Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
PowerBook Updates? Dual G4 Core?
- Thread starter MacRumors
- Start date
- Sort by reaction score
cmoney said:
The unbelieveably idiotic, dumbass part of the techworld maybe.
Whereas the rest of the techworld would be rejoicing and patting apple on the back.
jadam said:
No, Apple can't just declare a G4 a G5 without serious repercussions. The G5 stands for, among other things, 64-bit capability. There is no way that Apple would make a G5 that didn't offer some kind of 64-bit functionality similar to the 970FX. Let's take the tagline from Apple's own site: "PowerPC G5. Introducing the world's first 64-bit desktop processor." While not exactly a true statement, it does show you what they chose to emphasize above all else. 64-bit.
Clicking on the first technology page, the term "64-bit" appears six times on the first page. Apple's individual pages aren't very long, so six times is pretty significant. The other technological features on that intro page are the fast bus speed and the "Award-winning Logic" of IBM's POWER line. Apple would have to change that entire approach to the G5 and face the wrath of IT professionals, experts, and avid users worldwide. Even if you could argue that Apple would happily disregard its users, their business customers and marketing department absolutely would not stand for it.
Given that the 64-bit version of the G4 does not exist, the discussion pertains to the existing e600-derived models and other Freescale products, all of which are NOT 64-bit and will NOT be called G5 if used by Apple.
That would stink !
You mean they might try and water down a true G5 and call it like a G5M but at its core its a G4. That would stink!
Caiwyn said:
You mean they might try and water down a true G5 and call it like a G5M but at its core its a G4. That would stink!
Rather than rambling on about Dual Cores and G5, my 2c says that the closest possibility is still the single core 7448 chip with it's 200 Mhz bus which is pin compatible with the 7447. Last update on Freescale's page is from September 2004 so who knows now what stage sampling/production could be. Only hope we'll not see only minor speed bumps owise there's going to be quite some disappointment around here. 
billystlyes said:
Not water down a true G5, but take a dual-core G4 and call it a G5M. And yes, it would stink. I don't expect Apple to do it, but it's been two years since the introduction of the G5 processor, and Apple needs the marketing advantages that the G5 name brings. They won't call it a G5, but there's always the possibility that they'll give it a new, similar name, like a G5.
That said, there's really nothing wrong with the G4 design. Even in Tiger, 64-bit support in OS X is minimal at best, and most apps won't benefit from it. And clock-for-clock, the G4 is generally as good as a G5. The G5 can hit higher clock speeds -- but that advantage is probably going to be lost when fitting it into a notebook. So a dual-core G4 would certainly be nice. I just don't believe it's going to happen.
Dalriada said:Rather than rambling on about Dual Cores and G5, my 2c says that the closest possibility is still the single core 7448 chip with it's 200 Mhz bus which is pin compatible with the 7447. Last update on Freescale's page is from September 2004 so who knows now what stage sampling/production could be. Only hope we'll not see only minor speed bumps owise there's going to be quite some disappointment around here.
It would be more than a little sad to see the MPC7448 dropped into a PowerBook, though. They probably would not invest in the required components to make use of that FSB and instead we'd just get a marginally faster G4 with the same old FSB. I hope Apple bought controllers, proper PCB, and other components for the upgrade to use the new 200MHz bus.
Secretly, though, I really hope that the new 7448 is meant for the iBook, while the PB gets something a little more potent...
Because it wasn't designed for laptop. But that's irrelevant.dongmin said:
It's enough for me that they've been wrong before. I'm not saying I think it's going to be Dual G4s on Tuesday, just that your arguments against a D-G4 don't amount to anything particularly exciting.dongmin said:
Everything I've heard about this chip so far makes it absolutely perfect for a laptop. Low power consumption, higher performance, integrated memory controler, pcie controller, and gigabit ethernet. Someone posted before me that the dual core G4 is similar in power consumption to the single core due to the smaller process (90nm). This is very believable.dongmin said:
But the high integration means fewer chips, which gives an even greater power savings. Every chip on the board has a certain ambient power consumption due to leakage even when it isn't being used. Just combining 2 chips into one saves power. We're talking about combining Ethernet, Cpu, and bridge/ram controller; that's three chips.
