Apple introduces Power Mac G5 Quad and Power Mac G5 Dual

Apple today unveiled its new Power Mac G5 desktop line featuring the Power Mac G5 Quad, providing quad-core processing with two 2.5 GHz dual-core PowerPC G5 processors. All Power Mac G5 models now feature dual-core processors, a new PCI Express architecture and higher performance graphics options including NVIDIA’s Quadro FX 4500, bringing the industry standard for workstation graphics to the Mac.

“The Power Mac G5 Quad delivers the workstation performance our creative and scientific customers demand,” said Philip Schiller, Apple’s senior vice president of Worldwide Product Marketing in the press release. “With quad-core processing, a new PCI Express architecture, and the fastest workstation card from NVIDIA, the new Power Mac G5 Quad is the most powerful system we’ve ever made.”

With two 2.5 GHz dual-core processors, each with 1MB of L2 cache, the Power Mac G5 Quad delivers breakthrough performance on applications used extensively by creative professionals and scientists. For example:

• Final Cut Pro 5 encodes SD video content up to 60 percent faster on the new Power Mac G5 Quad when compared with the Dual Processor 2.7 GHz Power Mac G5
• Adobe Photoshop runs a series of 45 commonly used filters and actions 43 percent faster on the new Power Mac G5 Quad when compared with the Dual Processor 2.7 GHz Power Mac G5
• Adobe After Effects renders video effects up to 69 percent faster on the new Power Mac G5 Quad when compared with the Dual Processor 2.7 GHz Power Mac G5
• LightWave 3D renders animation up to 59 percent faster on the Power Mac G5 Quad when compared with the Dual Processor 2.7 GHz Power Mac G5
• The scientific benchmark Bioinformatics Benchmark System (BBS version 3) completes a comparison of genomic sequences up to 39 percent faster on the Power Mac G5 Quad when compared with the Dual Processor 2.7 GHz Power Mac G5

The Power Mac G5 line offers leading-edge expansion with industry standard PCI Express architecture, providing four expansion slots to support high-performance video and audio devices and multiple standard graphics cards to drive an array of up to eight displays. The Power Mac G5 delivers up to 16GB of 533 MHz DDR2 SDRAM and now features two Gigabit Ethernet ports — ideal for customers working in an Xsan environment. Providing industry-leading connectivity and high-performance I/O, every Power Mac G5 also includes one FireWire 800 port, two FireWire 400 ports, four USB 2.0 ports, two USB 1.1 ports, optical and digital audio input and output, and built-in support for AirPort Extreme and Bluetooth 2.0+EDR. The new Power Mac G5 line also includes Apple’s next generation mouse, Mighty Mouse, featuring up to four programmable buttons and an ingenious Scroll Ball that lets users scroll in any direction — vertically, horizontally and even diagonally.

The new Power Mac G5 comes standard with either the NVIDIA GeForce 6600 LE with 128MB of video memory or the NVIDIA GeForce 6660 with 256MB of video memory, both provide support for dual-displays and Apple’s breakthrough 30-inch Cinema HD Display, now priced at $2,499. The NVIDIA GeForce 7800 GT with 256MB of video memory and the NVIDIA Quadro FX 4500 with 512MB of video memory are both available as build-to-order options. The Quadro FX 4500 is ideal for demanding animation, special effects and scientific visualization applications such as Alias’ Maya and DeLano Research’s PyMol. The Quadro FX 4500 can drive up to two 30-inch Apple Cinema HD Displays and includes a stereo 3D port to connect goggles for stereo-in-a-window applications.

The 2.0 GHz and 2.3 GHz Power Mac G5 Dual are available now and the 2.5 GHz Power Mac G5 Quad will be available in early November through the Apple Store, Apple’s retail stores and Apple Authorized Resellers.

Included software: Mac OS X, Spotlight, Dashboard, Mail, iChat AV, Safari, Address Book, QuickTime, iLife (includes iTunes, iPhoto, iMovie HD, iDVD, and GarageBand), iCal, DVD Player, Classic environment, Art Directors Toolkit X, FileMaker Pro Trial, GraphicConverter, Microsoft Office 2004 for Mac Test Drive, OmniGraffle, OmniOutliner, QuickBooks New User Edition, Zinio Reader, Xcode Development Tools.

The Power Mac G5 Dual, with a suggested retail price of $1,999 (US), includes:
• dual-core 2.0GHz PowerPC G5 processor
• 512MB of 533 MHz DDR2 SDRAM expandable up to 16GB
• 160GB Serial ATA hard drive running at 7200 rpm
• NVIDIA GeForce 6600 LE with 128MB of GDDR SDRAM
• three open PCI Express expansion slots: two 4-lane slots and one 8-lane slot
• dual Gigabit Ethernet ports
• 16x SuperDrive with double-layer support (DVD+R DL/DVD±RW/CD-RW)
• ships with Mighty Mouse and Apple Keyboard

The Power Mac G5 Dual, with a suggested retail price of $2,499 (US), includes:
• dual-core 2.3 GHz PowerPC G5 processor
• 512MB of 533 MHz DDR2 SDRAM expandable up to 16GB
• 250GB Serial ATA hard drive running at 7200 rpm
• NVIDIA GeForce 6600 with 256MB of GDDR SDRAM
• three open PCI Express expansion slots: two 4-lane slots and one 8-lane slot
• dual Gigabit Ethernet ports
• 16x SuperDrive with double-layer support (DVD+R DL/DVD±RW/CD-RW)
• ships with Mighty Mouse and Apple Keyboard

The Power Mac G5 Quad, with a suggested retail price of $3,299 (US), includes:
• two dual-core 2.5GHz PowerPC G5 processors
• 512MB of 533 MHz DDR2 SDRAM expandable up to 16GB
• 250GB Serial ATA hard drive running at 7200 rpm
• NVIDIA GeForce 6600 with 256MB of GDDR SDRAM
• three open PCI Express expansion slots: two 4-lane slots and one 8-lane slot
• dual Gigabit Ethernet ports
• 16x SuperDrive with double-layer support (DVD+R DL/DVD±RW/CD-RW)
• ships with Mighty Mouse and Apple Keyboard

Build-to-order options include up to 16GB of NECC or ECC DDR2 SDRAM, up to two 500 GB Serial ATA hard drives running at 7200 rpm, NVIDIA GeForce 7800 GT with 256MB of GDDR3 SDRAM, NVIDIA Quadro FX 4500 graphics card with 512MB of GDDR3 SDRAM, AirPort Extreme and Bluetooth module, Apple Wireless Keyboard and Apple Wireless Mouse, Xserve RAID and Apple Fibre Channel PCI Express Card.

