Motorola and PowerBook, iMac Delays [Archive]

solvs

Jul 28, 2003, 08:16 PM

Well, point-by-point:

1. Duh. Moto had some problems, 'Books are late. Maybe late Sept. If we're lucky. Duh.

2. No one ever said there was a target date for either. If the 970 is ready to sell by Sept, they are on schedule. Moto may be right on schedule with their 7457s for all we know. But they probably aren't. They're probably late, as usual. That's why we hate them.

3. Rebates end in Sept., coinciding with several trade shows. If they've got nothing to show, they will suffer. We will b*tch (and we will b*tch). And have every right to. But I'm, thinking Sept. anyway.

4. Please stop saying "c'mon" Apple and threatening to buy a Dell. Accept it, new 'Books in Sept. If we're lucky. If you're going to buy a PC, go ahead. No sarcasm, it's just a computer. Personally, I hate Dells. My Stepmom's is terrible, and has LITERALLY 1 hour of battery life. even when it steps-down to half the CPU speed and dims the screen.

5. The current 'Books aren't that bad. No one likes to buy something that will be "outdated" in a few months, but such is technology. The 12" still seems to be doing rather well. I wouldn't wait this long just for a minor speed bump.

6. No dual 'Books any time soon. No way.

7. G5 'Books will be nice, but I'm not holding my breath. I hope I'm wrong.

8. G3 with Altivec is called Mojave. So far it's a just rumor. The Gobi does not have Altivec (that we know of), and will be about 1.1 GHz. Not a rumor, but we don't know the specifics yet. Except that it will be coming later in the year. Possibly. We hope.

9. iBooks were updated in April. eMac in May. No updates for awhile. iMacs first, probably after the P'Books. Hopefully at higher speeds than 1.3 GHz. xServes, too. Hopefully G5s.

10. IBM doesn't suck. Moto does. IBM saved Moto's @$$ during that whole 500 MHz G4 debacle. If IBM has the same issues, they will suck, too. And we will hate them. Let's hope it doesn't come to that.

I really don't want to have to hate IBM too.

jaedreth

Jul 28, 2003, 08:23 PM

You do not seem to understand the dynamics of the situation, nor Motorola's shaky past with Apple.

And Apple cannot ditch plans that have been in the works for long periods of time, drop them at a hat, and then somehow shoot out a machine with no guts...

Apple can't ship a machine without a processor.

Apple can't use G3, the logic board is for G4.

Apple can't use G5, the boards won't be ready for half a year.

Apple is stuck over a barrel and it *IS* because of Motorola.

Motorola promised to ship the chips by mid May. It's now July, almost August, no chips.

So please refrain from purely derogatory and inflamatory comments that are devoid of any actual facts.

Thank you.

Jaedreth

wizard

Jul 28, 2003, 08:26 PM

With the current hardware that Apple has to choose from I don't see an immediate delievery of SMP notebooks. That doesn't mean that it is a bad idea by any means.

Once you'r running an OS like OS/X, SMP becomes a very credible feature on any PC. But like everything else in life you need to get by the engineering trade offs.

Thanks
Dave

Originally posted by sockseller
Can someone explain to me why people want dual-chip Powerbooks?

Maybe if Moto or Apple can cut the power consumption and radiant heat by a half...

windwaves

Jul 28, 2003, 08:45 PM

jaedreth,
it is still Apple's and Apple's only screw up. As simple as that.

jaedreth

Jul 28, 2003, 08:46 PM

You know, Apple does *want* to do Dual Proc G4's. However, they can't do that if Motorola can't supply an abundance of chips. Right now, they are still so far behind schedule it's not funny.

So even if we do get a PB G4 update, it's not going to be dual, thanks to Motorola.

So the PowerBooks will move to G5. At *that* point it will be an even further possibility of having dual processors. Not unless they wind up using low power consumption ones, at lower frequencies.

Now, if Apple could secure lower power chips, lower heat, lower voltage requirements, then it might be able to pull this off.

Imagine having PowerBook G5's available in 1GHz, 1.3 GHz (12"), 1.3GHz, 1.6 GHz (15"), 1.6GHz, Dual 1GHz (17").

So with this, Apple could offer a DP powerbook. Also, a DP 15" could be a configure to order via the Apple Store, if Apple wanted to be so bold. But they would have to use lower rating chips to do so.

Jaedreth

19823

Jul 28, 2003, 08:59 PM

tazznb

Jul 28, 2003, 09:15 PM

Originally posted by Powerbook G5
______________________________________

This could also be a time for a "new" G4 with the IBM G3+altivec. Wouldn't that help everyone? Give IBM's new plant more business to keep them more efficient, give us a quicker and higher yield of products, less delays, more updates, show Motorola we don't need to deal with their BS, etc. It seems like IBM already has the ability to crank out G3s that could spank a Motorola G4, all they need is to add the altivec and they'll really fly.

Correct me if I'm mistaken, but isn't a G4 basically a slightly improved G3 with Altivec?

macphoria

Jul 28, 2003, 09:25 PM

If a single processor has sufficient clock speed, dual processor is really not necessary. Also, for the sake of battery life, it is better to have single, efficient processor.

Personally I do not care if Apple does come up with dual processor laptop or not, as long as it performs well and has decent battery life.

Having said that, MOTOROLA IS NOT HELPING THE SITUATION MUCH!

For AGES, Motorola had Apple stuck around 500MHz with their G4 and even then caused delays because they couldn't ship enough chips on time. If you think Motorola is not at fault for causing trouble for Apple, slower sales (Power Macs) or losing market share (lack of speed upgrade on G4's) then you really have not followed Apple products long enough.

jaedreth

Jul 28, 2003, 09:38 PM

Yeah, anyone who has followed Apple since before the PPC era knows the score...

I remember when the 020 was a currently shipping chip.

And I agree about the battery issue. However, certain recon hints that Apple may be looking at alternatives to batteries. I really do expect the G5 to use this new technology if Apple has in fact already found a supplier, and if Apple does, dual processors will not be a problem.

But that's pure rumor and speculation at this point.

Jaedreth

saint.duo

Jul 28, 2003, 09:43 PM

No, but heat may be...

Originally posted by jaedreth
I really do expect the G5 to use this new technology if Apple has in fact already found a supplier, and if Apple does, dual processors will not be a problem.

macphoria

Jul 28, 2003, 09:44 PM

Correct me if I'm mistaken, but isn't a G4 basically a slightly improved G3 with Altivec?
Yes it is. AltiVec is a type of SIMD (Single Instruction Multiple Data) that is designed to help processing multimedia (movies, music, etc) data informations.

But AltiVec is Motorola's version of SIMD and IBM has its own that it has implemented on their 970/G5 processor. IBM never embraced AltiVec and stuck with G3 development without SIMD. But there has been rumors about IBM developing G3 with their own SIMD (Gobi or something rather). So technically if IBM does build such processor, it could be could be considered a variation of G4.

Some people who consider G3 from IBM to be ideal for mobile computing because of good performance, energy requirement and etc. So it is not entirely without hope or expectation that IBM does develop its own G4 that can outperform G4 from Motorola.

JtheLemur

Jul 28, 2003, 09:47 PM

Go Motorola!

new user

Jul 28, 2003, 09:56 PM

awinn233

Jul 28, 2003, 09:57 PM

Simply unacceptable, Motorolla. I'm tierd of it. Apple's products have suffered because of you for the last time! :mad:

windwaves

Jul 28, 2003, 10:06 PM

new user,
if I were you I'd wait, especially considering that you can "afford" to wait. Wait and see what happens, till school starts. I have a feeling the current PB's are pretty archaic and would not keep up adequately with new OS as well as other software developments.

Wonder Boy

Jul 28, 2003, 10:11 PM

I come back from cape cod and this is what i have to read? Curse you motorola and your inferior products!

macphoria

Jul 28, 2003, 10:11 PM

it is apple's fault, they should have plan b ready the second it was obvious that mot wasn't going to be delivering the goods, probably was a couple of weeks before june wwdc.

plan b could have included cosmetic 15pb updates, like bluetooth, extreme, and firewire 800. not the biggest deal but sufficient to convince some folks to buy now instead of waiting (and waiting and waiting) and not think they will be getting an obsolete machine.

instead of being severely outdated, the 15 pb would just be outdated as the rest of the line.

it is silly to partially update the 15 pb, then updated again with the rest of the line, but a consumer oriented company would have considered that option, especially considering past delays.

either way, i'm waiting until a couple of days before school starts and i'm buying whatever the 15 pb product is out there. unless some one here would like to argue that the upgrades and price increase of the 15 ai doesn't make up for the cost saving of the 15 ti right now (with ed discount). anyone saying that?
Cosmetic update for 15" PowerBook?

The real issue at this point is the performance of Apple products which is the area that Apple has been lagging (Motorola being a significant factor, if not the factor). Cosmetic upgrade at this point is pointless. Apple already introduced Power Macs with G5. People are expecting G5 PowerBooks, which is probably another year away from materializing. They need to boost performance on G4 PowerBooks. Mere cosmetic update is pointless. And if they want to update their PowerBook lines, they need latest processors for design and configuration. Apple probably already did worked that out with Motorola, except now Motorola can't deliver the parts.

Apple has been working with IBM and Motorola in partnership developing PowerPC processors. Only member that has consistently caused problems is Motorola and that is a fact. There has been many rumors about Motorola working on G5 but that never came through. IBM at least, even though it took them a while, delivered 970/G5.

Gyroscope

Jul 28, 2003, 10:18 PM

Originally posted by macphoria
Yes it is. AltiVec is a type of SIMD (Single Instruction Multiple Data) that is designed to help processing multimedia (movies, music, etc) data informations.

But AltiVec is Motorola's version of SIMD and IBM has its own that it has implemented on their 970/G5 processor. IBM never embraced AltiVec and stuck with G3 development without SIMD. But there has been rumors about IBM developing G3 with their own SIMD (Gobi or something rather). So technically if IBM does build such processor, it could be could be considered a variation of G4.

Some people who consider G3 from IBM to be ideal for mobile computing because of good performance, energy requirement and etc. So it is not entirely without hope or expectation that IBM does develop its own G4 that can outperform G4 from Motorola.

There are more differences between G3 and G4 than just Altivec.
G3 was basically derived from PPC 603. G4 borrowed lot from its PPC 604 father :).eg FPU unit, that is considered to be much better than G3 one. I'm just sayin' that if you slap altivec unit on G3 it doesn't necessarily turns it into G4.

macphoria

Jul 28, 2003, 10:26 PM

There are more differences between G3 and G4 than just Altivec.
G3 was basically derived from PPC 603. G4 borrowed lot from its PPC 604 father .eg FPU unit, that is considered to be much better than G3 one. I'm just sayin' that if you slap altivec unit on G3 it doesn't necessarily turns it into G4.
Yes indeed. But hopefully you got the point I was trying to make, which is IBM has moved its G3 development along (which by the way can outperform Motorola's G4 at similar clockspeed in some non-AltiVec tasks) and could come out with G3 with SIMD that has potential to outperform Motorola's G4.

Gyroscope

Jul 28, 2003, 10:46 PM

Originally posted by macphoria
Yes indeed. But hopefully you got the point I was trying to make, which is IBM has moved its G3 development along (which by the way can outperform Motorola's G4 at similar clockspeed in some non-AltiVec tasks) and could come out with G3 with SIMD that has potential to outperform Motorola's G4.

