I don't get it. Where did this rabbit pop out of the hat?:
http://www.theregister.com/2006/02/07/ibm_power6_show/
Why didn't Apple know about this being developed, or did they? Will the IntelMac end up being a flash in the pan? Sounds like the hardware of the future could change with the wind.
According to the linked article, the Power 6 is a year and a half away from production. That's a looong time to go without a speed bump in the G5s.
The article also implies the power consumption of the Power 6 will be kept at least similar to the G5 (ie - parity, or consuming only slightly mode) making it completely useless for laptops. Apple would have been up the creek for portables without the Intel switch.
There's also no mention of compatibility with the G5, which was derived from IBM's Power 4 architecture.
Doesn't sound to me like it'll make any difference to Apple's plans.
is this stuff PPC compatible?
is what the PowerPC is based on. Basically, what IBM does it take the POWERx (x denoting generation) scale it down a bit for use in desktop or embedded applications and usually, but not always with Motorola, markets it. The POWER chips usually find themselves inside something along the lines of an S/390 (they are now referred to as z/series) mainframe, AS/400 (don't remember what series they are now) database box, or RS/6000 (called P series now, IIRC) Unix box. The G5, or PowerPC 970, is a stripped down Power5. The Power5 chips from the start were dual and quad core. For cost and heat considerations, the Power Mac G5 used a scaled down single core chip.
IBM is always developing new technology. It's just what they do.
Will the the Power 6 become the next 68060? Something that is shrouded in awe and mystery down at the 68KMLA.
As IBM will be using this chip. If I recall correctly, the 68060 gained little to no traction at all in being used in any capacity. However, those of us on the consumer level will likely never see these products outside of a data center. The Power6 family will power IBM's bread and butter servers, database boxes, and mainframes, and likely a scaled down version will power Xbox 720.
Probably not. It's got a real market and purpose, though the 68060 had some, IBM isn't betting the arch on other vendors. They've got plenty of machines to stuff it in that they build. Even if they scaled it back a bit to be more of a workstation CPU they'd likely sell their own workstations with it. They've sold 604 based PPC RS/6000 workstations. I used to own a 7248 40P, IIRC. It ran AIX nicely.
I don't think anybody is going to stuff a POWER6 into a quadcore G5 without some major surgery, so it might get a bit of the '060 effect, but since the QC G5 is dang nice as-is...
It's becoming pretty clear from the various benchmarking frenzies that the Intel macs just don't have much if any on the G5, core-for-core. (Yes, a Core Duo is faster then a single G5, but the dual core models hold their own pretty well and the quad-core Dual-package model slaughters it.) It's also interesting how *after* Apple made its "We're switching, screw you IBM announcement" IBM publicized the existance of low-power G5 variants which are "in the ball park" of the Pentium M/Core Duo.
So why would Apple switch if PowerPC is at least comparable, and why would they seemingly willfully ignore improvements in the technology? (Another dirty little secret is that the G5 CPU is actually *cheaper* then the Intel replacement. Not by a little, but by a *lot*. Like 1/3rd the cost.) Obviously this has nothing to do with CPU performance. This is all about Apple exiting the computer design market. They just don't want to have to worry about making the nuts-and-bolts parts like motherboard chipsets anymore. IBM isn't interested in doing that work (which isn't cheap) for them and Intel is clearly willing to bend over backwards to make them happy. (And it doesn't cost them anything, really, since they're selling the same stuff to Dell, Sony, Toshiba, HP, etc, already.)
IBM makes "real computers". (Servers, mainframes, embedded products, military and aerospace processors, network backbones, etc.) They're not interested in making home computers or "toys", except where such products might overlap with one of their "real" markets. An interesting example of such overlap would be the "Cell" processor, which even before coming to market in the Playstation 3 is being prepped for use in network servers and embedded applications:
http://www.theinquirer.net/?article=29575
Intel's primary focus on the other hand is providing parts for "small computer" (commodity desktop, laptop, and small server) manufacturers It's a huge part of their business, and they spend a *lot* of time and money supporting their customers with technical, design, and marketing support. Apple is basically a home computer and entertainment product manufacturer. Which company does it make more sense for them to work with? Seems pretty obvious, really.