See the new Apple Power Mac G5 models here.

Advertisement: Order the new Power Mac G5 from the Apple Store now. Dual-core PowerPC processors, a modern PCI Express architecture, and wicked-fast workstation graphics. From $1999. Free shipping.

Related MacDailyNews articles:
First benchmark tests of Apple’s new Power Mac G5 dual-core machines – October 19, 2005

57 Comments

  1. What will be funny to see is all the people who say “I have been waiting for a QUAD G5 with better grafx before I upgrade.” Next you’ll hear, “Well what I really wanted was a Quad Xeon”

    This is a nice system, too bad the Powerbooks are still so far behind.

  2. Dual dual CPUs on the top end PowerMac is great, but who ever heard of the clock speed going DOWN on a new pro computer?? Though it’s not a big drop (from 2.7 to 2.5), you have to wonder what’s next, 2.3 GHz?
    Stumped,
    Jake

  3. $3299 on the high-end!?!?!

    Man, that should’ve been $2999 instead.

    Oh well — I guess if you’re at somewhere around $3G already, what’s another $300, eh?

    Also, it’ll be really interesting to find out what the cost is to max-out the RAM in these things. 16GB? Yikes.

  4. Apple will gouge you for $11K plus for 16 GB, so it will probably be about $5-6 K from other vendors

    Glad the 23″ and 30″ monitors came down a bit. Still probably can’t justify the 30″, though…gotta spend the extra $ on RAM

  5. So we wait about 12 months for a decent PowerMac update, and to reward us they jack up the pricing by 10% on the top end. Did they do this because “it’s the top model”, or because of chip “supply constraints” or because they wanted to piss off Think Secret who’d mentioned a $3K price point? Or is it simply a different version of the iPod tax, which is also 10%. Or is this to compensate for the falling G5 sales numbers over the past 12 months? Or is just for pure profit?? (again partially due to falling numbers.)

    Ok, rant over (mostly). I’m just jacked off coz I am seriously considering buying one, but another $300 plus the associated $30 extra in tax kinda sucks. Kinda like they’re saying “Here’s a hot product you’ve been waiting a year for, on, and check out the great new price!” Suckers.

    Still, it’s a nice update at the top end, but the bottom/middle haven’t really changed chip speeds (although I’m guessing these newer chips are faster internally… anyone know?) It’s great that the bottom model isn’t crippled anymore too (assuming there are no hardware bugs of course.)

    Keep in mind folks, that in another 12 months, dual core will be normal.

    I blinked twice at the graphics – it’s a 6600, not a 6800 graphics card. D’oh.

  6. Anyone else notice how Apple has already replaced the new Power macs as its front page highlight on its website and now is back to video iPod and iMac???
    Apple knows very, very, very few consumers are interested in these pricey computers….

  7. Re: Jake –

    Even though the clock speed is lower than the older models, the dual core configuration will significantly outperform the single core version. Think of it as having 2 small V6 engines in your car as opposed to one big V8.

    What you should be stumped by – and what stumps me – is that the Xbox 360 has a Triple Core PowerPC CPU (the same family as the G5 CPU) running at 3.2GHz…

  8. Queezie,

    All of today’s announcements were aimed at “Pro” users. And yes, there is a much smaller market for “Pro” users who need this bleeding edge performance. But those who like myself make a living by using this stuff are willing to pay for it.

    It is nice to see that Apple still has us “Pro users” in mind… Afterall, we are the ones who kept the company alive before the iPod and iMac..

  9. why can’t apple throw a dual core into the imac (for a price). That would knock me off my 3.5 to 4 year replacement schedule. Are they too hot? I thought they were comparable. Don’t they have more sophisticated power management? Are they too scarce?

  10. Buffy – I don’t think anyone who does their homework will be wishing for a dual-core Xeon anytime soon. I just read a benchmark report on this beast, and about all they do is double the heat. (OK, a slight exaggeration, but it was a very disappointing benchmark.) (Love that OSX built-in spell checker, or I woulda really clobbered that word exaggeration.)

    I’ll be reporting my experiences with that new quad G5 when it arrives. Been waiting for it for over a year now.

    MW – needed, as in, “What’s needed now is a stock jump of 5 bucks to get a free Quad.”

  11. b,

    Because the iMac is a consumer level machine, and most consumers do not need or would use dual core performance. If they do, well then they should buy a Powermac. This is what keeps the consumer and pro machines distinguishable.

    And yes, ther are too hot. The Quad Powermac uses a liquid-cooled radiator…. Would never work in an iMac.

  12. Trevor,

    You’re right – iMac is consumer level, which is why they won’t for the shortterm include a dual core CPU. What I can’t work out is why the $1700 consumer model includes such options as wireless airport and bluetooth, and why on a $3300 computer you *STILL* have to pay another $99 for them.

    If this isn’t gouging pro’s what else is?? Of course, perhaps Apple figures that many pro’s don’t want/need that stuff, but to charge those who do yet another $100 is pretty shitty.

    It appears that recently, Apple’s consumer products have been very good value. The top end are like luxury cars, where even a fairly basic option jacks up the price significantly (relatively speaking)

  13. Trevor,
    In a couple years, all new personal computers will be dual core. I’m not a “professional” user but I would love the benefit of a dual core in an imac. In fact, I’d probably rather have a dual core 1.6 ghz then a 2.1ghz single core. I think consumers need the dualness more than the ghz.
    If your talking about Apple’s need to differentiate their consumer and pro models, that’s just marketing. Also, Apple may feel that 2.1ghz “sounds” better than 1.6ghz dual core, that’s marketing too.

  14. Never… buy RAM memory from Apple. Not that their RAM has any issuses, rather that Apple charges 2 to 3 times and more of the going rate for RAM. Buy your memory somewhere else and you’ll have the same quality and guarantee but, you’ll have a wad of cash left over.