Point certainly taken. Lets just hope that IBM will be able to come up with something in some reasonable time, that would replace Motorola CPU's in Apple's consumer boxes.This thing with Motorola inability to deliver is getting downright pathetic.

jbomber

Jul 28, 2003, 10:57 PM

well this just sucks worse than anyone could have imagined.... i think i'll go turn on the oven and climb in.

Ok I'm over that.

How the HELL did this happen? could apple really have let motorola jerk them around for this long without having a plan b? if the rumors are saying that production is delayed until mid september, what's to suggest that they'll sort out whatever problem they're having between now and then? They're already months and months overdue. I don't see how another 2 weeks is going to fix whatever new catastrophe they've encountered.

It seems to me that laptops are probably going to get a tiny speed bump if anything at all, by overclocking the existing chips, and then another (hopefully) sizeable price drop. Once the deadline approaches and motorola drops the ball again, we'll be looking at early next year for an updated machine. By then, who knows. Hopefully apple put plan b into effect as soon as they got wind of motorola's foul-up.

kshkval

Jul 28, 2003, 11:00 PM

just checked the apple store and it's off limits, it seems... maybe some PB or other news on the way???

Nebrie

Jul 28, 2003, 11:01 PM

Originally posted by kshkval
just checked the apple store and it's off limits, it seems... maybe some PB or other news on the way???

Loads fine for me. You ran into one of their maintainence periods.

macphoria

Jul 28, 2003, 11:14 PM

IBM G3 750VX "Mojave"

http://www.bayarea.net/~kins/AboutMe/GIFs/CPUs.html

http://www.macrumors.com/pages/2003/05/20030509020831.shtml

"750VX ("Mojave") is an Altivec enhanced G3 to debut at 1.5GHz"

That sounds promising right about now.

new user

Jul 28, 2003, 11:22 PM

The real issue at this point is the performance of Apple products which is the area that Apple has been lagging (Motorola being a significant factor, if not the factor). Cosmetic upgrade at this point is pointless.
--------------------------------------------------------

the real updates, like waiting for motorola, apple has no control over. what they do have control over, the minor cosmetic changes, should have been plan b. hell, to make it up for the users, they could throw in some rams or something. it isn't cosmetic or pointless for the people who want to buy now. my point is that even if apple was at the mercy of outside vendors, like motorola, just dropping the price and nothing more isn't enough.

my only hope is that with all the rumors leaking about the delays and such, is that the problem has been fixed and the new machine will be expected very soon. of course, that is just a fancy way of saying, wait.

robodweeb

Jul 28, 2003, 11:25 PM

Originally posted by jbomber
How the HELL did this happen? could apple really have let motorola jerk them around for this long without having a plan b?

Well, Apple has been paying attention to things other than growing processing power: redesigning the interior of a computer to fit a hemispherical shell, iPods, music stores, retail stores, etc. None of which do the thing we've been wanting most: faster, more capable computers.

Apple (including Steve) does deserve partial blame for this debacle ... not as much as scum-a-rola ... but deserves some. Competent management would not have allowed a single point of failure to persist for so many years.

sacrilicious

Jul 28, 2003, 11:33 PM

Vigilante justice at this point?

jbomber

Jul 28, 2003, 11:34 PM

Originally posted by robodweeb
Well, Apple has been paying attention to things other than growing processing power: redesigning the interior of a computer to fit a hemispherical shell, iPods, music stores, retail stores, etc. None of which do the thing we've been wanting most: faster, more capable computers.

Apple (including Steve) does deserve partial blame for this debacle ... not as much as scum-a-rola ... but deserves some. Competent management would not have allowed a single point of failure to persist for so many years.

That's true. Apple was irresponsible for not putting together a few contractual failsafes with Motorola. Anything to keep them motivated, since they're obviously draggin their sorry @$$es with these new chips.

DrGonzo

Jul 28, 2003, 11:40 PM

http://www.2trak.com/thread-sucks.jpg

macphoria

Jul 28, 2003, 11:43 PM

This article was in MacWorld magazine, August 2001. Two years ago.

http://www.macworld.com/2001/08/bc/buzzstandbycpu/

bretm

Jul 28, 2003, 11:50 PM

Originally posted by Waluigi
By next year this is what I want their product line to look like (all processors are G5's):

PowerMac:
Dual 3GHz
Dual 2.7GHz
Dual 2.4GHz

PowerBook
17" Dual 2GHz
15" Dual 2GHz
15" 1.8Ghz
12" 1.6GHz

iMac
17" 2GHz
15" 1.8GHz

iBook
14" 1.4GHz
12" 1.4GHz
12" 1.2GHz

eMac
1.4GHz
1.2GHz

If they went all G5, they would have no dependencies on motorloa anymore. It would be really cost efficient to have only one chip to make rather then 3. It would also be better for apple to fully migrate their apps and what not to the G5. It would also help sales of the iBook, because no one wants a G3 now that the G5 is out.

If IBM can keep on making faster processors on a regular basis, the powermac line should have no problem keeping on top, with the latest and greatest features too. This means performance all across their lines, with big profits too. It can be a win-win situation, although it is more of a dream now...

--Waluigi

I just don't understand what this relationship between screen size and processor size is. Maybe I'd like a dual 2ghz 12" powerbook. Screen size (on a pb at least) is a portability issue. Why is the processor size linked with the screen size?

Freg3000

Jul 29, 2003, 12:00 AM

Originally posted by bretm
I just don't understand what this relationship between screen size and processor size is. Maybe I'd like a dual 2ghz 12" powerbook. Screen size (on a pb at least) is a portability issue. Why is the processor size linked with the screen size?

The logic is that if you are getting a Dual 2 GHz, you obviously need power for running apps like Photoshop and FCP. In that case, you need a lot of real estate, so a bigger screen would naturally follow. More power ----> Power user ------> Pro Apps ------> Big screen. It's just a logical conclusion based upon a few simple deductions.

jbomber

Jul 29, 2003, 12:02 AM

i don't understand what the obsession is with dual processors is. isn't one of the true marks of a great portable its battery life? Dual processors will cut your battery to a couple of hours at best. not terribly efficient. plus the heat would likely fry your khaki dockers... if you need that kind of processing firepower, a desktop is the way to go.

iB24

Jul 29, 2003, 12:10 AM

Originally posted by jbomber
i don't understand what the obsession is with dual processors is. isn't one of the true marks of a great portable its battery life? Dual processors will cut your battery to a couple of hours at best. not terribly efficient. plus the heat would likely fry your khaki dockers... if you need that kind of processing firepower, a desktop is the way to go.
LOL

I totally agree.

mnb

Jul 29, 2003, 12:23 AM

Motorola has always dropped the ball. It's always lagged behind in speed, delivered small quantities, and delivered them late.

IBM typically has rose to the task and met or exceeded expectations.

If these companies were soldiers, Motorola would either be court-martialed or constantly on KP duty, while IBM would get lots of Liberty Passes and medals for meritous conduct.

If they were my customers (which they are to Apple), Motorola would have been fired and IBM would have been given a nice, fat, long term contract.

Motorola makes pretty good phones. Their committment to computer CPUs is weak and their performance shows it. The only reason I can see that Apple hasn't dumped them is that due to their phone/pager/etc background, they do pretty good at designing (not making) low power chips that are required for laptops.

ryaxnb

Jul 29, 2003, 12:23 AM

Perhaps the most annoying part is that Motorola could do it if they cared. I bet they could make at least a 1.2 Ghz G4 for laptops. In any case, AltiVec isn't the way to go. It was good in 10.0/.1 but now a 2.4 Ghz speed would be much better.
Updated: Also, I have love for the G3 - it provides plenty of power for my needs (and I should know, I have an iBook) - and I would actually be disappointed in an AltiVec update, at least until programmers (like those of Office, iPhoto, etc) release sufficently AltiVec-enhanced programs.

Dros

Jul 29, 2003, 12:32 AM

Originally posted by Freg3000
The logic is that if you are getting a Dual 2 GHz, you obviously need power for running apps like Photoshop and FCP. In that case, you need a lot of real estate, so a bigger screen would naturally follow. More power ----> Power user ------> Pro Apps ------> Big screen. It's just a logical conclusion based upon a few simple deductions.

I'm a person for whom the logic you give doesn't apply--computational biology calculations need speed, but not the screen real estate. But, I'm sure Apple knows their target consumers. The limited model set does make it hard for people in the minority, though.

I'm happy that we seem to be approaching a time when the difference between a high and low end processor is getting smaller. Processors running at 300,400 and 500 MHz have a big percentage difference compared to 1.6, 1.8 and 2GHz processor. Of course, putting duals in changes the speed gap as well.

Powerbook G5

Jul 29, 2003, 12:39 AM

I'm sure the logic behind dual processors in the bigger models is more from the fact that it can't be as easy to cram two into a 12" with very limited space as it would be to cram them into a 17" that would have the room to spare. Pop open a 12" Powerbook and take a good hard look at where you would consider it possible to add a second processor, then feel how hot it is with just that one processor in it as well. I'm sure in a perfect world, we could imagine all computers having dual processors. Hell, put 4 processors into a Cube! It's just not as easy actually applying that in the real world with factors such as heat, power, space, price, etc.

rotorblade

Jul 29, 2003, 01:10 AM

I knew there were going to be serious problems when Motorola hired three of Bill Gates cousins and placed them in top positions throughout the company back in 1997. Seems the war didn't end in '97! It was just starting!
.
Unfortunately, none of this is true. Damn. Can't even come up with a good excuse for all the b.s. we've been enduring over the years. This blows!

jbomber

Jul 29, 2003, 01:40 AM

all-in-all it's a bad turn of events especially in light of the past few years of Apple Bliss that we've had. There's no excuse, but the fact that things had been so good for so long makes it stick out.

Sol

Jul 29, 2003, 01:45 AM

I agree that Motorola has let Apple down with yet another delay but what is the alternative at this point? G5 PowerMacs have not even shipped yet and nearly every other portable processor is x86 based. I have hoped that Transmetta's processors would one day make it into Apple laptops simply because of their superior heat ratings but that dream never materialised.

As for Windows-running laptops, buy them at your own risk. No matter how much faster their processors are supposed to be, they do not compare well to the iBooks and PowerBooks when you look at the overall build quality, IO ports, and last but not least, the operating system itself.

When it comes to portable systems Apple's products are still the standard the others try to reach. If the best the competition can do is use a processor with a higher GHz rating (that runs even hotter) then the situation Apple finds itself in today is not as bad as this forum's participants make it out to be.

globalstar

Jul 29, 2003, 01:48 AM

Hy guys. First let me say that this is my first post in the forums. I discovered macrumors a few weeks ago and have been lurking about since.

Anyway I was just thinking that if Steve hoped to unveil the new PB-s at WWDC, then when put into a situation where he had no processors to put into them and having to deliver a keynote that was scheduled, they had to bring something else out. Something big. The G5.

This being the "year of the laptop" - yeah right - I really believe that they wanted to bring the PB-s out not the G5.

Actually it wouldn't have made too much sense to bring them out if they had some other wonderwul products. Why? Because apple knows that announcing something and not being able to deliver it for some 2(!) months is not good for business.

We already know that the Xserve's sales have plummeted. The people here are now thinking to forego the next PB update and wait for the G5 etc.

G5 is good publicity and has given Mac afficionados some hope for the future, but releasing it more than 2 months before selling it (or well, delivering) is nt good business.

I beleve Steve knows that and must be really unhappy with this. As we are...

uberman42

Jul 29, 2003, 02:45 AM

My gosh...the venom being spewed in this forum...

