5Ghz G4s are at 167Mhz). That is a pretty hefty increase, so they should perform quite well...

    But having said that, I (who own a 1.5Ghz 15' Powerbook) would rather see the G5 in a laptop.

    While the G5 is a GREAT processor, the 64bitness is not it's strongest points. The G5 is simply a very efficient processor... But still, most apps are not CPU constrained, but Bandwidth constrained. In fact, I would bet that a 2Ghz G4 on a 200Mhz bus would still be outperformed by a 1.6Ghz G5 on a 800Mhz bus.

    What I would like to see would be a ~ 1.5Ghz G5 laptop with a 1Ghz bus and a 7200rpm disk and a good video card. They could do that right, and not have heat problems. But it wouldn't be cheap... :) The common 7200 RPM drives and high end video use a lot of power and generate a lot of heat...

    Interestingly, if you don't believe how much things OTHER than the processor can impact performance, if you have a PowerBook (especially one with a 4200RPM drive) - Run XBench, note the score, then plug it into a 7200RPM firewire drive, and run it again (using the firewire drive instead of the internal). Note the score. That one little change should give you a pretty good speed boost!
    • "What I would like to see would be a ~ 1.5Ghz G5 laptop with a 1Ghz bus and a 7200rpm disk and a good video card. They could do that right, and not have heat problems. But it wouldn't be cheap... :) The common 7200 RPM drives and high end video use a lot of power and generate a lot of heat..." So you want the 8461D dual-core, then? What Freescale announced today is that they'd be at least 1.5ghz, with an on-die memory controller than can access up to 667mhz RAM (replacing the "bus" in roughly the same way
  • So, can I replace the 800MHz G4 in my iBook with one of these?
    • I was just thinking the same thing. I picked up a 1 GHz iBook G4 today. (Open box demo unit woo hoo, typing onit right now - my first G4). Obviously, it won't be convenient. I assume it isn't just a zif socket, but hypothetically, if I was good with a soldering iron, would it go? Is there any way to get the old CPU out without breaking it? Not that my current one is slow for my use. (I've heard good things about X Code, so It will be a coding box, and Safari is pretty nice for a web browser... No ga
    • by Anonymous Coward
      Nope... well not easily at least.

      The CPU in the iBook is integrated to the main logic board.

Those who do things in a noble spirit of self-sacrifice are to be avoided at all costs. -- N. Alexander.

Working...