  15. Wow. I think I need to become a professional prognosticator. Pinned the PowerMacs right on. I’ll be curious to see if I got the PowerBooks right. ” width=”19″ height=”19″ alt=”LOL” style=”border:0;” />

    Okay, done with horn tooting. Time for some fun conspiracy theories.

    Wander over to Apple’s site and check out the benchmarks. A few entertaining things to note:

    1. Apple is no longer benchmarking against Intel. The main PowerMac page pits the Quad against the Dual 2.7GHz.
    2. Apple has a special “Upgrade” section pitting the various PowerMac models against a dual 1.42GHz PowerMac G4.

    I gotta admit, I’m vaguely intrigued by this.

    For point #1, is Apple not benchmarking against Intel because Intel and Apple are friends now? Or do the Quad Macs get smoked and no amount of creative benchmarking can obscure this fact?

    For point #2, is Apple trying to drag any laggards that are sitting on PowerMac G4s and waiting for Intels into the camp? Maybe get the “even” upgraders? (I, myself, am an “odd” upgrader–G3 to G5 to, probably, G7)

    To wander off on a different tack, Apple still smokes Dell on price. I’d love to see a performance comparison, but the PowerMac G5 Quad is $3399 with 1GB of RAM. The Dell Precision 470 with two dual-core 2.8GHz CPUs, 1GB of RAM, a 250GB hard drive, a 48X/32XCDRW AND 16XDVD+/-RW w/Sonic DM and DVDit! SE, CyberLink PowerDVD, a 256MB PCIe x16 nVidia Quadro FX 3450, Keyboard, Optical Mouse, and no floppy drive comes to $5454 at Dell’s Small/Medium Business Store. That’s a difference of $2055! Heck, you could buy another dual 2GHz for that kind of savings!

    (Entertaining aside: I tried to configure the Dell with 512MB of RAM, like the Mac. It told me the memory was “incompatible.” When I upped it to 1GB, no problem. According to Dell, they’re both DDR2 400Mhz ECC. Why do people trust these guys?!)

  16. I was waiting on these and I have to say I’m not as enthusiastic as I thought I would be.

    $300 more than the previous top of the line. – Love to see 3rd party speed comparisons with the older line. Bet the loss of MHz and the inability of most applications to really take advantage of an extra cpu/core will make this upgrade not much of one.

    $800 more for an extra chip + 200MHz… steep by any comparison, also this is a single cpu going for the price of a duelie. Seems Apple found a way to increase the profit margin on the PowerMac.

    Also, the line seems like a slick way to guarantee another simple ‘step up” upgrade in another six months prior to the Intel switch, when they all become “Quads”.

    Is still may buy one of course. It’s a sickness.

  17. All exciting stuff, but just one thing… I’m not sure about the name of the photo app… I mean… aperture? Seriously? I know it’s a little adolescent, but couldn’t they have thought of a less giggle-worthy name? I can see the joke posts and headlines now.. “I have a big problem with my aperture”, “please help me instal plugin in my aperture” , “Apple’s aperture feels good to computer geeks” etc etc.

  18. invasifspecies – glad to educate you:
    Aperture is a photography term….

    The word “aperture” means an opening, from the Latin aperire, to open.

    In optics, it is that portion of the diameter of an object-glass or mirror through which light can pass free from obstruction. It is equal to the actual diameter of the cylinder of rays admitted by a telescope.

    In photography specificly, the aperture defines the size of the opening in the lens, which in advanced cameras can be adjusted to control the amount of light reaching the film or digital sensor (CCD or CMOS). In combination with variation of the shutter speed, and variation in film speed (ISO), this will regulate the photograph’s degree of exposure to light. Typically, a fast shutter speed will require a larger aperture to ensure a sufficient exposure to light, just as a slow shutter speed will typically require a smaller aperture to prevent excessive exposure to light.

  19. Peter you forgot to add the top of the line Quadro video card to your configuration which makes the Mac only $400 not $2000 less than the Dull.

    Bottom line I just bought a 2.3 Dual, upped the RAM to 4GB, tossed a 500 GB HD in there and with Apple Care walked away for $3398+tax (thank god my wife is in grad school, I love that edumacation discount).

  20. “but who needs 8”

    Actually, if you want to do a 3×3 mosaic (7680 x 4800) you would need 9. Thus eight is still not enough.

    Yet it is a step forward from the 1988 version Macintosh II which could support six concurrent displays. (Although I know of no one outside of Apple engineering staff who could get more than five to work concurrently — install that sixth card and it all went blank even after repeated attempts and lots of tinkering.)

  21. Peter says: “… is Apple not benchmarking against Intel because Intel and Apple are friends now? Or do the Quad Macs get smoked and no amount of creative benchmarking can obscure this fact?”

    I think you may need to invert your thinking – the fact that Apple isn’t comparing them with Intel PCs anymore isn’t because the Quad G5s are getting smoked. It’s because the Intel machines aren’t looking too good in comparison.

    I gaurantee, once the tech sites get a hold of one of these monsters and puts it head-to-head against Intel’s best dual cored, hyperthreaded rigs (which would be their closest equivilent to having 4 CPUs on one system), the results will be in the G5’s favor.

    This indicates two things: One is that the PowerMac (and probably the XServe) will be the last products to be ‘Macintel’d’. Not that Apple will keep upgrading it with the best that IBM has to offer (see how they ignored the dual core G4 on the PowerBook for comparison), just that the performance boost we see here will actually carry for quite a while before Apple gets around to designing a system around an Intel x86 CPU that can actually surpass it.

    Two is – broken record time – people will finally, hopefully, start realizing and accepting that Jobs sold us a wagon load of manure regarding the performance ‘dead end’ that the PPC architecture was leading him to. This transition to Intel (and their hardware DRM) is a ‘money-move’ for video content – nothing more, nothing less.
    ” width=”19″ height=”19″ alt=”cool smirk” style=”border:0;” />

  22. Odyssey67, Jobs never said that the POWER-architecture represented a “dead end”. The implication was that IBM could design chips that kicked Intel’s ass all over the place, if tripling your power bill and having a room-size refrigeration unit were no problem. What Jobs said was that “per watt” Intel had a better road map. Which means that things like PowerBooks will never get faster without the switch. It was always implied that Pro-level desktops will continue to use IBM chips for at least a year or two after the first Intels come out. Don’t be surprised if MacWorld shows off Intel Powerbooks, Intel iBooks, and Intel Minis (even if they don’t actually ship for a month or two after the show).