It makes proper business sense for apple to continue to clear the channel of the existing product. Remember the days when everyone screamed the sky is falling when apple had over a months worth of inventory. Apple is smart, they know you guys are waiting for the new goods. But they also know it makes good business to sell as much as they can of what they have already made. It's being fiscally responsible.

You guys wanna see the new powerbooks? Buy the rest of the gigs and 867s that are in stock...open sesame....

So continue to add coins into your piggy banks, go back to your 667 Mhz Pbooks (which are pretty good btw), do some yoga, align your chakras, and breathe. Patience is a virtue. And when the time comes for a new Pbook and you are wowed, all is forgiven for another 6-9 months.

tduality

Jul 29, 2003, 02:58 AM

syclone

Jul 29, 2003, 03:43 AM

I've followed these forums for a while, but never posted before, but i just wanted to chime in with my 2 cents this time.

I'm going to be mad at Motorola right now, just because i need someone to be mad at right now, and can't bring myself to blame apple.

My situation is this. I'm starting my second year of undergrad in three weeks. At the beginning of last year I bought a shiny new iMac G4 with the Superdrive. I was going to get the 17" I kept waiting and waiting and waiting. All of the stores had display units, but I couldn't get my hands on one until 2 weeks after the year started. I broke down and bought the 15" at a local store on the day before classes started. Overall, I have been more than pleased with the purchase. I ddon't want to give up my hemispherical treasure. I have yet to even use the DVD burner. But the thing is too much of a pain in the ass to bring to and from school. The only way to really pack it for a long car trip is in the original box which is clunky and damn heavy. Anyway, the family here needs to replace the old beige G3 - my mom never upgraded coz there was no Quark XPress for OSX until a month ago, but thats another story.

Anyway, I need a laptop for school. When the 12" and 17" came out I was psyched that the 15" would come up to match them really soon. But that never happened. I've settled on a 15" - It's going to be mostly a desktop so the 12" is too small, and the 17" is way out of my price range.

But I really wish Apple could at least announce the damn thing so I could preorder or wait to get a current 15" at a big discount. The big selling points for me are bluetooth, airport extreme, and usb 2.0. If they could just put those in the aluminum case with the cool ney keyboard for around the same price I'd be sold. Hell, even if I could just get that much guaranteed for a month from now it would be great.

But there is simply no way in hell I'm going to pay more than I did for my iMac a year ago for a slower PB with fewer features. Of course with the games Apple was playing last year I should have expected it. Then it was the display manufacturers, now its Motorola.

Almost makes me want to switch to Dell. That is if my roommates didn't just up and refuse to power on after a few drops of water on the keyboard, and another close friends bluescreened in IE after about 15 minutes or so every time I tried to use it. Dell took the truly broken one after about 2 hours of arguing on the phone kept it for 2 weeks, and charged a healthy sum to fix it without ever explaining what they did. The one buggy with software just got a note explaining that it was a common problem that would be updated soon that never was.

I wouldn't blame people for leaving Apple at this point - If I didn't have so much invested in software on this thing I would. But whatever you do, don't get Dell. At least when Apple finally gets these G4's, they will work right, unkuke Dell who rushes their suppliers for stuff thats junk in the first place. Then again, the PB5300 did catch fire from faulty batteries...

Phil Of Mac

Jul 29, 2003, 05:05 AM

Originally posted by solvs
I really don't want to have to hate IBM too.

Then again, hating IBM would have a rather nostalgic feel to it :)

Originally posted by macphoria
Yes it is. AltiVec is a type of SIMD (Single Instruction Multiple Data) that is designed to help processing multimedia (movies, music, etc) data informations.

In the end, yes, but it's not directly designed for that. It's *directly* designed to process vectors. Vectors are numbers with directions, whereas normal numbers without directions are called scalars. A direct application of vectors is in physics. If I am traveling 20 MPH east and throw a baseball to the north at 50 MPH, for instance, calculating the direction and speed of that baseball relative to the ground would be a vector operation.

WorldMage

Jul 29, 2003, 05:40 AM

Originally posted by sockseller
Can someone explain to me why people want dual-chip Powerbooks?

Maybe if Moto or Apple can cut the power consumption and radiant heat by a half...

I think the idea is that you use two slightly lower power processors. Then when you are on battery you can shut one of them off and get better run time than one high speed CPU (that lacks elaborite step down speed controls) or leave both on and keep around lots of spare batteries :)

Heat and interference are still issues in such a compact form factor, and this is likely the reason the 17" is often mentioned as a canidate for dualies.

crenz

Jul 29, 2003, 06:46 AM

This one definitely cracked me up:

Apple CEO and co-founder, Steve Jobs, is said to be so irate over the issue, he often uses profanity when expressing his displeasure with the semi-conductor sector of Motorola.

crenz

Jul 29, 2003, 07:03 AM

Actually, how does this fit together with the fact that the 7457 was originally to be shipped in Q4 2003? Does anybody know when Motorola moved that to Q3? If that was not too long ago, I would find it difficult to believe that Apple wanted to announce new PBs at WWDC.

Lanbrown

Jul 29, 2003, 08:35 AM

Originally posted by Mudbug
I'd like to be the first to say:

I knew it was motorola all along.

my 2¢

Motorola just recently announced third quarter availability for the 7457, before that, it was the fourth quarter.

Maybe they should just outsource chip production. Ti would be a good source, better fab equipment and better technology as well.

Lanbrown

Jul 29, 2003, 08:42 AM

Originally posted by nydoofus
Actually, I don't know how bad their new phones are, but their older phones used to be quite good. My 5 year old Startac is still functioning quite well, despite being dropped many many times.

Maybe a parallel can be drawn here. I don't know.

Actually, in many cases their quality was lower then the competition. The failure rates were far higher and they had software problems that took them years to finally acknowledge and fix.

As soon as they started to produce digital phones, their quality went down hill. No one could beat their analog phones; those were virtually indestructible.

Lanbrown

Jul 29, 2003, 08:51 AM

Originally posted by NHMac
Showing some extreme ingorance here... is there any IP that Moto or IBM has that would prevent another high volume chip maker, say AMD, Intel, Nvida, from making a G4/G5 chip? For Apple? They seem to have the processes down, why not leverage the volume and expertise they have in mass producing many, many processors?

Well, if you look at the PowerPC it deals more with the instruction set then the actual processor. There are many different PowerPC processors out there. Some networking gear uses a PowerPC processor. The G5, which is the IBM 970, uses the Power4 core, which IBM owns. Much like Motorola owns the G4. The instructions sets are pretty much the same, but are different chips. Motorola would have to allow someone else to produce the chips.

If they were going to outsource it, it would be IBM or Ti. Ti is actually ahead of IBM in terms of chip technology. Setting up a pre-production run would take several months at a minimum.

rjwill246

Jul 29, 2003, 08:59 AM

Originally posted by jbomber
That's true. Apple was irresponsible for not putting together a few contractual failsafes with Motorola. Anything to keep them motivated, since they're obviously draggin their sorry @$$es with these new chips.

But the vendor of the new processors would have been...........?
Moto simply couldn't/can't make them. Apple and Moto have contracts in place, the lower prices on the G4' is probably part of the 'penalty' for not meeting Apple's needs... but that still doesn't change the fact that Moto can't do any better. Reports abound on their terrible harvesting of quality chips. IBM is the answer, it seems and we have to wait til the G5s come pouring out of Fishkill. There is not ONE other solution to this dilemma.
So, I don't agree that Apple management is to be held responsible for this. The alternative was to produce nothing since Apple can't make Moto improve on the physical disaster of Moto's manufacturing setup. At least Apple tried to improve on the rest of the computer line during this 'wait and see' period. It's far from wonderful but before we go nuts about this, has anyone's computer stopped doing what it did well, before we knew the G5 was coming?

Little Endian

Jul 29, 2003, 10:02 AM

Yes Motorola is crappy at meeting deadlines for Processor revisions and have had rampants supply problems from the get go back all the way too the G4's of 1999. Motorola has had a history of taking time with the G4 chip as far as improvements and meeting supply needs.

Apple should have been well aware of this and planned accordingly by using the G4 chip in the most resourcefull ways by equipping all it's G4 computers leaving the factory with as much power as possible to lenghthen the product lifespan this is especially true with the G4 based imac and PowerBook lines. Seeing how both these machines are non upgradeble and Knowing Motorola's delays Apple should have equipped it's entire PowerBook and imac lines with level 3 Cache. Faster Graphics Cards would have helped as well.

Ex.
The 12" Powerbook 867Mhz G4 should have been released from the get go with at least 1Mb of level 3 cache instead of none and should have utilized the 167Mhz bus instead of the 133Mhz bus. Apple could have also done better by using a faster chipset than the Nvidia Geforce 4MX 420 Go chipset. They should have offered it with at least the 440MX Go or better yet the Radeon 9000Pro Mobility. Also they should have sported FW800 and USB 2.0 from the get go.

The 17" 1Ghz should have also been equipped from the get go 2Mb of L3 cache which would have been Possible. Also Apple would have been wise to use a faster chipset than the Geforce 4Mx 440Go . Apple should have Used the Radeon 9000Pro Mobility which does perform better than 440MX which is odd considering the mid range Powerbook 15" offers the Radeon 9000Pro. An even better option would have been to use something like the Geforce Ti4200 Go or equivalnent.

Putting 5200 RPM notebook drives also would not have hurt either.

imac 17" 1Ghz should have also shipped from the beggining with 1MB of level 3 cache and maybe even the 167Mhz bus.
Also Apple could have choose to use a faster Graphics Card than the Geforce 4MX in the imac which is a 440MX to be exact. It would have cost apple very little to implement the 460MX card or for that matter just use the Radeon 9000Pro.

Bottom line is Apple can control things on their end of the stick, they just choose not too. Often times they hamper themselves from the Get go. Steve should have foreseen delays form Motorola and Should have made sure all 2003 G4 machines leaving Apple's doors were equipped to be the longest Cycle holdout attempt ever.

Phinius

Jul 29, 2003, 10:06 AM

In February of this year Motorola had a press release (http://www.motorola.com/mediacenter/news/detail/0,1958,2322_1901_23,00.html) announcing the 7457. It clearly states that the chips would become available in the fourth quarter. This is now the beginning of the third quarter. As far as I can tell the supposed delays are just wild speculation based on rumors.

Motorola has been slow getting a PowerPC processor made on a 130-nm process. But that is true about all of their chips and not just the ones made for Apple. This slow progress is probably greatly due to Motorola losing money for several quarters in a row. There have been severe cutbacks at Motorola and slowing down the pace of manufacturing process advancement was probably greatly influenced by this.

Phinius

Jul 29, 2003, 10:22 AM

In October of 2000 IBM announced the 750FX. IBM stated that it would run up to 1 GHz. Now IBM lists the topend speed for the 750FX as 900 MHz. Just recently IBM announced that the 750GX will be produced in December of this year and it will run up to 1.1 GHz. Great, the 750 G3 gets to 1 GHz over two years after IBM announced a 1 GHz version. This has caused a delay in upgrading the speeds of iBooks. So I'd hardly call IBM a company without disappointments for Apple.

Abstract

Jul 29, 2003, 11:01 AM

I think IBM is better than Moto.

And yes, this is like an abusive relationship where Apple just can't seem to fully detach itself from Motorola. Its sad when it happens to people, but its bad business to stick with your mate after all this.

jettredmont

Jul 29, 2003, 11:44 AM

Originally posted by jaedreth
Well, Motorola *did* fire a bunch of people.

But they didn't hire back.

They are understaffed.

I bet their MOS's, especially MOS 13 are hemorraghing money.