And just what exactly is wrong with having 4 ethernet controllers? Of course you only need one, so disable the other three. It's a major waste of money to redesign a chip just to get rid of three ethernet controllers when you can just not use them. In fact I'd place a handsome wager than you can completely disable all three in firmware and eliminate what little extra power they consume when not in use.
While it costs nothing to turn them off, it would likely cost well over $1M to make a separate chip. Making the masks for a new chip costs a ton of money, but stamping out chips afterward is dirt cheap. It's far more cost efficient to combine the volume from both markets (laptop and embedded). You don't pay for the extra mask fee, and you get higher volume on the one you already made which lowers the price per part.
This is really the only argument that holds water in my opinion. I personally don't know much more about the Dual core G4 than I've read from these forums, so I don't have any feel for credibility of the publicly announced sampling/shipping period for these parts.dongmin said:
So while I think this is a strong point, consider this line of reasoning: isn't freescale an ex-division of motorola that had trouble with delayed production of the G4? They may have lied about how soon the parts will be ready to make sure they delivered on time or earlier, in an attempt to reestablish credibility. Perhaps they want to announce availability the same day apple ships them in their laptops (or more likely, Apple wants them to wait).
It's only a theory, but it's no less credible than "new powerbooks next tuesday".
-kev
Dalriada said:Rather than rambling on about Dual Cores and G5, my 2c says that the closest possibility is still the single core 7448 chip with it's 200 Mhz bus which is pin compatible with the 7447. Last update on Freescale's page is from September 2004 so who knows now what stage sampling/production could be. Only hope we'll not see only minor speed bumps owise there's going to be quite some disappointment around here.
The single-core 7448 is the most likely candidate for the next round of updates. Unlike the dual-core G4, the single-core 7448 is supposed to be available right about now. Of course, if that's what Apple is going to put into the powerbook, then all we'll see is a speedbump to 1.67 Ghz. The 1 MB of L2 cache should offer a significant improvement in performance, though. As for the 200 Mhz bus... Apple would have to redesign the current powerbooks for that to be effective. And if they're using the 7448 as a drop-in replacement, they may not go to that trouble. But who knows?
panphage said:
Jesus... I leave you people alone for a month or two so that I can enjoy time with my girlfriend and I come back to this.
Yes, the MPC8641D and MPC8641 aren't pin compatible with the previous generation of MPC7xxx (also known to Apple customers as the G4), but neither is the 970 or its derivatives. No matter what way they move with a new processor, there will have to be a new motherboard design in order to adopt the chip. However, unlike the 970, the design for an MPC8641D implementation could be vastly simpler and require far fewer components on the board, since the PCI-e bridge, 4 gigabit NICs with hardware encryption acceleration, a DDR2 667mhz memory controller for each processor, 1MB L2 cache per core with sharing to make it effecively 2MB of cache, and RapidIO serial bus. That right there eliminates nearly all of the functions of an independent system controller (like the 970 PowerMac's U3 and the iMac's U3-Lite) and takes out an additional source of heat. Factor in the removal of NIC hardware and you're adding on the savings, especially when you consider the power saving inherent in a move to DDR2. Oh, and all of that goes onto a chip running at the same - or less - heat as a G5 without any of the SoC features in the dual-core Freescale design.
Also, what in the world makes you think that Apple doesn't have a few of these now? Freescale can make limited runs of engineering samples and Apple has always been one of their big customers, especially in a market segment that something like the e600 line seems to be aimed at.
dguisinger said:
I can name several reasons, actually.
1) Battery life and cooling - The 2" thick computer that you're referring to is constantly hooked into wall current for power supply and therefore needs not conserve energy for anything but thermal concerns. It also has three blowers that shove air through the relatively open (compared to a laptop) chassis, using convection in a way that Apple couldn't without making the PowerBook into something thicker than it currently is. Current laptops are already hot on the underside, but using chips that push serious wattage (the 970 is roughly twice as hot as the 7447A at any core clock that would be competitive) would make this worse.
2) Supply - There's already a problem with keeping G5s thoroughly in stock and that won't at all be helped by the addition of another product in the same line. This seems to be improving, but the PowerBook is one of Apple's highest volume sellers.