  23. “Peter you forgot to add the top of the line Quadro video card to your configuration which makes the Mac only $400 not $2000 less than the Dull.”

    Fair enough. I didn’t stop to think about the graphics card on the Mac–I just tried to match the memory and picked the cheapest one from Dell’s mix. So, I’ll add in the Quadro FX 4500 and watch the price of PowerMac G5 jump to $5049.

    Of course, I’ll also try to be fair by giving the Dell the same Quadro FX 4500 and I’ll watch the price of the Dell jump to $6069.

    So the Mac is $1020 less than the Dell. So forget buying another G5 with the extra money–you’d have buy two low-end Mac minis instead.

  24. invasifspecies….still taking wedding photos, eh?

    Wait, I know, I will say something clever like:
    “if you kiss my aperture you’ll see it tastes of sarcasm.”

    Guess nobody else echoed your attempt at humor or sarcasm. Oh well, back to the dark room….

  25. “[…] the fact that Apple isn’t comparing them with Intel PCs anymore isn’t because the Quad G5s are getting smoked. It’s because the Intel machines aren’t looking too good in comparison.”

    Well, the obvious question is, if the Intel machines aren’t looking good in comparison, why isn’t Apple promoting this? Is it because Intel and Apple are friends now? Is it because Apple doesn’t want the Intel machines to look bad because Apple has publicly said that they are moving to Intel?

    I agree, though, that the PowerMacs will be the last machines to go to Intel.

    But I don’t agree with the conspiracy theory as to why Apple did it. I agree that it was about money, but not in the way you think.

    Essentially, Apple was paying for PowerPC development. IBM had the brains and the fabs and what IBM wanted Apple to do is pay for the development of the chips that Apple wanted. If IBM felt they could use them, too, they’d give Apple a price break. IBM wanted the same model that they had with Sony and Microsoft.

    Apple decided it would be cheaper to switch to Intel than to pay money to IBM to stay ahead in the race. Without Apple paying money and pushing IBM/Freescale for faster chips, the PowerPC will stagnate on the desktop, Intel will surpass them, and Apple will jump aboard with Intel. Remember that Apple doesn’t have to pay Intel to design faster CPUs–there’s already a huge market for faster x86 chips.

    There’s no DRM conspiracy, it’s just plain economics.

  26. TommyBoy says: “Jobs never said that the POWER-architecture represented a “dead end”. The implication was that IBM could design chips that kicked Intel’s ass all over the place, if tripling your power bill and having a room-size refrigeration unit were no problem. What Jobs said was that “per watt” Intel had a better road map.”

    What he said was ‘We don’t know how to build the great things we want to build with PPC’ (not a direct quote, but pretty close). And yet both IBM and Freescale have lowpower single core AND dual core versions of their CPUs available right now. And yet every supercomputer manufacturer that choses the Xserve G5 cites it’s energy consumption and low temperatures as the first two reasons why. And yet Intel’s roadmap isn’t as good as even AMD’s in terms of performance features, and that company is closing the gap fast on the power front as well.

    Look, the PentiumM/Centrino chipset is a fine piece of hardware, and clearly the best to be had for a laptop today. But it’s based on Pentium III technology, and soon will be matched by it’s most direct x86 competitor. If you’re telling me that IBM somehow couldn’t do the same thing, when they’ve been a pioneer in their own low-power areas (like Silicon On Insulator technology), I just don’t buy it.

    “Which means that things like PowerBooks will never get faster without the switch.”

    Powerbooks could have ALREADY gotten faster – had Apple used the dual core G4s that are available right now. The fact that they didn’t had nothing to do with technology – it had everything to do with market timing. You can say, well they committed to Macintel so they shouldn’t spend any more money on PPC based Powerbooks. Or you can claim that they won’t bring a dual core G4 to market regardless, when the next version of the PB will be a dual core Macintel that it can be directly compared to. Either way, tin hat or no tin hat, technology limitations are clearly not at issue.

    “It was always implied that Pro-level desktops will continue to use IBM chips for at least a year or two after the first Intels come out. Don’t be surprised if MacWorld shows off Intel Powerbooks, Intel iBooks, and Intel Minis (even if they don’t actually ship for a month or two after the show).”

    Right you are. I think this exactly how it will play out as well. However, I think the G5 line-up is left for last simply because the direct comparisons with their Macintel successors would be even more unfavorable without a period of ‘stasis’, similar to that which Apple applied to the G4 notebooks.

    The dualie G4s would have been uncomfortably close to the dualie x86s that are coming, in all ways but absolute power consumption (I should point ot that I’m now on my 4th battery hour on my brand new iBook as I write this, so G4s aren’t exactly hurting on that front even now). But the dual/quad G5s are going to especially need a ‘cooling off’ period. When the benchmarks come in, I think its capabilities are going to be so impressive, that people are going to wonder why the low power version of the 970FX (maxing at something like only 25watts) couldn’t be similarly enhanced (i.e. dual core’d), and thus smoke a few Pentium Ms in the process.

    Jobs’s “Performance per watt” spiele was a smoke screen. Neither IBM nor Freescale were as stuck in molasses on that front as we were led to believe. Like I said – Macintel is about the video, not the performance.

  27. (MDN made me wait to post this…)

    “[…] the fact that Apple isn’t comparing them with Intel PCs anymore isn’t because the Quad G5s are getting smoked. It’s because the Intel machines aren’t looking too good in comparison.”

    Well, the obvious question is, if the Intel machines aren’t looking good in comparison, why isn’t Apple promoting this? Is it because Intel and Apple are friends now? Is it because Apple doesn’t want the Intel machines to look bad because Apple has publicly said that they are moving to Intel?

    I agree, though, that the PowerMacs will be the last machines to go to Intel.

    But I don’t agree with the conspiracy theory as to why Apple did it. I agree that it was about money, but not in the way you think.

    Essentially, Apple was paying for PowerPC development. IBM had the brains and the fabs and what IBM wanted Apple to do is pay for the development of the chips that Apple wanted. If IBM felt they could use them, too, they’d give Apple a price break. IBM wanted the same model that they had with Sony and Microsoft.