As they are, they're leasing unused space in their three Austin facilities to other companies, to help keep rent... In short, they're going downhill. Perhaps if they focused on one thing... Cell phones, and ditched everything else, they could stay competitive. Maybe.

Jaedreth

Perhaps they should sell their processor business (including plants) to AMD ...

As much as I despise Motorola, Apple needs multiple CPU sources. Perhaps "needs" is a bit strong, but it's certainly better to have two suppliers than one (you can draw a very strong parallel to the G4 conundrum ... at the beginning, it was IBM that said it wouldn't do the G4 or Altivec, leaving Apple with a single-supplier situation if it wanted to actually use its one processor advantage ...)

Motorola just "going away" is bad for Apple. Very bad. Even worse than them just being second-rans that Apple might be able to fall back on if IBM completely screws the pooch on the G5s.

Of the likely contenders, an AMD buyout would be both most likely (Intel is unlikely to want Moto's SC business) and the most Apple-friendly (Apple and AMD are on pretty good terms, and AMD has a record of funding processor development that Moto can't touch).

Brings a whole new light to those old "Jobs meeting with Ruiz" Apple-AMD rumors, huh?

jettredmont

Jul 29, 2003, 11:45 AM

Originally posted by nydoofus
Doesn't IBM have a G3+Velocity Engine? (Gobi i think). Apple must have some contract with Motorola keeping them with Moto.

Not officially, and according to rumors this is a ways from release.

Yes, Apple likely has contracts with Moto (although I think Apple could easily get out of them given Moto's repeated business failings).

jettredmont

Jul 29, 2003, 11:49 AM

Originally posted by AidenShaw
Wow, IBM has yet to meet a target date for the PPC970, yet you'll trust them with everything.

I pity poor IBM if they're a day late with the PPC970, you hyenas will tear the flesh from their bones in an instant! ;)

Well, IBM has a long history of meeting or beating target dates.

Motorola has a long history of missing target dates.

This is an opinion with a long history. If IBM were to miss its date on the 970, that would be forgiven (to some extent) because it will be a relatively new phenomenon and unlikely to happen again.

jettredmont

Jul 29, 2003, 12:03 PM

Originally posted by bretm
I just don't understand what this relationship between screen size and processor size is. Maybe I'd like a dual 2ghz 12" powerbook. Screen size (on a pb at least) is a portability issue. Why is the processor size linked with the screen size?

Bigger screen == bigger case (although some space may be reduced in thickness, width and depth of the notebook are directly related to screen size ... until oled's fold out from a smaller enclosure of course ...)

Bigger case == more optimal component spacing

more optimal component spacing == easier heat dissipation

better heat dissipation == can use more heat-generating components (higher-frequency processors)

&&

Bigger case == bigger battery

bigger battery == can use more energy-consuming components (higher-frequency proessors)

whatever

Jul 29, 2003, 12:10 PM

Originally posted by mrsebastian
give me a f$@king breaK!!! i don't care who apple wants to blame (ie: motorola) but in the end, why can't get their ***** together and release a product on time for once?! i've been waiting forever for a 15" albook and it looks like i'm gonna be waiting some more. i love apple and i ain't switching anytime soon, but for the company that's the most user friendly in the computer world, they sure are sorry when it comes to producing and distributing hardware!
sincerely,
tired of the same old "delayed again" story

I don't believe Apple has announced a new Powerbook, so I'm not sure what the delay you referring to is. The current Powerbooks are selling well, so why would they upgrade them for no reason? But then again are you one of those who will complain when Apple releases new G4 Powerbooks and state that they're going to wait for the G5s?

Whatever

jettredmont

Jul 29, 2003, 12:18 PM

Originally posted by Phil Of Mac
In the end, yes, but it's not directly designed for that. It's *directly* designed to process vectors. Vectors are numbers with directions, whereas normal numbers without directions are called scalars. A direct application of vectors is in physics. If I am traveling 20 MPH east and throw a baseball to the north at 50 MPH, for instance, calculating the direction and speed of that baseball relative to the ground would be a vector operation.

In case you're not joking:

You have a good description of vectors in physics; unfortunately "vectors" in computers are a completely different beast.

In computing terms, a "vector" is an arbitrarily-sized array of values. A "vector unit" will typically deal with vectors of 32-bit ints, 16-bit ints, 8-bit bytes, and 32-bit floating point numbers more efficiently than by dealing with each element of data individually. In Altivec, the vector processing size is 128 bits, so if you have 32-bit ints in your vector the Altivec unit can process those ints four at a time (4x32 = 128). However, you can only do this when you are doing the same thing to every int (for instance, multiplying each by 3). Another name for a vector unit is "SIMD", which stands for "Single Instruction, Multiple Data".

This is a good thing for "multimedia" applications as they often do the same thing on multiple bits of data. Other apps ... well, there's not much opportunity for SIMD in a word processor, for instance.

No, I have no idea why they chose the name "vector" for this, but they did, and we must live with it :) The term persists in several computer languages as well (C++ and Java, for instance).

Snowy_River

Jul 29, 2003, 01:49 PM

Originally posted by new user
they should have plan b ready...

Originally posted by robodweeb
Competent management would not have allowed a single point of failure to persist for so many years.

With regards to these comments, don't you understand that the IBM G5 (PPC970) is Apple's plan B? It is their answer to the 'single point failure'? It was about two years ago that Moto truly demonstrated their failure, when they discontinued their development of the G5 (PPC8500?). That was when (if we are to believe what evidence we have at our disposal) Apple approached IBM to start development of the PPC970.

It's clear that Apple is trying to move away from Moto entirely. However, this is a process that takes long term planning. On the plus side, if the rumors of Mohave are correct, then we might well see it in the near future, as IBM will have a 'G4' type chip for Apple to use for it's lower end products.

Originally posted by Phinius
In October of 2000 IBM announced the 750FX. IBM stated that it would run up to 1 GHz. Now IBM lists the topend speed for the 750FX as 900 MHz. Just recently IBM announced that the 750GX will be produced in December of this year and it will run up to 1.1 GHz. Great, the 750 G3 gets to 1 GHz over two years after IBM announced a 1 GHz version. This has caused a delay in upgrading the speeds of iBooks. So I'd hardly call IBM a company without disappointments for Apple.

This is something that it's hard to blame IBM for. I think that this is likely the fault of Apple marketing and Moto. Apple marketing has long held that consumer level products shouldn't outperform their Pro-level machines. So, because the G4 was lingering at a maximum of 800 MHz by 2001, there were not going to be any 1GHz G3 to be found. Even now, the fastest PowerBook is at 1GHz, so you don't find anything faster the 900MHz in the iBook (which is now the only G3 processor in the Apple line).

This is a long standing issue that has been debated ad-nausium. Some may not agree with Apple's choices in this area, but at least try to understand them.

jbomber

Jul 29, 2003, 02:10 PM

Originally posted by Lanbrown
Motorola just recently announced third quarter availability for the 7457, before that, it was the fourth quarter.

I was always a bit skeptical of that announcement. It seemed kinda weird for them to bump expectations up from Q4 to Q3 unless they'd had a few huge breakthroughs. that kind of thing only tightens the timeline of expectation and magnifies any slowdown in production. They should have kept the Q4 marker, and then we'd all be pleasantly surprised when Motorola beat the predictions. Now instead of being the hero, they're the goat.

sedarby

Jul 29, 2003, 02:57 PM

Originally posted by Dros
I'm a person for whom the logic you give doesn't apply--computational biology calculations need speed, but not the screen real estate. But, I'm sure Apple knows their target consumers. The limited model set does make it hard for people in the minority, though.

I'm happy that we seem to be approaching a time when the difference between a high and low end processor is getting smaller. Processors running at 300,400 and 500 MHz have a big percentage difference compared to 1.6, 1.8 and 2GHz processor. Of course, putting duals in changes the speed gap as well.

I think the answer goes more to economics. Would Joe Consumer really get it if the 12" costs more than the 17" even if the specs are greater on the 12"? Remember Apple deals in consumer products and must appeal to the lowest common denominator meaning bigger is better rules. It would be nice to have a 15" dual 2 Ghz machine but unless the 17" is at least that powerful it won't happen.

WM.

Jul 29, 2003, 03:39 PM

Originally posted by tazznb
Correct me if I'm mistaken, but isn't a G4 basically a slightly improved G3 with Altivec?
(lost my first reply with login weirdness--it makes no sense to me: I typed in my password, but when I went to post it asked me to login again; usually when I do that I get a "no thread specified" error and lose my post, so I hit "back", and thus lost my post anyway...)

Anyway, the answer is "sort of". I'm not sure about the 603 vs. 604 thing someone mentioned--I have heard that thrown around but I've also seen a rebuttal to it.

But the other key difference is the FSB--the G4 has a different (better) one. No, not in terms of MHz. :) It has some features that make it more efficient than the G3's. You should be able to find out more about this in any Developer Note about a G4 machine (pick one at random).

FYI: the G3's bus is "60x"; the G4's is "MPX" or "MaxBus".

HTH
WM

Phil Of Mac

Jul 29, 2003, 04:00 PM

The G3 has lineage from the 603e. The G4 has lineage from the G3, except with a much better floating point unit and AltiVec. I think 604e was a dead end :)

Originally posted by Lanbrown
Maybe they should just outsource chip production. Ti would be a good source, better fab equipment and better technology as well.

Or IBM? :)

Originally posted by whatever
I don't believe Apple has announced a new Powerbook, so I'm not sure what the delay you referring to is. The current Powerbooks are selling well, so why would they upgrade them for no reason? But then again are you one of those who will complain when Apple releases new G4 Powerbooks and state that they're going to wait for the G5s?

It's rather standard in the computer industry to update your products often. It's called innovation :)

Originally posted by jettredmont
In case you're not joking:

You have a good description of vectors in physics; unfortunately "vectors" in computers are a completely different beast.

In computing terms, a "vector" is an arbitrarily-sized array of values. A "vector unit" will typically deal with vectors of 32-bit ints, 16-bit ints, 8-bit bytes, and 32-bit floating point numbers more efficiently than by dealing with each element of data individually. In Altivec, the vector processing size is 128 bits, so if you have 32-bit ints in your vector the Altivec unit can process those ints four at a time (4x32 = 128). However, you can only do this when you are doing the same thing to every int (for instance, multiplying each by 3). Another name for a vector unit is "SIMD", which stands for "Single Instruction, Multiple Data".

This is a good thing for "multimedia" applications as they often do the same thing on multiple bits of data. Other apps ... well, there's not much opportunity for SIMD in a word processor, for instance.

No, I have no idea why they chose the name "vector" for this, but they did, and we must live with it :) The term persists in several computer languages as well (C++ and Java, for instance).

Ach! Well, I guess that's that then. I would have thought it was really cool if they could calculate physics vectors like that. Well, thanks for the clarification.

Phinius

Jul 30, 2003, 09:58 AM

Originally posted by jaedreth
Apple can't ship a machine without a processor.

Have you ever stopped to think that maybe Apple has had production problems making computers?

Apple can't use G3, the logic board is for G4.

The 12" iBook and 12" PowerBook are almost exactly the same size. There should no technical reason why Apple could not put a G3 inside a 12" PowerBook other than the fact it will run at a much lower frequency and its performance will be much slower.

Apple can't use G5, the boards won't be ready for half a year.

Where does Apple state this? This is obviously all based on rumors.

Apple is stuck over a barrel and it *IS* because of Motorola.

You could also say that Apple is stuck over a barrel with IBM and Motorola. There are two companies that make PowerPC processors and not just one. I would hardly call IBM's 750 processors a match in performance to Motorola's 7500 series chips.