Working with Intel lets them buy pre-debugged hardware designs off the shelf, stuff them in a pretty case, and slap an Apple logo on it. They don't need to worry about all that other hardware minutia it takes to build a high-performance computer anymore. All they need to do is write software and do marketing. There's always been speculation about Apple turning into a "software company", and discontinuing their hardware products. This Intel agreement lets them do that. Yes, their software still only runs on machines with Apple logos on them, but Apple doesn't have to design those systems anymore. So, yes, after 2007 they *will* just be a software company. The extra money they're paying Intel for design support is probably more then made up by the fact that soon they'll be able to fire their entire computer hardware engineering division. And it also makes it easier at some point in the future to drop or spin off the software division entirely, should making Macs cease being a profitable business. Keeping the option open to switch back to PowerPC would be *way* too costly, so I wouldn't hold your breath.
Anyway. That's why you'll never see a Power6 Mac, even if the CPU debuts at 10 Ghz, runs a hundred times faster then the "Intel Core Octo", and draws half-a-watt of power. Apple won't have the engineers to make home computers out of it. C'est la vie.
--Peace
Also I think Apple was thinking of iPods or whatever and not just macs when they switched to Intel.
All that low watt ppc chip supply problems hype was just BS in my opinion.
I know several ppl who uses the New Intel Mac's and they love the speed they get with x86 coded apps. It is a litle slower than a g5 on rosetta appps but it is doing emulation.
I don't think CPU speed or power consumption are the end of the story here, either.
Apple's primary motivation is to make money. I'm not passing judgement here, just stating fact.
Let's look back at the early-to-mid 1990's: Apple's in the middle of a switch from 68k to PowerPC. Why did they do this? We were told that it would mean cheaper Macs and better performance. RISC was all that and a bag of chips. For the next 5 years, Apple bled money and lost users left-right-and-centre. It did, however, force the Mac faithful who had a 68k Mac to buy a new one after the release of OS 8.5, if they wished to remain current.
Fast-forward to today. OS X runs on practically any New-World machine. Albeit, you need XPostFacto to run Tiger on a non-firewire equipped Mac, but the possibility does exist. While far (FAR) from a speed-demon, a Beige G3 can run Panther or Tiger reasonably well for light usage. So, while you're shelling out $150 to Apple for a new version of OS X, the real money lies in the hardware, and Apple knows this. So, every decade or so, they switch hardware platforms which (eventually) forces people to upgrade. OS X will continue to provide PPC support probably into 10.6, or *maybe* 10.7. It's really hard to say at this point, but I would wager that OS X will continue to be produced for PPC's for another 2 years or so.
I'm basing this on the 68k/PPC switch: The first PPC's came with OS 7.1, the last 68k support was with OS 8.1. Mind you, Apple continued marketing 68k machines for 2 years after the PPC switch, which would mean that 68k support continued for ~2 years after the last 68k machines were discontinued. Based on that, and Apple's claim that the entire lineup will be Intel by mid 2006, we can expect PPC support into 2008. This all depends on how Steve wants to proceed with this, and personally, I'm beginning to hoard PowerPC applications (Commercial, Free/Shareware, etc.) and OS updates for the day that they're no longer available.
However, this isn't switching from one obscure platform to another, this is going from a relatively closed archetecture to the most-used Personal Computer platform on the planet. Apple's gonna lose their grip on the Mac OS, as it becomes increasingly easier to install it on generic PC's. Apple's really playing with fire here, let's hope that Steve doesn't *completely* burn the shack to the ground.
Cheers,
The Czar
First we will loose the G3's in 10.5. nobody will say anything at the time, except for G3 owners. Everybody else will say "don't you think it's time to upgrade?".
Then, in 10.6, everything pre quicksilver will be cut out. Repeat above.
10.7 ~ 10.8 will be G5 only. Sorry, MDD Dual owners. Your computer is now obsolete.
And the last nail in the coffin for PPC will be ... OS 11, or whatever it will be called.
Why do I say this? History.
OS 7.6 was 32 bit clean only.
OS 8 = Quadra only.
OS 8.5 = PPC only. So sorry, mr. owner of a 3 year old performa/lc. time to upgrade.
OS X = G3 only, and too bad for first gen PB G3 owners. Aparently a ~$6,000 investment wasn't enough to get you into OS X 3 years later.
I have what I expect to be the last PPC rev of the 12" powerbook, and given that it's a 'pro' level machine, I wouldn't expect support for it to be cut any earlier than complete PPC support.
That being said, I expect 10.5 to be PPC compatible, and 10.6 to be Intel only. I don't see a point in stretching the transition to more than that.