    Apple decided it would be cheaper to switch to Intel than to pay money to IBM to stay ahead in the race. Without Apple paying money and pushing IBM/Freescale for faster chips, the PowerPC will stagnate on the desktop, Intel will surpass them, and Apple will jump aboard with Intel. Remember that Apple doesn’t have to pay Intel to design faster CPUs–there’s already a huge market for faster x86 chips.

    There’s no DRM conspiracy, it’s just plain economics.

  28. Odyssey67 you are short-terming your outlook. Yes, today the G5 is a much faster product. And I honestly don’t know why a FreeScale dual-core was not put in the new Powerbooks (highly disappointing).

    But Jobs was talking road map. Road map refers to chips that aren’t in production yet. Jobs obviously was sold by Intel on something that is still in the design stage.

    Whether Jobs was sold a bill of goods or whether Intel will come through remains to be seen.

    Fact, IBM took months longer to come through on the G5 than Jobs hoped. Fact, IBM still hasn’t come up with the 3 GhZ G5 (even though M$ hopes to start shipping them in Xbox 360s in six weeks). Fact, IBM still doesn’t have a G5 chip that will work in a PowerBook (that heat sink on the Quad is massive).

    Do you think a 4″ thick laptop is a reasonable choice?

    Bottom line, if running Intel processors then the argument come strictly down to one point “it’s the OS.”

  29. Peter says: “Well, … if … Intel [is]n’t looking good in comparison, why isn’t Apple promoting this? Is it because Intel and Apple are friends now? Is it because Apple doesn’t want the Intel machines to look bad because Apple has publicly said that they are moving to Intel?”

    You asked this before – but the last question you ask here is the answer. Apple won’t trash Intel anymore. Being on the cusp of transitioning to their product, that would be stupid. Once you’ve made the decision of going to Intel in the first place, you won’t help yourself sell computers in the future by publically making yourself look dumb for doing so.

    “I don’t agree with the conspiracy theory as to why Apple did it. I agree that it was about money, but not in the way you think. Essentially, Apple was paying for PowerPC development. IBM had the brains and the fabs and … wanted Apple to … pay for the development of the chips … Apple wanted. If IBM felt they could use them, too, they’d give Apple a price break. IBM wanted the same model that they had with Sony and Microsoft. Apple decided it would be cheaper to switch to Intel than to pay money to IBM to stay ahead in the race. “

    Generally, this is how IBM conducts business – they are CPU designers/manufacturers for hire. And this was fine for Apple when they commissioned them to make the G5. Suddenly, it wasn’t. Why? Because they resented not having a 3Ghz CPU? Everyone in the industry was having similar problems, so Apple didn’t lose anything there. Is Apple more hard up for cash now than they were in 2002? No – in fact they have even more in the bank ($6-7 billion), and were recently willing to give Samsung a huge chunk of it to help them make NAND flash memory. Cost per CPU? IBM has some of the highest yields per silicon wafer in the business (thanks to the small size of their CPUs), and many knowlegable sites have said that Apple’s terms with IBM were extremely favorable (BTW – THIS is the advantage of paying your development costs upfront; the CPU supplier doesn’t have to recoup their own investment by charging more per CPU). So the costs are probably a wash when compared to whatever Intel is charging. In short, the argument that the ‘financials’ of being with IBM were egregious, or more favorable with Intel, just fall short.

    As for IBM insisting Apple use a common Sony/M$ design; not possible – those designs are way too different. Xenon is a 3 core PPC that has fewer integer registers than normal, and Cell … is Cell. There’s nothing else like it. Nintendo has a more standard, faster clocked PPC, so if there was anyone IBM would have ‘insisted’ on Apple pairing with it would’ve been them. But the fact is, IBM is making all of the above, in addition to the G5s (dual and low power single core versions) and their own Power5! Clearly, IBM will make anything that anybody wants.

    “Without Apple paying money and pushing IBM/Freescale for faster chips, the PowerPC will stagnate on the desktop, Intel will surpass them … Remember that Apple doesn’t have to pay Intel to design faster CPUs–there’s already a huge market for faster x86 chips. There’s no DRM conspiracy, it’s just plain economics.”

    No conspiracy, eh … what do you call the MPAA? Their incessant lobbying in Washington? The legal actions against private citizens? And if you seriously think the Hollywood studios aren’t pushing for hardware DRM, then you my friend have been asleep. Don’t mean to be harsh, but you should have heard about all this by now.

    You might be right about Freescale’s offerings for the desktop stagnating, with their relationship with Apple completely severed. But IBM is now dealing with a company called Mercury (I believe) that intends to use Cell for servers and workstations. IBM is obviously going to keep making Power5 based servers and supercomputer clusters. So I think they will be putting out some good stuff for the forseeable future.

    Would it have been better with Apple in the fold? Undoubtably, but that cuts both ways. IBM may not push things as far in the future as they might have in partnership with Apple, but then Apple’s decision gaurantees THEY won’t be on the receiving end of that advancement either. And since Intel has already rebuffed Apple’s requests for preferential delivery of their next gen CPUs, the likelihood they will have much influence on Intel’s future plans is small – just like their marketshare, which is ultimately how Intel ranks their customers.

    That’s the main reason I regret their decision to go ‘generic’ – I think it hurts both Apple and IBM, and by extension all of us, by not continuing what was one of the few partnerships that actually derserved the ‘synergy’ moniker. The only ones who really make out in the end now are Intel and the movie studios.

    Oh, and as for economics, let me reiterate; Apple pays for faster x86 chips the same way all of Intel’s customers do – via the price of the chip.

  30. what about X-Serves? I bet a dual core in one of these puppies will fly!

    the dual g5 fan sounded like a 747 when it kicked in to high, what to these sound like, a Saturn V rocket taking off??? i bet the whole building will shake ” width=”19″ height=”19″ alt=”smile” style=”border:0;” />

    will I be able to buy these at the iPod store??? (a good day today for Apple COMPUTER..they really still do make and sell computers!)

  31. TommyBoy says: “you are short-terming your outlook. Yes, today the G5 is a much faster product. And I honestly don’t know why a FreeScale dual-core was not put in the new Powerbooks (highly disappointing). But Jobs was talking road map. Road map refers to chips that aren’t in production yet. Jobs obviously was sold by Intel on something that is still in the design stage.”