Motorola is just coming out of severe losses for the last two years. Projects were put on hold and cutbacks were made. Apple didn't help the situation any with falling Mac sales.

Motorola promised to ship the chips by mid May. It's now July, almost August, no chips.

No, Motorola had a press release which stated the chip would be sampling in March to major customers and it would go into production in the fourth quarter of this year. That's in writing from Motorola and what you are stating seems to be based on information from rumor sites.

So please refrain from purely derogatory and inflamatory comments that are devoid of any actual facts.

Almost all of what you have stated is based on rumors and not facts.

Phinius

Jul 30, 2003, 10:13 AM

Originally posted by Snowy_River
With regards to these comments, don't you understand that the IBM G5 (PPC970) is Apple's plan B? It is their answer to the 'single point failure'? It was about two years ago that Moto truly demonstrated their failure, when they discontinued their development of the G5 (PPC8500?). That was when (if we are to believe what evidence we have at our disposal) Apple approached IBM to start development of the PPC970.

Jon Rubenstein has publicly stated that Apple started working with IBM three years ago to get the Power4 made into a PowerPC processor. Regardless if Motorola had a next generation of PowerPC processor available, Apple was intent on getting a Power4 derivitive.

It's clear that Apple is trying to move away from Moto entirely.

That is not clear whatsoever. Apple would be foolish to base all of their computer production around a single source for processors. Motorola has plans for a dual-core G4 and there are virtually no other major customers that would be interested in it other than Apple. Motorola would not commit to such a large project unless Apple agreed to buy large quantities of these upcoming processors.

This is something that it's hard to blame IBM for. I think that this is likely the fault of Apple marketing and Moto. Apple marketing has long held that consumer level products shouldn't outperform their Pro-level machines. So, because the G4 was lingering at a maximum of 800 MHz by 2001, there were not going to be any 1GHz G3 to be found. Even now, the fastest PowerBook is at 1GHz, so you don't find anything faster the 900MHz in the iBook (which is now the only G3 processor in the Apple line).

What nonsense. If IBM believed there was a great opportunity to make a large amount of money selling higher performance G3 processors to Apple, then IBM would have produced them. Obviously IBM did not see any great opportunity for fortunes to made selling greatly faster G3 chips only to one customer that had falling sales.

IBM has made the commitment to make a PowerPC version of the Power4 chip due to IBM also wanting to use it their own computers. If Apple would have been the only customer for it, then IBM would be much less interested in making it for the prices that Apple would be willing to pay.

Phinius

Jul 30, 2003, 10:30 AM

Originally posted by jettredmont
Well, IBM has a long history of meeting or beating target dates.

This is an opinion with a long history. If IBM were to miss its date on the 970, that would be forgiven (to some extent) because it will be a relatively new phenomenon and unlikely to happen again.

In October of 2000 IBM announced that the 750FX would hit speeds up to 1 GHz and as of yet there has been no 1 GHz 750FX produced. In fact IBM is now stating that the 750FX will only go to 900 MHz.

You are conveniently ignoring the fact that IBM has up until now been unable to produce a PowerPC processor that can match the speeds of Motorola's PowerPC processors. Yet you claim IBM is the good guy and Motorola is the bad guy. Geez...

Bruja

Jul 30, 2003, 04:34 PM

No matter who's to blame, the bottom line is this:

Apple should've done something to not let this happen. They knew that Moto was jerking them around so they should've had a plan B. Shifting the blame is weak and it makes one wonder if there really was a 15.4" ??

All of us whom are waiting for Apple to release this wonderful Pbk. have been dupped. If Mr. Jobs was so upset and unhappy earlier this year and later last year, then why could'nt he forsee this happening?? To re-coup some of the losses from people changing to PC's maybe Apple should seek an action against Moto for lost revenue becuase the 4th Quarter is going to really hurt them.

Another piece of food for thought: If they decide to release it in Q4, then the technology would still be behind because there are Dell Laptops and (GACK) Gateway Lappers that are ahead of Apple's 15"in terms of features and at a lower price. Apple's product still looks better!

:eek:

Snowy_River

Jul 30, 2003, 04:49 PM

Originally posted by Phinius
What nonsense. If IBM believed there was a great opportunity to make a large amount of money selling higher performance G3 processors to Apple, then IBM would have produced them. Obviously IBM did not see any great opportunity for fortunes to made selling greatly faster G3 chips only to one customer that had falling sales.

IBM has made the commitment to make a PowerPC version of the Power4 chip due to IBM also wanting to use it their own computers. If Apple would have been the only customer for it, then IBM would be much less interested in making it for the prices that Apple would be willing to pay.

Uh.. I don't see how you're not actually agreeing with me. Apple didn't want a faster G3 processor because of the marketing decisions and where it would put it relative to the G4. Thus, IBM didn't make a faster G3.

Yes, of course IBM is planning on using the G5 (970) in their own computers. But, to what extent did they ever use the G3? And, regardless of whether or not IBM plans on using the G5 in their own computers, it is clearly a chip that has been aimed directly at Apple's needs.

jaedreth

Jul 30, 2003, 04:51 PM

Keep in mind, IBM's G5 *is* Plan B. They rushed to IBM when Moto said they couldn't make a G5 because they couldn't work out design flaws. (Does this surprise anyone?) The powerbook issue really could not have a plan b. If Moto gave Apple enough warning that they wouldn't have a G5 at all (like they are under contract to do so), then Apple could have partnered with IBM sooner, and thus they might have been able to get a logic board design for PB G5 sooner than this January, however, Apple didn't have any indication back in mid May of *how* late Moto was going to be on delivery, thus they wouldn't have had the foreknowledge to have rushed development of a PowerBook G5 which could be the only Plan B to this mess, which realistically, was never possible in the first place. Basically, Moto has stuck it to Apple, and BAD.

And just wait until Apple posts profits for Q4. You should see legal action against Moto after said announcement.

Jaedreth

tychay

Jul 30, 2003, 05:33 PM

Originally posted by Phil Of Mac
In the end, yes, but[ Altivec is] not directly designed for that. It's *directly* designed to process vectors. Vectors are numbers with directions, whereas normal numbers without directions are called scalars. A direct application of vectors is in physics.[/B]

Whoa! This is almost like saying that derivatives are mathematical representations of slopes of a curve when we are talking about options and futures.

The vector we are really talking about in Altivec is a nx1 matrix of numbers. In the case of computers n is 128 and the numbers are bits (1 and 0)--or an exchange of these values as we'll see in an example below. Now a vector in physics is a nx1 matrix of numbers. In the case of your example, n is 3 (1 magnitude and 2 direction if in three dimensions). But it can also be 4 (adding time to position, or energy to momentum for Special Relativity), or 6 (3 position and 3 momentum values in advanced classical mechanics), or perhaps 10 (I'm told String Theory uses this, I guess that's the number of dimensions you need so that everything gets symmetric), or sometimes infinity (a "wave function" in Quantum Mechanics is an infinitely-dense, infinite-dimensional "vector" in something called a Hilbert Space). The fact that both physics and computational "vectors" are representable by a nx1 matrix is the only reason they're both called vectors.

The aspect of "vector computing" that is interesting is SIMD, or the ability to apply a single operation on multiple sets of data (in parallel). Lets say we are talking about integers (typically 32 bits) we can perform the same operation of 4 of them at once--a 4x1 "vector" of numbers. If that represents 4 pixels that we are applying some filter to... you see the advantage of being able to perform that operation on 4 pixels of data at a time instead of one at a time: "vector computation". In the case of Velocity/Altivec/VMX (all brand names for the same thing) we can do this processing without cost; in the case of 3DNow/MMX/SSE, we lose the floating point unit to do so (we also have a smaller register space because we have to combine 2 floating point register into a single vector register)--I think you can intuit that vector processing likes a lot of registers to drop intermediate results into.

A program is "vectorizable" to the extent that it can be mapped onto this sort of computing. For example, if the next action the program depends on the result of the previous action, then what is the advantage of being able to perform the same action on multiple sets of data simultaneously? Software which deals with manipulating large data sets (scientific computation, photo editing, video editing, audio processing, rendering) is usually what sees the biggest gain. Deciding whether or not the user clicked on the "OK" or "Cancel" button in Word is not going to see any benefit by vectorization.

In other words, some software we use today (iTunes, iMovie, Photoshop, rendering windows in Mac OS X, etc.) runs faster with such a unit available, but not so for Office and our database-based web application--the last case deals with large amount data, but not manipulating it.

Where physics "magnitude and direction" enters into it... I fail to see.

Take care,

terry

Phil Of Mac

Jul 30, 2003, 07:07 PM

It's already been established that physics vectors and chip design vectors are different things. I had been confused because they were both called "vectors", and that in both cases they were different than "scalars".

That said, perhaps it's time to note that the 970 was supposed to debut at a maximum of 1.8 GHz, instead it started at 2.0 GHz. IBM is overachieving, which is what we all want to see. It's a far cry from anything Motorola ever did ever since cancelling the Book E G5/8500 three years ago.

tychay

Jul 30, 2003, 07:39 PM

Originally posted by Phil Of Mac
It's already been established that physics vectors and chip design vectors are different things. I had been confused because they were both called "vectors", and that in both cases they were different than "scalars".

Yes, so I noticed after the fact. However my post explains why they are both called "vectors" even though they mean different things. Also I distinguish between vector computation (a la supercomputers) and SIMD (a subset of vector computation) as well as the differences between PowerPC's SIMD implementation and the ones in Intel/AMD (known as MMX, SSE, or 3DNow).

That understand is important when comparing p****s size as well as when making a purchase of your next computerr--it's important to understand what exactly you are paying extra bucks for.

Also note that there are significant advantage for vector processing in scientific computation (which is related to your comment and perhaps a cause of your confusion).

One thing I didn't mention is that Intel's compiler (ICC) automatically detects at compile time whether or not vectorization will cause a speedup in code and will auto-vectorize parts of code to take advantage of it. This is why Pentium SPECmarks are so high (Dell and others run SPECmarks on ICC, not gcc like Veritest did for Apple). I imagine that VisualAge (IBM's compiler) does something like this because IBM sits on all the patents on this sort of thing. There is a note in Ars Technica that IBM engineers announced the possibility of such a thing for gcc if the gcc team will accept the patches (this is not foregone as gcc is based on portability so most accepted optimizations are not back end).

However, it does leave the option of Apple supplied patches to gcc so us common folk (with Developer Tools or XCode which uses gcc as its compiler) can get SIMD speed without any extra programming work. Apple already maintains small patches to many open-source code bases (many of which are available on their Darwin website) including gcc so this is a possibility.

Sol

Jul 30, 2003, 08:13 PM

Thank you guys for the explanations on the word 'Vectors'. Now could you could please tell us all why 'Chips' is used to describe computer parts when we all know that the edible chips were here first?

jaedreth

Jul 30, 2003, 08:37 PM

Well, the chips are produced on wafers...

Nabisco-TM *heh*

No clue, really...

Jaedreth

Phinius

Jul 31, 2003, 01:13 AM

Originally posted by Snowy_River
With regards to these comments, don't you understand that the IBM G5 (PPC970) is Apple's plan B?

What utter hotwash! Apple tried from the very beginning of hearing about the Power4 to get it made into a PowerPC processor. It was never 'plan B'. Apple is trying to get the fastest PowerPC processors that can be made.

It was about two years ago that Moto truly demonstrated their failure, when they discontinued their development of the G5 (PPC8500?). That was when (if we are to believe what evidence we have at our disposal) Apple approached IBM to start development of the PPC970.