Frankly, I see support (current versions of OS X) for PowerPC macs to continue until 2009 or 2010. Granted, the support for new features on OS releases wil vary based on the machine. Given the relatively limited installed base of G5 users (only iMac and Power Mac), it's doubtful that Apple would go that hardcore to alienate buyers of G4 class machines. Also, considering how similar in architecture the bulk of the G4 towers is, if Apple kills support for one, they kill it for all. If anything, I see the blue and white G3/PCI graphics G4 and the Pismo to be the machines cut from the supported hardware list come 10.5. The slot load iMacs and dual usb iBooks are probably going to be the lowest end hardware supported. (I can easily see a 500 mhz or better system requirement) As for 10.6, who knows? We know precious little about 10.5 already!
It's all speculation at this point, but my personal suspicion is that Apple's target date for seriously losing interest in PowerPC is year-end 2009, as a number of bits of evidence point to them considering a three-year obsolesence cycle acceptable.
(As a "for instance", it's notable that the G3 iBook motherboard replacement program was only good for three years after the purchase date of the laptop. Clearly they have no interest in someone running a piece of their hardware for longer then that.)
As for "cut-off" dates, my crystal ball tells me that if Apple were to disenfranchise another large group of machines the only logical cut-off would be machines without Airport Extreme. ("UNI-N" based machines.) Discontinuing G4 support isn't really doable because they've sold G4 Powerbooks and iBooks right up to the switchover, but it *might* be fair game to cutoff Pismos, Titaniums, G3 iBooks and iMacs all in one fell swoop. If they did that it probably *won't* be in 10.5, but it could *possibly* be 10.6 if there is a PowerPC version. The interesting thing to ponder about such a cutoff would be if G4 towers would be excepted, because *all* except the very last models designed were UNI-N based. They excepted the B&W G3 and "Yikes" G4 from the "Paddington" cutoff in 10.4 after all. (The cull that claimed the original iMac, iBooks, and Lombard.)
(Cutting off the B&W and Yikes honestly would of been a great idea, but I know that politically it would of been unacceptable to kill a G4 while they were still selling them. They'll have no such qualms soon.)
In any case, the real question is whether anyone keeps producing applications for the older machines. Developers, at least the ones that make it through the transition, will probably want to forget about debugging endian issues as soon as possible.
--Peace
Yes, but those of us who are bloody determined should be able to find a way to install 10.6, and maybe even 10.7/8 on our PowerPC rigs through XPostFacto (like BornAgain to run OS 8.1 on an 030). It won't be until PowerPC support is *completely* cut from OS X that we'll be *forced* to buy an Macintel.
Reading through some of the earlier posts, something popped into my mind. Apple's going down the road HP paved with red-ink and layoffs. Instead of being an innovator in the PC marketplace, they're now simply a commodity PC producer (and a lack-lustre one at that). I wonder if Apple is not headed for the same fate?
Cheers,
The Czar
Hehe... Not quite correct. I've got OS X 10.2 running on my Power Mac 6500, which uses a 250mhz PowerPC 603e CPU.
nice point... dunno how I missed this thread for so long... thanks for keeping it alive.
OS X 10.3 is a mature OS in its own right. UNIX itself hasn't changed all that much at the user level in the last 30 years. What can Apple possibly come up with to make 10.6 oh so entirely insanely better than 10.3... compared to how MacOS Sys 9 is so much better than MacOS Sys 6? It's going to need a 3 dimensional holographic display and a direct to brain interface or something...
How long did Video Toasters keep on working at a pro level after the orig. Commodore (the Co.) was long gone? Considering that OS X has solid UNIX underpinnings, I think we're entering a whole new frontier where old computers keep on chugging for maybe ... 20 years, and who knows? If the next OS doesn't run on your mac, is that really the push off the cliff to buy new hardware? When 99.9% of the software still runs on the 'old' OS? AFAIK, the only apps that require 10.4 and G5's are Apple's pro apps, and successes have been made in hacks to allow even them to run on the G4 (Aperture I know has been hacked to run on slow G4s, and apparently, it runs just fine).
Sometimes... the advance of technology boggles my mind. Think about the speed of the computers that OPENSTEP ran on... 25-75 MHz! I don't remember thinking those machines were slow. So... now the processors are running... I don't know, 2000% faster... and ... I don't really see the difference. As the hardware gets faster, the OS and apps get bigger and need that superfast processor just to meet the status quo of user experience. I realize that OPENSTEP isn't quite OS X... but what would it feel like to be running OPENSTEP proper on a 2.5GHz processor (when it runs nice and spiffy on a 100MHz processor).
Anyway... Apple can shift directions, mid-stream, at the drop of a hat, whatever, and not lose ground. They've shown they can do this, and if it makes sense to do so, they will again.