    Actually I’m longer-terming it. Long term I see nothing – especially from Intel – that comes close to being as revolutionary as the PPC based Cell. Short term, it’s usefulness is marginal (gaming), but only because A] it’s a ‘revision a’ design right now, and B] no one has quite figured out what to do with it. Nevertheless, people are working on both problems as we speak, and no other CPU has it’s capabilities.

    Powerbook didn’t get the dual G4 because it would have made a relatively ancient CPU design look real good … too good in comparison to what is being billed as an improvement in the Macintel. I’ll also give a nod to the fact that it would have cost Apple more money and time than they wanted to spend in upgrading the MoBos FSB and so forth, since Macintel will be coming first in laptops in just under a year. Still, if Macintel is such an improvement, they could have migrated the architecture down to the iBook line and fully amortized the costs. And they would have, if dual G4s were clearly inferior to dual x86s. They didn’t do that, and I think this shows Apple didn’t want any direct comparisons between them. So much for advancement!

    As for roadmaps – To the extent roadmaps are reliable, there is nothing Intel is claiming to do that AMD isn’t already implimenting, or on par with bringing to market. In fact, in some important ways (like on die memory controllers) AMD’s roadmap makes Intel’s look weak. IBM is doing about the same stuff as the both of them, but IBM will build whatever you want, so a roadmap is less relevant for them. Freescale will now be making some of the most powerful cable TV boxes and automobile ‘engine management systems’ that the world could ever have expected, but …

    “Whether Jobs was sold a bill of goods or whether Intel will come through remains to be seen.”

    Jobs was not sold the bill of goods. WE were sold the bill of goods. Jobs knows exactly what he is doing, and exactly why he is doing it. But he wasn’t going to come out and say to the ‘fanbase’, “Hey, we’re moving to an inferior architecture, and to the second best version of it, just so that I can aquiese to the wishes of the Hollywood movie companies, who want to shackle you all to their way of doing things, and in return I get to continue building a media distribution/creation empire.” Jobs himself may even be rationalizing how this will ultimately be a good thing, since a financially stronger Apple SHOULD be able to do some cool things. He might be right, but he’s getting in bed with some lousy partners to do it. And, frankly, ends rarely justify means.

    “Fact, IBM took months longer to come through on the G5 than Jobs hoped. Fact, IBM still hasn’t come up with the 3 GhZ G5 (even though M$ hopes to start shipping them in Xbox 360s in six weeks). Fact, IBM still doesn’t have a G5 chip that will work in a PowerBook (that heat sink on the Quad is massive). Do you think a 4″ thick laptop is a reasonable choice?”

    Fact – IBM had exactly the same problems ramping up clockspeeds that the whole industry had – ESPECIALLY Intel (how quickly everyone forgets their promises of 4-5Ghz P4s). AMD hasn’t hit 3gigs, yet no one is coming down on them as they spank Intel’s ass in every performance measure that matters. These new dual core G5s are going to do the same thing. The XBox and Cell are more ‘clean sheet’ designs, with lessons learned from what went wrong with G5, so of course they’re going to be faster & run cooler. The G5 is still a better general purpose CPU then either of them (though Cell has room to improve), and Power5 is better still. In the end, if performance was really their main concern, Apple had many more options with PPC and IBM then Jobs admitted to. Ask yourself why.

    As for the G5 Powerbook – at the same time IBM announced the dual core 970 last spring, it announced a low power single core version of the same (25watts max). A single core G5 PB probably would be here by now, IF Apple hadn’t dumped IBM a month before these CPUs showed up. And, I would suppose, that if Apple had stayed as a customer, a dual core version of the low power G5 would probably be coming out of the Fishkill factory by the beginning of next year. No more big cooling rigs in PowerMac.

    “Bottom line, if running Intel processors then the argument come strictly down to one point “it’s the OS.””

    I think I know what you meant here – and we all had better hope so.


  32. Actually I’m longer-terming it. Long term I see nothing – especially from Intel – that comes close to being as revolutionary as the PPC based Cell. Short term, it’s usefulness is marginal (gaming), but only because A] it’s a ‘revision a’ design right now, and B] no one has quite figured out what to do with it. Nevertheless, people are working on both problems as we speak, and no other CPU has it’s capabilities.

    You’re one of the few remaining people still sucked up into the Cell hype propagated by Sony’s initial unveiling, probably because you haven’t really kept up with recent events concerning this “revolutionary” chip.

    Cell is extremely difficult to program for, and ever since developmental kits were sent out developers have been complaining because of this very complexity….it turns out taking advantage of Cell’s vaunted hypothetical processing power can only occur in very limited situation that game developers want (FEAR’s physics caused a huge slowdown on the SPE’s, and the game actually ran faster with slightly cut down physics running on the PPE), leaving game code to depend on the PPE’s absymally slow speeds the vast majority of the time.

    From demo’s seen so far, and from individual developer interviews, most consoles magazines agree Sony has fallen short of the hype as the most visually complex and advanced game engines will be showcased on the Xbox360, not PS3. Developers do not like programming for 7 individual SPE’s that cannot even perform the tasks they want, and that’s one of the reason why Sony is having a tough time keeping big name developers like Electronic Arts, and surprisingly Konami, onboard.

  33. “Generally, this is how IBM conducts business – they are CPU designers/manufacturers for hire. And this was fine for Apple when they commissioned them to make the G5. Suddenly, it wasn’t. Why?”

    Very simple–because Apple had given IBM a hunk of money to develop the roadmap. IBM failed to do so. Furthermore, IBM was being wooed by Microsoft and Sony and was not paying attention to Apple’s concerns, since IBM was seeing tons of dollar signs from Playstation and Xbox sales. Again, the amount of money that IBM will make of those two devices makes PowerMac G5 sales look like a drop in the bucket.

    IBM failed. IBM sold Apple a bill of goods and didn’t deliver. What Intel did was buckle down and go in a different direction. What IBM did was go look for other customers to sell their current wares and not devote the resources to solving Apple’s problems, which is what Apple had paid them to do. IBM did try to interest Apple in the cell chip as sort of a consolation prize–“No, we didn’t accomplish what we said we’d do, but can we interest you in this?”

    Again, why should Apple throw good money after bad?