Apple approached IBM to get a PowerPC version of the Power4 long before Motorola gave any sign of delaying development of the next generation of PowerPC host processors. Even if Motorola did make a new PowerPC processor Apple would still have gone ahead with trying to get IBM to make a cut down version of the Power4 chip.

It's clear that Apple is trying to move away from Moto entirely.

Baloney! Apple will still be buying PowerPC processors from Motorola as long as Apple is still buying PowerPC processors. Apple is simply not going to limit themselves to only one supplier of processors. The same thing was the case with Apple still buying processors from IBM when the G4 came along. The G4 was just as clearly superior in performance to the IBM's G3 as the G5 is to the G4, yet Apple continued to use the G3.

On the plus side, if the rumors of Mohave are correct, then we might well see it in the near future, as IBM will have a 'G4' type chip for Apple to use for it's lower end products.

IBM's 750 series of PowerPC processors are far inferior in performance compared to the 7500 chips from Motorola. The 750FX is up to 900 Mhz on a 130-nm process size and the 7457 will top out at 1.8 GHz, that's double the frequency of the 750FX.

This is something that it's hard to blame IBM for. I think that this is likely the fault of Apple marketing and Moto. Apple marketing has long held that consumer level products shouldn't outperform their Pro-level machines. So, because the G4 was lingering at a maximum of 800 MHz by 2001, there were not going to be any 1GHz G3 to be found. Even now, the fastest PowerBook is at 1GHz, so you don't find anything faster the 900MHz in the iBook (which is now the only G3 processor in the Apple line).

What a hoot! The reason IBM's 750 series of PowerPC processors have not progressed beyond Motorola's is simply due to economics. Apple's sales are too small to justify investing enough money to make substantial upgrades to the architecture of the 750 chips. As for the 970, it's essentially just a cut down version of the Power4 chip. IBM just tacked on Altivec to it. There simply is not enough money to be made in creating a whole new architecture exclusively for Apple's small marketshare.

Phinius

Jul 31, 2003, 01:19 AM

Originally posted by Phil Of Mac
That said, perhaps it's time to note that the 970 was supposed to debut at a maximum of 1.8 GHz, instead it started at 2.0 GHz. IBM is overachieving, which is what we all want to see. It's a far cry from anything Motorola ever did ever since cancelling the Book E G5/8500 three years ago.

Motorola's 7455 processor was supposed to top out at a maximum of 1 GHz. Motorola supplied Apple with 1.42 GHz 7455 processors. That's 42% above the listed maximum frequency. That's also about a tripling of frequency since the G4 processors first became available. Motorola's upcoming 7457 processors are listed for a maximum of 1.33 GHz and yet internal Motorola documents state that it will top out at 1.8 GHz on a 130-nm process. Clearly you are ignoring the reality of the situation.

Also, it took 41 months for PowerMac G4s to go from 500 MHz to 1.42 GHz. That's an increase in frequency of 2.84X. The Pentium 4 increased in frequency 2.1X in 31 months. Clearly Motorola has been advancing the frequency of the PowerPC processors supplied to Apple at a slightly faster pace than Intel has with the Pentium 4. This whole idea of Motorola dropping the ball on this is just a bunch of baloney.

Motorola has not taken in enough money from sales to Apple to justify the torrid pace of capital investments in chip making or processor architecture upgrades that Intel is making, but then no chip manufacture has been able to match the pace that Intel is setting in this regard.

jaedreth

Jul 31, 2003, 02:02 AM

Motorola (http://jasonc.home.texas.net/motorolaLogo.gif) needs to be tarred and feathered, but unfortunately, that's all I could do with GraphicConverter in a short of time as I had before going to bed.

Jaedreth

panphage

Jul 31, 2003, 03:10 AM

phinius, you don't find it even the slightest bit convenient that no G3 has ever been released that equals the clock speed on whatever G4 was in the powerbook at the same time? Even though IBM said it had gotten the G3 to 1ghz two years ago? Do you think apple would allow the ibook (the only apple machine still on the G3) to be running a 1.5ghz G3 while the powerbook was hanging out at 1.0? Did you notice that the second the powerbooks went to 1.0, the ibooks bumped up to 900mhz? No, this doesn't prove anything. But it sure looks suspicious to me. Then again, IBM doesn't have a ton of motivation to spend money improving this chip, something like 200,000 units last year from apple, plus however many sold to upgrade makers (which doesn't seem like it can be a whole lot.)

Sun Baked

Jul 31, 2003, 03:26 AM

Originally posted by Phinius
In October of 2000 IBM announced that the 750FX would hit speeds up to 1 GHz and as of yet there has been no 1 GHz 750FX produced. In fact IBM is now stating that the 750FX will only go to 900 MHz.

You are conveniently ignoring the fact that IBM has up until now been unable to produce a PowerPC processor that can match the speeds of Motorola's PowerPC processors. Yet you claim IBM is the good guy and Motorola is the bad guy. Geez... The G3 at Motorola and IBM are the low power PPC chips, ramping up the speed would eat up the power advantages of these things.

They're ramping the speed up slowly while maintaining the low single digit Watt figures.

If you look at some of the integrated CPUs IBM and Motorola are offering in the PPC line, you'll probably see the power curve start sub 1W -- but they definitely won't be 1+ GHz chips.

Phil Of Mac

Jul 31, 2003, 04:08 AM

I've decided Phinius is actually the Iraqi Information Minister who has been recently employed at Motorola.

Apple asked IBM to make the 970 as a Plan B either in case the Book E 8500 was cancelled by Motorola, or because it was cancelled. The Book E 8500 was cancelled by Motorola. Which is why we have the 970.

Oh, and incidentally, Motorola spent an entire year twiddling their thumbs as the G4 stayed at 500 MHz interminingly. And this was after the G4 was released and then speed-dumped by 50 MHz because an "errata" prevented the G4 from going above 450 MHz. And that was after the G4 was scaled down from the multicore beast it once was into a meek G3 with AltiVec and a better FPU.

And this was after Motorola dropped the ball on G3 production, necessitating that IBM do all the work.

Yes, Apple needs a second chip maker, optimally. But Motorola has failed. Any product using a Motorola processor will likely be delayed and held back. The need to dump Motorola outweighs the need to have a second chip maker.

abdul

Jul 31, 2003, 05:21 AM

i dont know why people say that its not possible at the current time to annonce a g5 pb (no-1s saying release, but just to annonce it) everyone talks about heat dissapation issues and the fact the pm g5 hasnt been released yet (will be by aug/sept)

if apple do want to use a g5 it is possible at the g5 @ 1.2GHz dissapates 19w which is acceptable, no-1 argues with the fact that @1.8GHz is dissaptes too much heat (49w) for a portable. By apple using the G5 would also be great marketing.....here come the g5 the same famly of chip which is in 'the most powerful computer'.

then they can move the ibook to g4 so peeps can have a superdrive and use idvd, keynote and other products which say alvitec recommended.

only if......
:rolleyes:

Phinius

Jul 31, 2003, 10:07 AM

Originally posted by Phil Of Mac

Apple asked IBM to make the 970 as a Plan B either in case the Book E 8500 was cancelled by Motorola, or because it was cancelled. The Book E 8500 was cancelled by Motorola. Which is why we have the 970.

Well why don't you show everyone your inside information about this since you claim to know this. The 8500 was never intended for Apple's use since it is not a host processor. It was and still is listed on Motorola's website as a product. Don't let the facts confuse your incredible judgement.

Also, Motorola's G5 processors were never intended to perform at the same level as the 970. So even if Motorola was now making a G5 processor, Apple would still want the 970 to be the topend chip for the pro Macs.

Motorola designs the host processors used in Apple's computers with the intent of moving them down through the PowerPC line eventually. By selling the PowerPC host processors in other markets, Motorola can keep low prices on the processors that Apple uses. If Motorola designed a processor for Apple's exclusive use, then the cost per chip would have to be raised significantly because of Apple's small market share in personal computers.

Oh, and incidentally, Motorola spent an entire year twiddling their thumbs as the G4 stayed at 500 MHz interminingly.

Do you realize that every chip manufacturer has delays? Intel was originally supposed to release the upcoming Prescott in the third quarter and then it was pushed back to the fourth quarter of this year. IBM's 750FX has not yet been reached 1 GHz as the company had announced in October of 2000.

And that was after the G4 was scaled down from the multicore beast it once was into a meek G3 with AltiVec and a better FPU.

Show us where that is stated. It seems you get most, if not all, of your information from rumor sites.

And this was after Motorola dropped the ball on G3 production, necessitating that IBM do all the work.

The 'G' designation that Motorola uses stands for the generation of the chip architecture. IBM and Motorola design their chips independent of one another. Simply because Apple uses the term G3 or G4 does not indicate that IBM and Motorola have the same chip designs. The only thing common between the IBM and Motorola PowerPC processors is the instruction set and not the chip architecture. This is much like what happens with AMD and Intel desktop processors. What's in common between the two companies designs is the instruction set.

Yes, Apple needs a second chip maker, optimally. But Motorola has failed.

What chip maker would step in and make a processor exclusively for Apple's tiny market share? The answer is none. Apple has to use Motorola for a second PowerPC chip suppler due to the fact there are only two PowerPC processor manufacturers. If Apple turns to another computer processor manufacturer it will more than likely be Intel and that would occur after IBM or Motorola can no longer supply competitive processors to what Intel can offer. Apple would simply be buying yet another off-the-shelf component to make sure that they can compete in personal computers.

Any product using a Motorola processor will likely be delayed and held back.

Show everyone where that has consistently occured with Motorola over the years.

A Motorola executive recently stated that they intend to double the frequency of the PowerPC processors about every 18 months. Judging from other statements that the company has made it would appear that Motorola intends to take the G4 to a listed 2 GHz on the 90-nm process sometime next year. That would occur from about a 35% increase from the process shrink and also up to a 20% increase in frequency with the addition of a low-k dielectric coating. A low-k dielectric was used on some of the G4 chips that were supplied to Apple on the 180-nm process and it will be more extensively used as Motorola gets more experience in using it.

Motorola delayed getting to the 130-nm process and the next generation of PowerPC processors probably due to being severely in the red for several fiscal quarters in a row. It would have been foolish of the company to keep spending at the rate they were and expect to turn a profit. There were cutbacks to get back in the black and the chip making division was not immune to this. Incidently, Apple's falling sales did not encourage Motorola to keep increasingly pump money into making major upgrades to the host PowerPC processors.

It's more than likely that Apple and Motorola worked out a plan where Motorola would use its limited chip making funds to work on increasing the frequency of the G4 chips. Motorola did this by adding SOI, low-k dielectric and boosting the voltage of some of the processors that Apple bought. In the end the G4 on the 130-nm process will be at about 10% lower frequency than the G5 that Motorola had planned. Also because Motorola will be able to shrink the G4 significantly on a 90-nm process, two G4 processors will be put on one chip. If Motorola was making the G5 that would be much less likely to have happened. There is more than one way to obtain performance, having a older chip architecture does not necessarily mean a slower overall performance.

Phinius

Jul 31, 2003, 10:30 AM

Originally posted by Sun Baked
If you look at some of the integrated CPUs IBM and Motorola are offering in the PPC line, you'll probably see the power curve start sub 1W -- but they definitely won't be 1+ GHz chips.

Look on IBM's website, the 750GX will be produced on the current 130-nm process and it will run up to 1.1 GHz.

Man, I don't know where you guys get your misinformation.