3) Most Pentium 4 laptops are hot, heavy, battery-chewing luggable desktop replacements, not a true portable that can be taken anywhere and used at will. They are loud, obnoxiously warm, and not at all equal to the level of design that Apple has been known for. I doubt people want a ten pound PowerBook.
1) Boohoo. Planning on using more than 8GB of RAM in your laptop sometime soon? If not, then this is such a non-issue that it's a joke. The only reason that 64-bit processors give any benefit in the x86 world is that AMD designed their Opteron and Athlon 64 lines with extra registers that only work in that mode and not traditional 32-bit.
2) Once again, boohoo. The dual-core 8641D is likely to beat the hell out of any single processor system to begin with. Why neuter a product ahead of time by intentionally limiting it? I say that they run with it across the entire pro line if IBM can't keep the 970 competitive, especially since I'm about to hit your next point...
3) Apple has already said that they forsee a future with both IBM and Freescale, just as things have been since the days of the G3. That being said, the compiler updates are probably far more minor than anything that had to be done for the G5's introduction, since the e600 still uses the full G4 instruction set with additions for the new features. In the way that the Pentium M is the Pentium 3 Super Ancestor Sparkle Edition, the e600-based 8641D will be largely compatible with existing code. Even if it were a problem, I seem to recall that Apple has little issue with fat binaries and. more recently, point-of-install compiling and optimization.
On the contrary, adopting the G5 for the sake of marketing is the boneheaded choice in this matter, especially with the obvious advantages that one can gain from SoC in a portable format. Were Apple to pass this chip up, it would be as ridiculous as the wait for updates from the original G4 to the G5 (which took two years of behind the scenes engineering and pre-planning). Of course, if they can pull it off with a laptop and an unannounced chip one time, I don't at all see why people think they can't do it again with another. After all, the G5 was released by IBM but never attributed to Apple, only to have Steve drop the bomb at WWDC.
khammack said:
While this is remotely possible, I think it's highly unlikely. Lying about your expected timeframe is never good business. If Freescale says it's not going to happen until 2H 2005, people will make buying decisions based on that information. And if a buyer needs a chip sooner, they'll choose something competitive from a different supplier instead.
Freescale wins nothing by delivering ahead of schedule. The best way to reestablish credibility is to deliver on time.
G-M
If Apple does use the dual core G4, they might brand it as simply G-M (abandoning numbers as Intel did with the P-M)... by branding the processor as G-M, Apple would differentiate it from the Powerbook G4 to emphasize a significant upgrade, but would not have to "muddy" the G5 name by branding a non-64-bit processor as a G5.
If Apple does use the dual core G4, they might brand it as simply G-M (abandoning numbers as Intel did with the P-M)... by branding the processor as G-M, Apple would differentiate it from the Powerbook G4 to emphasize a significant upgrade, but would not have to "muddy" the G5 name by branding a non-64-bit processor as a G5.
GonzoRob said:
Thank you GonzoRob and Caiwyn.
I was going to write the same thing, but you said it first !
@ Thatwendigo : I'm with you all the way (you have been saying all this or how long now
). You didn't mention the timeframe though. I personally think it's a bit early for dual core this or next week/month.
QFace said:
Except that shortly thereafter, both General Mills and General Motors would file lawsuits. I think they could develop a market for the "M" series in their mobile lines and "G" in their desktop lines, which would help alleviate future pressures to keep up when the engineering just didn't support it.
Edit: On a related note, after writing the above post, I had a semi-delirious vision of a PowerBook M1 with a new low-profile keyboard (that would both reduce the thickness and also keep dirt and dust and hair out from under the keys). Just think, a keyboard you could wipe! I don't know how they'd work out getting just enough travel so you could feel the key moving. But it would be fun...
Edit 2: Maybe a reassuring but quiet keyboard "tap" noise you could choose to disable with a button on the keyboard. Something hardware, like a small piezo-electric device that wouldn't require any software support, sort of like the old-school PC speaker. It would also make the computer more stealthy in conference rooms! Oh man I'm having fun with this.