    “No conspiracy, eh … what do you call the MPAA? Their incessant lobbying in Washington? The legal actions against private citizens? And if you seriously think the Hollywood studios aren’t pushing for hardware DRM, then you my friend have been asleep. Don’t mean to be harsh, but you should have heard about all this by now.”

    Oh, I agree that the Hollywood Studios are pushing for hardware DRM. And, if they are successful, Apple will have to deal with this. Yet, why does that require Apple to go with Intel? Gee, USB–an Intel technology–doesn’t require an Intel CPU. I’m sure that Apple could include Intel’s DRM chip, if that become the standard, on a PowerPC motherboard.

    Also, maybe I am out of the loop, but I thought it was the RIAA that is pursuing legal actions against private citizens. And, personally, I’m all for it. You solve people problems with people. You solve technology with technology. Trying to use technology to solve people problems doesn’t work, as we’ve seen time-and-time again.

    Don’t get me wrong–I, too, believe it was a bad decision on Apple’s part. While Mac OS X’s superiority is unquestionable, it will ultimately come down to some hard numbers. And it will be much more difficult for Apple to compete in the professional arena with essentially the same hardware as Dell. I can go to the purchasing people and show them benchmarks saying the Mac is faster than a PC when running Photoshop. When a Mac and a PC are the same thing, it’ll make it a bit harder for me to say, “I need a Mac to run Photoshop.”

    Again, I think Apple looked at the stuff they wanted to build. They didn’t have much faith in IBM or Freescale’s ability to get them there before Intel and AMD offered PC builders the same thing. Whether it will work out is anyone’s guess–it’s a gamble, no question. But, as they say, the definition of insanity is doing the same thing over and over and expecting different results. Why is Apple to believe that IBM will be able to produce what Apple needs on time and on target? Why is Apple to believe that Freescale will produce what Apple needs on time and on target. Intel may not either, but based on past performance, IBM and Freescale aren’t the most trustworthy partners.

    I like a good conspiracy theory as much as the next person. But I think the reality is far more mundane. Trust me, while I was at WWDC this year, the place was totally abuzz with all sorts of great conspiracy theories. But Occam’s Razor works–the simplest solution is probably the correct one.

  34. Sammy says: “You’re one of the few remaining people still sucked up into the Cell hype propagated by Sony’s initial unveiling, probably because you haven’t really kept up with … events concerning this “revolutionary” chip. Cell is extremely difficult to program for, and … developers have been complaining because of this … taking advantage of Cell’s vaunted … processing power can only occur in very limited situation … (FEAR’s physics caused a huge slowdown on the SPE’s, and the game actually ran faster with slightly cut down physics running on the PPE), leaving game code to depend on the PPE’s absymally slow speeds the vast majority of the time.”

    I’ve kept up on recent events, and you’re right about the developers, and about Cell’s current limitations. However, I’ve also been looking into how a Cell is made, what modifications could be implimented as a result, and whether anyone else beyond the game developer community has taken up the task of using Cell in other ways.

    I don’t believe the current ‘rev a’ version is to be the last word on Cell. It’s a fairly modular design, with a PPC mother chip surrounded by (in PS3) 7 SPEs. The die size is large, but fixed, and IBM/Sony/Toshiba decided to maximize the parallel computing capabilities of Cell (for the purposes of multimedia) by using the real estate for SPEs, and skimping on the mother chip’s transister count devoted to cache size and integer registers. The latter are crucial for a general purpose CPU, and thus the current Cell – without some fancy programming – falls flat in that role. Now, there is a company (Mercury) that is working on overcoming those limitations, developing admittedly ‘higher math’ oriented servers and workstations by leveraging Cells massively parallel design. IBM is working closely with them. If they find success, it’s expected that IBM will tweak Cell further for them in any way they can to make it more suited for their purposes, even if that means making Cell more ‘GP’ oriented.

    Is that possible? Yes – thanks to Cell’s modular design. Essentially, for the sacrifice of an SPE or 3, you can pump up the mother chip in ways that make it more GP capable. More cache, standard integer register allotment, even dual cores are all possible. At that point Cell can be coded for as easily as any PPC based CPU, and yet would still have all the parallel computing features and a few SPEs with which to expand its capabilities. With future lithography improvements (i.e. going from 90nm to 65nm to 45nm) this scenario is made even easier … provided IBM has a partner willing to do something with such a CPU.

    I think IBM simply assumed Apple would be going down that road with them. Maybe this seemed such a no brainer that they temporarily stopped stroking Jobs in other ways. I don’t know. But what I do know is that Cell’s overall features are – despite the current version’s limitations – the most advanced thing out there in CPU design.

    “From demo’s seen so far, and … developer interviews, most … agree Sony has fallen short of the hype … the most … advanced game engines will be showcased on the Xbox360 … Developers do not like programming for 7 individual SPE’s that cannot even perform the tasks they want, and that’s one of the reason why Sony is having a tough time keeping big name developers like Electronic Arts, and … Konami, onboard.”

    Yes, this could definitely be a better situation for Sony. But remember that the Emotion Engine CPU of PS2 was a bitch at first too. Console economics being what it is, all companies try to come up with an architecture that may be hopelessly outside the mainstream now, for the purpose of having years and years where the devs can tease out ever better performance as they get used to it, WITHOUT actually having to upgrade anything. Sony went more overboard than M$ or Nintendo on this, and is paying the price, but I think they saw the future of Cell being bright enough to weather these kinds of problems in the present. Also, even the 360 isn’t being lauded for it’s easy game development (it has many of the same integer and cache problems as Cell), especially when it comes to making legacy x86 based games run on it. Nintendo’s CPU – basically a hopped up (3+Ghz) 970 – is the most mainstream of the three (and probably the one IBM thought would hold Apple over while everyone figured out how best to proceed with Cell).

    Regardless – I realize we could debate all day on this stuff – I’ve studied the topic as much as the next guy, and I still see PPC, and specifically Cell in the long term, as a better future to base CPU architecture on.

  35. Peter says: “Very simple–because Apple had given IBM a hunk of money to develop the roadmap. IBM failed to do so. Furthermore, IBM was being wooed by Microsoft and Sony and was not paying attention to Apple’s concerns, since IBM was seeing tons of dollar signs from Playstation and Xbox sales. Again, the amount of money that IBM will make of those two devices makes PowerMac G5 sales look like a drop in the bucket. IBM failed. IBM sold Apple a bill of goods and didn’t deliver.”