Snowy_River

Jul 31, 2003, 03:53 PM

Gee, mighty argumentative, aren't we?

Originally posted by Phinius
IBM's 750 series of PowerPC processors are far inferior in performance compared to the 7500 chips from Motorola. The 750FX is up to 900 Mhz on a 130-nm process size and the 7457 will top out at 1.8 GHz, that's double the frequency of the 750FX.

I just thought that I'd comment on this. This, based on what information we have available, is simply not true. If it were true, then we wouldn't see 800MHz iBooks outperforming 867MHZ PowerBooks in benchmark tests in everything but Altivec tasks. That rather clearly shows that the 750 series is not inferior to, oops, sorry, I was comparing it to a 7400 series chip. My mistake. I guess I don't have access to the specs on the, to date, vapor-ware 7500 series PowerBooks.

For that matter, I haven't even hear of a 7500 series chip. Isn't the latest greatest chip that we know exists (seeing as how you have summarily dismissed any knowledge that comes from rumors) the 7455? Isn't that still a 7400 series chip? Well, given that, I guess a 7500 series chip from Motorola might outperform a 750 chip from IBM... if it ever comes into existence.

But, the simple fact is that even if Moto does get the 7457 chip out, we have no reason to think that it will outperform, per clock cycle, the 7455. And we already know that the 7455 does not outperform, per clock cycle, the 750FX, except in Altivec tasks. So, if a 750VX (Mohave) is produced by IBM with an Altivec compatible vector unit on it, it will likely outperform the 7457 per clock cycle, and might even be capable of reaching higher clock speeds. Further, IBM's plant technology is far advanced over Moto's and they can therefore produce such chips for less. So, Apple would have the option of having a G4 class of chip that could outperform the latest, greatest G4 from Moto for less money.

Finally, I think that I'd like to agree with what others have said. Yes, Apple wants to have more than one supplier, but Moto isn't a good candidate. They have publicly stated that they are getting out of the desktop processor business. And their chip fab capabilities are falling way behind the cutting edge, where Apple needs a chip supplier to be.

Whether Apple chooses to drop Moto or not is not up to us, and we can really only sit by as observers. But, I, for one, think that it is in Apple's best interest to do so.

WM.

Jul 31, 2003, 06:22 PM

Originally posted by panphage
Did you notice that the second the powerbooks went to 1.0, the ibooks bumped up to 900mhz?
Actually, they didn't. They went to 800 MHz when the 867 MHz/1 GHz PowerBooks were introduced, and to 900 MHz about five or six months later.

HTH
WM

macphoria

Jul 31, 2003, 07:02 PM

But, the simple fact is that even if Moto does get the 7457 chip out, we have no reason to think that it will outperform, per clock cycle, the 7455. And we already know that the 7455 does not outperform, per clock cycle, the 750FX, except in Altivec tasks. So, if a 750VX (Mohave) is produced by IBM with an Altivec compatible vector unit on it, it will likely outperform the 7457 per clock cycle, and might even be capable of reaching higher clock speeds. Further, IBM's plant technology is far advanced over Moto's and they can therefore produce such chips for less. So, Apple would have the option of having a G4 class of chip that could outperform the latest, greatest G4 from Moto for less money.
Absolutely.

Finally, I think that I'd like to agree with what others have said. Yes, Apple wants to have more than one supplier, but Moto isn't a good candidate. They have publicly stated that they are getting out of the desktop processor business. And their chip fab capabilities are falling way behind the cutting edge, where Apple needs a chip supplier to be.
I agree. It would be better for Apple to have multiple suppliers, not just for the sake of options, but also to promote bit of competition between suppliers to improve their products which in turn help Apple. But Motorola simply has not been able to step up to the plate and do something.

But as much as I would like Apple to dump Motorola, I wonder what such action would mean for Apple's future. Apple will have to rely solely on IBM's efforts. However at the moment, IBM seems more than capable of handling the task.

jaedreth

Jul 31, 2003, 07:47 PM

IBM is more than capable of handling the tasks...

IBM makes as yet the worlds fastest computers. Nobody can touch them when it comes to their Mainframes. Nobody.

AIX is the most secure and stable System V variant.

Apple would do well to just drop Motorola entirely, and trust in IBM. Especially if further technology exchange were to occur.

There were speculations years ago of kicking Moto out of AIM, and putting in AMD. That's not gonna happen these days...

Nor would it be good for Apple. Not unless Apple were going to make actual PC's that run an AMD version of OS X (not Intel)...

But I don't see Apple going into that market...

Unless Apple did something very devious, and didn't sell Mac OS X for AMD separately in the box... But again, that would still never happen.

So who would replace Motorola? Who could do as good a job as IBM? In my opinion, no one. We don't need no water let the MF burn. Burn MF, Burn...

Jaedreth

Phil Of Mac

Jul 31, 2003, 09:28 PM

Phinius, I suggest you learn how to use quotes.

Originally posted by Phinius
Well why don't you show everyone your inside information about this since you claim to know this. The 8500 was never intended for Apple's use since it is not a host processor. It was and still is listed on Motorola's website as a product. Don't let the facts confuse your incredible judgement.

The 8500 was originally designed to be Apple's G5. It was codenamed Book E and was supposed to have come out a couple years ago. It didn't. Blame Motorola.

Originally posted by Phinius
Also, Motorola's G5 processors were never intended to perform at the same level as the 970.

That's because the 970 was designed at a later date, and intended for a later release. Chips actually perform at higher levels as time passes. It's called "innovation".

Originally posted by Phinius
Do you realize that every chip manufacturer has delays?

Staying at 500 MHz for twelve months isn't a delay, it's a catastophe!

Originally posted by Phinius
Show us where that is stated. It seems you get most, if not all, of your information from rumor sites.

Phinius, I suggest you remind yourself where this forum is hosted.

Originally posted by Phinius
The 'G' designation that Motorola uses stands for the generation of the chip architecture. IBM and Motorola design their chips independent of one another. Simply because Apple uses the term G3 or G4 does not indicate that IBM and Motorola have the same chip designs.

In the case of the G3, they did. The 603, 603e, 604, 604e, and G3 were all designed by IBM and Motorola at a joint facility. and manufactured by both.

Originally posted by Phinius
The only thing common between the IBM and Motorola PowerPC processors is the instruction set and not the chip architecture. This is much like what happens with AMD and Intel desktop processors. What's in common between the two companies designs is the instruction set.

Actually, AMD processors use their own proprietary instruction set. They have an extra unit to translate back into x86. This makes it feasible for AMD to replace Motorola--simply get AMD to produce PowerPC's, and they can sell them as x86 processors by attaching an x86 frontend to the processor.

Originally posted by Phinius
Show everyone where that has consistently occured with Motorola over the years.

1. Motorola promises a May release for the 7457. It's not out yet.
2. The 500 MHz for 12 months debacle.
3. Motorola dropping the ball on G3 production, necessitating that IBM take over all manufacture.
4. Failure to ever provide even compatibility with DDR SDRAM.
5. At initial release, the "errata" holding the G4 back to 450 MHz, necessitating the infamous "speed-dump."
6. Also at initial release, from apple-history.com: "There were extreme supply issues with the G4 initially, due largely to Motorola's inability to deliver the 7400 chips in adequate supply."

That's since Motorola became Apple's single supplier of high-end processors.

Originally posted by Phinius
A Motorola executive recently stated that they intend to double the frequency of the PowerPC processors about every 18 months. Judging from other statements that the company has made it would appear that Motorola intends to take the G4 to a listed 2 GHz on the 90-nm process sometime next year.

Wow. Sometime next year, Motorola will have a slower processor at the same frequency the G5 is at now! And by that time, it'll probably support a 250 MHz bus!

Originally posted by Phinius
There is more than one way to obtain performance, having a older chip architecture does not necessarily mean a slower overall performance.

You don't think a Motorola G4 next year will have a slower overall performance than a G5? The mythical dual-care G4 you keep blathering about still won't keep up with a G5. And if it follows Motorola's past history, it'll probably just be restricted to a single bus at one tenth the frequency, or be stuck at 1.5 GHz because of an errata.

I don't think Motorola could even add DDR support by 2004! Forget about dual-core. Motorola can't even reliably produce a single core processor.

WM.

Jul 31, 2003, 09:34 PM

Originally posted by jaedreth
...an AMD version of OS X (not Intel)...
What's the difference?

WM

daveL

Jul 31, 2003, 09:54 PM

Originally posted by jaedreth
IBM is more than capable of handling the tasks...

IBM makes as yet the worlds fastest computers. Nobody can touch them when it comes to their Mainframes. Nobody.

AIX is the most secure and stable System V variant.

Apple would do well to just drop Motorola entirely, and trust in IBM. Especially if further technology exchange were to occur.

There were speculations years ago of kicking Moto out of AIM, and putting in AMD. That's not gonna happen these days...

Nor would it be good for Apple. Not unless Apple were going to make actual PC's that run an AMD version of OS X (not Intel)...

But I don't see Apple going into that market...

Unless Apple did something very devious, and didn't sell Mac OS X for AMD separately in the box... But again, that would still never happen.

So who would replace Motorola? Who could do as good a job as IBM? In my opinion, no one. We don't need no water let the MF burn. Burn MF, Burn...

Jaedreth
Jaedreth, I disagree with some of your positions.

IBM commercial mainframes are not fast. There are a number of Unix servers in the marketplace that will blow the doors off mainframes, and that is not a new development. IBM mainframes still exist for two reasons: Software and reliability. Software in terms of OS capabilities, especially in terms of virtual machine partitioning and workload management, but also, very importantly, in support of legacy applications. Reliability in terms of dual cores that check each other for errors (only one core's worth of real work gets done), redundant io channels, etc. IBM has been pushing this technology down into their Unix servers, but it still isn't close to parity. I have a feeling that you will come back with the "G5 gov'ment super secret thing", but that's not germane to the topic at hand.

As far as AIX goes, please add "in my opinion" since you cannot prove your claim. Many people (Sun and HP come to mind) would disagree with you, but again nobody can prove their postion, one way or the other; it's a holy war, among many. Sun was actually a close partner with AT&T during the development of the latter releases of System V Unix, having contributed a newly re-architected VM subsystem, among other things. So, to give IBM the Unix crown when IBM was actually *very* late coming to the party is a bit presumptuous. Now, of course, they are embracing Linux, so it's not clear where they are going with AIX.

Trust IBM? Right. Once they have you, they'll make you squeal like a pig. If you don't believe me, ask one of their long term customers. Some of them have been clients of mine. They'll give you the sun, the moon and the star, for free, up front, but they'll rape you on the back end, when you can't get away. It's kinda like a drug dealer. NOTE: This is my opinion and nothing more, and I'm not saying their technology isn't great, it's more their business practices that I'm concerned with.

With regard to AMD, I don't think they would be a bad choice to replace Moto at all. They are pivital in the HyperTransport consortium, of which Apple is a member and, now, a technology consumer. They are well regarded in the industry for their high clock rate designs and interconnect technology. Finally, they are not direct competitors. Bingo, sign'em up! This, in no way, requires, or even suggests, that Apple port to x86.

Well, that's it for me. I hope you understand that I simply disagree. Everyone has their point of view.

macphoria

Jul 31, 2003, 10:08 PM

1. Motorola promises a May release for the 7457. It's not out yet.
2. The 500 MHz for 12 months debacle.
3. Motorola dropping the ball on G3 production, necessitating that IBM take over all manufacture.
4. Failure to ever provide even compatibility with DDR SDRAM.
5. At initial release, the "errata" holding the G4 back to 450 MHz, necessitating the infamous "speed-dump."
6. Also at initial release, from apple-history.com: "There were extreme supply issues with the G4 initially, due largely to Motorola's inability to deliver the 7400 chips in adequate supply."