I'd have thought we would have seen the small bump update to 1.5/1.67 GHz already if that was all the update was going to be. I too have a good feeling that dual cores may be on the way - I expect the delays to the line-up being updated is again not part of Apple's plan.
I too hope they don't call a dual core PowerBook a G5, but they may decide to emphasise the fact that it is a more significant upgrade by calling it the "PowerBook G4+" or the "PowerBook G4 DC".
I too hope they don't call a dual core PowerBook a G5, but they may decide to emphasise the fact that it is a more significant upgrade by calling it the "PowerBook G4+" or the "PowerBook G4 DC".
THIS is how things LIKELY are, read, you may find interesting (comments as well)?
I think the answer is simple.
Why would Apple produce new motherboards and processors (or more specifically pay another company to do so for them) for a technology that will likely be replaced with a Powerbook G5 in half a year? If Apple goes dual-core G4, they will be STAYING dual-core G4 for at least 1-2 years. It would be too costly to switch suddenly. Furthermore, we already KNOW they are making Powerbook G5's, evident from a previous rumor regarding a contract with Asustek and Quanta Computer (for refresh, check https://www.macrumors.com/pages/2005/01/20050114041111.shtml).
Apple already has a ton of G4's, wouldn't the smarter move be to overclock the current G4's a bit (hence a .17 upgrade) and add some features, fix all previous problems, possibly lower the price a bit, and therefore spur some more powerbook sales before completely switching to G5 eventually? What does everyone think the mac mini is but a genius move to get rid of more G4's by introducing a clever new product? Apple is making the transition to a new chip while trying to keep profits as high as possible.
I don't know what the upgrade will be to, but as has been rumored many times before, a nice, polished, speed bumped G4 powerbook seems like the obvious path. Enjoy.
I think the answer is simple.
Why would Apple produce new motherboards and processors (or more specifically pay another company to do so for them) for a technology that will likely be replaced with a Powerbook G5 in half a year? If Apple goes dual-core G4, they will be STAYING dual-core G4 for at least 1-2 years. It would be too costly to switch suddenly. Furthermore, we already KNOW they are making Powerbook G5's, evident from a previous rumor regarding a contract with Asustek and Quanta Computer (for refresh, check https://www.macrumors.com/pages/2005/01/20050114041111.shtml).
Apple already has a ton of G4's, wouldn't the smarter move be to overclock the current G4's a bit (hence a .17 upgrade) and add some features, fix all previous problems, possibly lower the price a bit, and therefore spur some more powerbook sales before completely switching to G5 eventually? What does everyone think the mac mini is but a genius move to get rid of more G4's by introducing a clever new product? Apple is making the transition to a new chip while trying to keep profits as high as possible.
I don't know what the upgrade will be to, but as has been rumored many times before, a nice, polished, speed bumped G4 powerbook seems like the obvious path. Enjoy.
Wendi's back! Been wondering where you've been, I've missed your insight.
Every news source that's reported on the 8641D has said it won't sample until 2H 2005. Apple might have some sort of pre-release engineering sample as you say, but I don't believe for a minute that it's going to be in the next powerbook revision.
Wouldn't all of these concerns be addressed by the possibility of a redesigned G5? Think Secret reported in September on Antares, as did Ars Technica. That's the dual-core G5 that's coming this year, also known as the 970MP. The single-core version of that chip would be the 970GX. Think Secret placed the release of that chip in "early 2005." If such a chip exists, it would address the heat concerns... and by using it only in the Powerbook line, it wouldn't disrupt the supply of the 970FX for the desktop machines.
Aside from that, I tend to agree with you. There's nothing wrong with the G4 architecture -- in fact, what I really want to see in the next powerbook revision is a faster hard drive, because that's where the real bottleneck is right now.
thatwendigo said:
Every news source that's reported on the 8641D has said it won't sample until 2H 2005. Apple might have some sort of pre-release engineering sample as you say, but I don't believe for a minute that it's going to be in the next powerbook revision.
thatwendigo said:
Wouldn't all of these concerns be addressed by the possibility of a redesigned G5? Think Secret reported in September on Antares, as did Ars Technica. That's the dual-core G5 that's coming this year, also known as the 970MP. The single-core version of that chip would be the 970GX. Think Secret placed the release of that chip in "early 2005." If such a chip exists, it would address the heat concerns... and by using it only in the Powerbook line, it wouldn't disrupt the supply of the 970FX for the desktop machines.