    Ah yes – the ‘poor neglected Apple’ theory. Look, considering Apple has had low power and dual core versions of both the G4 and G5 available to it since spring, I don’t buy this as legitimate. It works if you really want to believe His Steveness always says the truth whenever he speaks, but (>cough< no video iPod >cough<) I think that’s naive.

    “What Intel did was buckle down and go in a different direction.”

    Right – every one of their current hot ‘new’ CPU offerings for the next 3 years is based on PIII. Their different direction is commonly known as “backward”.

    What IBM did was go look for other customers to sell their current wares and not devote the resources to solving Apple’s problems, which is what Apple had paid them to do. IBM did try to interest Apple in the cell chip as sort of a consolation prize–“No, we didn’t accomplish what we said we’d do, but can we interest you in this?” Again, why should Apple throw good money after bad?”

    Any purchaser of AMD and Intel CPUs is subsidizing development costs – it’s factored into the price of their chips. IBM goes in a different direction (sorry); they say ‘pay us upfront and we’ll develop whatever you want’. Did G5 meet everyone’s expectations? No – it didn’t hit 3Ghz in a year, and it was hot. Newsflash: Intel had the very same kinds of problems with the Netburst version of P4. They went ‘back to the file cabinet’ with PIII to find a solution. Meanwhile, G5 has gotten over the same humps without having to abandon their basic design. So Apple wasn’t losing money – they may have lost patience, but not money. Does any of this mean IBM has been faultless in it’s partnership with Apple? No. Could they have worked harder at getting dual core and (especially) low power versions of G5 out sooner than this past Spring? Yes. But they are out there now, and considering that AMD and Intel are only recently getting a full head of steam up on the dual core front, and that PBs and iBooks are selling like hotcakes even with the current versions of G4 (which also has lower power and dual core versions available), I just don’t see Apple being as behind the 8Ball in these areas with PPC as they want us to believe.

    cont …

  36. Peter says: ” I agree that … Hollywood … are pushing for hardware DRM. And, if they are successful, Apple will have to deal with this. Yet, why does that require Apple to go with Intel? … I’m sure that Apple could include Intel’s DRM chip, if that become the standard, on a PowerPC motherboard.”

    From what I know – and I’ll be glad to be corrected if I’m wrong – Intel is offering their TPMs as a package deal with their own CPUs and approved MoBos only. Would they do otherwise? They have no need to gain marketshare – only to maintain it – and offering a feature that your competitors could potentially benefit from (thus take market share from you) would make no sense.

    Apple is already dealing with DRM – they back the BluRay standard in optical disks, which is the most restrictive of the two alternatives (no coinicidence there), and of course they have software based Fairplay going. But Fairplay is roundly critizied by these studios as being too lax, and too hackable. There’s a reason, beyond download times, that the new iTMS video offerings are offered only in less-than-standard definition quality. There’s simply no stomach for offering anything better, so long as hardware DRM is not a ubiquitous reality.

    We already know that Macintels will have TPM chips – ostensibly to keep OSX on Apple made machines. But mark my words – when these computers start coming online, you will then see iTMS video offerings really open up, as well as the introduction of a computer that takes what the new iMac introduces much further toward the ‘video server’ realm. But it will be a ‘Macintel Only’ feature, because only they will have hardware DRM of the type that the studios prefer, and that only Intel offers.

    “… maybe I am out of the loop, but I thought it was the RIAA that is pursuing legal actions against private citizens.”

    RIAA, MPAA, RNC, SPCA … whatever. You know what I mean.

    “And personally, I’m all for it. You solve people problems with people. You solve technology with technology. Trying to use technology to solve people problems doesn’t work, as we’ve seen time-and-time again.”

    It’s not that I don’t agree with your basic philosophy here – I just don’t think it encompasses all the problems. I don’t want anyone suing me, or any computer chip limiting me, for doing what I’m legally entitled to do. File sharing per se is not a harmful activity. Recording audio or video for my own use, even to loan to friends, is legal. I’m already tired of Big Brother. If Apple is getting in bed with the same forces that I think are a threat to me personally, as well as an impediment to technological development and the legal and beneficial dissemination of knowledge and art, then I think a skeptical eye cast their way is more than warranted. Maybe Apple will be the agent of change here, but if going to marginal technology in CPUs is the first step, I’m not encouraged.

    “Don’t get me wrong – I, too, believe it was a bad decision on Apple’s part. While Mac OS X’s superiority is unquestionable, it will ultimately come down to some hard numbers. And it will be much more difficult for Apple to compete … with essentially the same hardware … When a Mac and a PC are the same thing, it’ll make it a bit harder for me to say, “I need a Mac to run Photoshop.”

    Your example of direct comparisons with Dell and Windows being unavoidable with the Macintels is spot on. Apple better have their A-Game going from here on out, cause there’s no fudging the numbers anymore (Tiger’s lingering bugs are a little worrisome here as well). Although I must say, if you really believe Apple’s making a mistake in this regard … then why are we debating? We agree going with x86 could invoke more trouble than Apple is bargaining on. I just think their selling their collective souls to the Hollywood Media Gods is going to hasten that.

    “I like a good conspiracy theory as much as the next person. But I think the reality is far more mundane. Trust me, while I was at WWDC this year, the place was totally abuzz with all sorts of great conspiracy theories. But Occam’s Razor works–the simplest solution is probably the correct one.”

    Well, I see the mental gymnastics everyone is going through to help Jobs validate his ‘PPC couldn’t cut the mustard anymore’ claims as anything but simple. The dual core/dual CPU G5 PM is going to blow the doors off anything Intel has on the ground, and will for a while (which is why it will be the last to be ‘converted’). And IBM – despite their Jobs induced reputation – has not been known for sitting still on such things. Occam’s Razor would seem to lead to the conclusion that a massive Macintel transition for performance reasons is the least likely explanation for it. And as Sherlock Holmes famously said, “… when all other contingencies fail, whatever remains, however improbable, must be the truth.”

    Tin hat or not.
    ” width=”19″ height=”19″ alt=”wink” style=”border:0;” />

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.