Aside from that, I tend to agree with you. There's nothing wrong with the G4 architecture -- in fact, what I really want to see in the next powerbook revision is a faster hard drive, because that's where the real bottleneck is right now.
John Rivers said:I think the answer is simple.
Why would Apple produce new motherboards and processors (or more specifically pay another company to do so for them) for a technology that will likely be replaced with a Powerbook G5 in half a year? If Apple goes dual-core G4, they will be STAYING dual-core G4 for at least 1-2 years. It would be too costly to switch suddenly. Furthermore, we already KNOW they are making Powerbook G5's, evident from a previous rumor regarding a contract with Asustek and Quanta Computer (for refresh, check https://www.macrumors.com/pages/2005/01/20050114041111.shtml).
Apple already has a ton of G4's, wouldn't the smarter move be to overclock the current G4's a bit (hence a .17 upgrade) and add some features, fix all previous problems, possibly lower the price a bit, and therefore spur some more powerbook sales before completely switching to G5 eventually? What does everyone think the mac mini is but a genius move to get rid of more G4's by introducing a clever new product? Apple is making the transition to a new chip while trying to keep profits as high as possible.
I don't know what the upgrade will be to, but as has been rumored many times before, a nice, polished, speed bumped G4 powerbook seems like the obvious path. Enjoy.
Let me say again, the fact that its called G5 means absolutely nothing regarding what chip is being used. Nothing at all, regardless of whether or not you think that's just stupid.
And if the 8641D was used, then it wouldn't have to be fazed out in two years. A 64bit derivative is already in the works.
So it could still go either way.
iRobot said:
G5 is an architecture. Unless motorola matches ALL specs of the G5 architecture, Apple would never call it a G5. Why?
"This application requires a Apple Macintosh G5 computer" becomes "This computer requires an Apple Macintosh G5 computer powered by an IBM 970 or 970fx processor". Users are stupid; they will not destroy the simplicity that Steve Jobs brought by the small product matrix and Gx processor naming.
Further more, Freescale's products are being moved down the line into the low-end product range. Everyone talks about how Steve said IBM and Motorola were going to be used forever on the mac. Thats not what he said.
He said something more like "As we look forward to the next few years, we are very pleased with the roadmap from both IBM and Motorola". Guess what guys, we are there. It is nearly 2 years later. Motorola/Freescale is now in machines that either haven't been updated in over a year, or in machines that are bare-minimums. Its time to face reality: Steve didn't say Motorola was staying on the mac, he said he was happy with where they were going for a 2-3 year period... which was just enough to satisfy the low-end eMac and Mac Mini. I'm assuming they thought they'd be at a G5 by now on the PowerBooks, hence the lack of an update for a year.
iRobot said:
That's naive at best. G5 means several things. It means PowerPC architecture (G5 literally stands for "Generation 5 PowerPC processor"). It also means 64-bit. Apple has spent a lot of time, effort, and money equating the G5 with 64-bit capability. If you really think that Apple would dilute the G5 brand by using it to describe a 32-bit processor, you have a tenuous grip on reality.
iRobot said:
Once again, the 8641D is not due out for at least another six months. The 64-bit derivative isn't due until even later, whether it's "already in the works" or not.
Caiwyn said:
Note that the PowerPC architecture was created by Motorola/Freescale, and the current G4s, obviously, and future G4s are obviously based on that.
Furthermore. A dual core processor from freescale would be a next generation, and G5 would be perfectly reasonable, or G5M or something of that sort.
I would say that the only person who has a "tenuous grip on reality" is one who thinks that Apple would limit their chip choices because of a NAMING CONVENTION.
But anyway, a dual core "G4" would be better than any G5 that'll be available for a laptop anytime soon. So why are you all deadset against the very idea?
Sure, its almost certainly not ready yet. And I personally would be happily surprised if we even got a 7448 soon. But the dual core freescale processors are definitely prime candidates.