That's since Motorola became Apple's single supplier of high-end processors.

This is so true. And another proof is the time when Dual Processor Power Macs came out. It came out around the time G4 speed update wasn't going any where. Some do speculate Apple introduced Dual Processor Power Macs in order to keep up with the speed war with PC's and also to keep up its sales.

Phil Of Mac

Jul 31, 2003, 10:51 PM

Originally posted by WM.
What's the difference?

WM

AMD uses a proprietary instruction set they translate to Intel x86.

Mac OS X for the proprietary AMD instruction set wouldn't be nearly as good as an AMD-PPC though.

Phil Of Mac

Jul 31, 2003, 10:56 PM

Originally posted by macphoria
This is so true. And another proof is the time when Dual Processor Power Macs came out. It came out around the time G4 speed update wasn't going any where. Some do speculate Apple introduced Dual Processor Power Macs in order to keep up with the speed war with PC's and also to keep up its sales.

Right. Notice also how they originally made duals only of the lower end processors. The high end processors were too rare to do that with.

Ahh, Motorola. Too bad that such a long-time Apple partner has turned into what it is today.

Adobe75

Jul 31, 2003, 11:35 PM

This makes me want to get into the Mot presentation at the microprocessor forum this fall and just go up there and beat the crap out of the presenter.
On second thought... maybe just a loud obnoxious scene with lots of four letter words would sufice.
Go IBM!!

Snowy_River

Jul 31, 2003, 11:37 PM

Originally posted by Phil Of Mac
Actually, AMD processors use their own proprietary instruction set. They have an extra unit to translate back into x86. This makes it feasible for AMD to replace Motorola--simply get AMD to produce PowerPC's, and they can sell them as x86 processors by attaching an x86 frontend to the processor.

Hmm... I wonder what the possibility of AMD putting a PPC front end on their chips would be...

Phinius

Aug 1, 2003, 01:06 AM

Originally posted by Snowy_River

This, based on what information we have available, is simply not true. If it were true, then we wouldn't see 800MHz iBooks outperforming 867MHZ PowerBooks in benchmark tests in everything but Altivec tasks. That rather clearly shows that the 750 series is not inferior to, oops, sorry, I was comparing it to a 7400 series chip. My mistake. I guess I don't have access to the specs on the, to date, vapor-ware 7500 series PowerBooks.

The 750FX is made on a smaller process size than the 7455 chips. When Motorola produces the 7457 in the next couple of months the 750FX and the 7457 processors will both be made on the same process size. The 7457 chip will have a much higher topend performance than the 750FX because of this.

Oh, and yes I mistakingly stated 7500 series of chips instead of 7450.

But, the simple fact is that even if Moto does get the 7457 chip out, we have no reason to think that it will outperform, per clock cycle, the 7455. And we already know that the 7455 does not outperform, per clock cycle, the 750FX, except in Altivec tasks.

The 7457 will run up to 1.8 GHz on a 130-nm process and the 750FX tops out at 900 MHz. That's twice as much frequency for the 7457 compared to the 750FX. The 750GX is scheduled for production in December and it tops out at 1.1 GHz. The 750 series of chips are simply not a match for the 7450 series from Motorola when both are made on the same process size.

So, if a 750VX (Mohave) is produced by IBM with an Altivec compatible vector unit on it, it will likely outperform the 7457 per clock cycle, and might even be capable of reaching higher clock speeds.

IBM's 750 processors are far behind the performance of Motorola's PowerPC chips when comparing both processors on the same process size and are not likely to catch up anytime soon simply because there is very little financial incentive to make major changes to the chip architecture.

Further, IBM's plant technology is far advanced over Moto's and they can therefore produce such chips for less. So, Apple would have the option of having a G4 class of chip that could outperform the latest, greatest G4 from Moto for less money.

Motorola will have PowerPC processors in production soon in a state of the art fabrication facility owned by one of their chip manufacturing partners. So no, IBM will not have an advantage in this area over Motorola.

Finally, I think that I'd like to agree with what others have said. Yes, Apple wants to have more than one supplier, but Moto isn't a good candidate. They have publicly stated that they are getting out of the desktop processor business. And their chip fab capabilities are falling way behind the cutting edge, where Apple needs a chip supplier to be.

So what chip making company has the wherewithal to step in and replace Motorola for PowerPC chips? No chip manufacture is going to create a PowerPC chip architecture for Apple's exclusive at the prices that Apple would want to pay. Intel's pace of chip advancement and prices have almost completely eliminated any serious long term challengers.

Where has Motorola stated that they are getting out of the desktop processor business? For that matter where have they stated that they are even in the desktop processor business? It seems you might be basing this on rumors.

Phil Of Mac

Aug 1, 2003, 01:18 AM

Originally posted by Phinius
Where has Motorola stated that they are getting out of the desktop processor business? For that matter where have they stated that they are even in the desktop processor business?

They're not? Well, that explains why they can't produce any for Apple! :)

Phil Of Mac

Aug 1, 2003, 01:28 AM

Originally posted by Phinius
IBM's 750 processors are far behind the performance of Motorola's PowerPC chips when comparing both processors on the same process size and are not likely to catch up anytime soon simply because there is very little financial incentive to make major changes to the chip architecture.

In other news, the Intel 486 processor is not nearly as fast as the AMD Opteron. Obviously, Intel is inferior.

Originally posted by Phinius
Motorola will have PowerPC processors in production soon in a state of the art fabrication facility owned by one of their chip manufacturing partners. So no, IBM will not have an advantage in this area over Motorola.

Actually, they will. IBM's Fishkill facility redefined "state of the art" to a point where no other chip maker can reach them.

Originally posted by Phinius
So what chip making company has the wherewithal to step in and replace Motorola for PowerPC chips? No chip manufacture is going to create a PowerPC chip architecture for Apple's exclusive at the prices that Apple would want to pay. Intel's pace of chip advancement and prices have almost completely eliminated any serious long term challengers.

Which is why we need a chip maker that already has a line of processors that can be pressed into service as PowerPC's with slight modifications. IBM with their POWER line fit this. So does AMD.

AMD builds proprietary processors with an x86 frontend. Now either AMD is going to replace the x86 frontend with a PPC frontend, or they're going to simply change the internal architecture of all their processors to PowerPC, and continue to attach an x86 frontend.

Remember, this isn't insane. AMD does have a lot of PPC and RISC technology in their processors already. Would it really be that hard for them to produce a PPC? Wouldn't ANY chip maker be better than Motorola, at this point?

macphoria

Aug 1, 2003, 01:57 AM

Remember, this isn't insane. AMD does have a lot of PPC and RISC technology in their processors already. Would it really be that hard for them to produce a PPC? Wouldn't ANY chip maker be better than Motorola, at this point?
There has been RUMORS about Apple and AMD working together. To what end, that is unclear and total speculation.

However, it is conceivable that Apple is seeking another supplier due to lack of desktop processor development on Motorola's part. And AMD could be an ideal candidate.

But AMD could very well be hesitant (or totally against) getting involved in Apple which holds only small portion of market share. Motorola and IBM have interests in other markets beside desktop, whereas significant part of AMD's business is devoted to desktop. It only makes sense that AMD is not interested. Although AIA (Apple-IBM-AMD) would probably make a great team and something many Mac fans would love to see materialize.

Phinius

Aug 1, 2003, 02:11 AM

Originally posted by Phil Of Mac

The 8500 was originally designed to be Apple's G5.

No, the 8500 series are not even host processors. You can see the 8500 chip series listed on Motorola's website.

It was codenamed Book E and was supposed to have come out a couple years ago.

There was a press release about it a couple of years ago and it is still listed as a product on Motorola's website.

That's because the 970 was designed at a later date, and intended for a later release. Chips actually perform at higher levels as time passes. It's called "innovation".

The 970 is derived from the Power4 processors which are designed for highend servers. The 970 is not an exclusive design for Apple. It has a much higher performance than the Motorola design due to its core design being intended for the server market. IBM would not spend the money to design a chip of this caliber just for Apple's use. Whereas you expect Motorola to do just that and lose money at it trying to keep up with Intel.

Staying at 500 MHz for twelve months isn't a delay, it's a catastophe!

The 750FX was announced in October of 2000 and it has yet to reach the stated maximum of 1 GHz. That's longer than 12 months and yet you have nothing but excuses for IBM letting down Apple by not stepping up and matching or outperforming the Motorola processors in the last few years. Clearly Motorola has created the top performing processors for Apple in the last few years and not IBM.

This makes it feasible for AMD to replace Motorola--simply get AMD to produce PowerPC's, and they can sell them as x86 processors by attaching an x86 frontend to the processor.

AMD won't do that because they are having enough trouble just trying to keep with Intel. It would be a losing proposition for AMD to make a exclusive processor for Apple and tacking a PowerPC and x86 processor together would make a chip that would be a poor compromise for both markets and more expensive to boot due the larger size of it.

1. Motorola promises a May release for the 7457. It's not out yet.

Hogwash! Motorola stated in a press release announcing the 7457 that it would sample to major customers in March and would be produced in the fourth quarter of 2003. I'd expect Apple to announce products that use the chip in the next few weeks.

3. Motorola dropping the ball on G3 production, necessitating that IBM take over all manufacture.

If IBM was so much better at making PowerPC processors than Motorola then why didn't the IBM PowerPC processors pass the Motorola processors in performance in the last few years? And why did Apple reduce the IBM processors to only the iBooks if IBM was clearly the superior manufacturer?

4. Failure to ever provide even compatibility with DDR SDRAM.

The IBM PowerPC processors did not move to DDR memory either.

That's since Motorola became Apple's single supplier of high-end processors.

Now why do you suppose that Apple would have continued with Motorola if the company bungled just about everything as you claim? It's obvious that IBM has had every opportunity to step up to the plate in that time and come out with a PowerPC processor that could match or beat what Motorola was delivering.

Motorola has done such a awful job of designing chips that IBM is using Altivec in the 970. It would seem that since IBM is so superior to Motorola that they could come up with something better than Altivec, now wouldn't they? Or could it be that there is so little money in designing chips exclusively for Apple that IBM had to tack on Altivec?

Wow. Sometime next year, Motorola will have a slower processor at the same frequency the G5 is at now! And by that time, it'll probably support a 250 MHz bus!

Actually next year the 7457RM will support DDRII, have a onboard controller and a bus speed considerably faster than 250 MHz.

The Motorola processors do not need to be as fast as the G5 since they will likely be used in low to mid range Macs.

You don't think a Motorola G4 next year will have a slower overall performance than a G5? The mythical dual-care G4 you keep blathering about still won't keep up with a G5.

What's hilarious about your arguments is you don't seem to have a clue to the fact that the G5 is supposed to be the top performer in Macs and Apple also needs lower cost processors for the low to mid range models also. The G5 will not likely be used throughout Apple's computer models partially due to the higher costs of the chips.

Phinius

Aug 1, 2003, 02:36 AM

Originally posted by Phil Of Mac
In other news, the Intel 486 processor is not nearly as fast as the AMD Opteron. Obviously, Intel is inferior.

Intel stopped making 486 processors years ago. IBM has had every opportunity to pass Motorola in host PowerPC processor performance and have failed up until now.

Actually, they will. IBM's Fishkill facility redefined "state of the art" to a point where no other chip maker can reach them.

That's hilarious. Your blind to what the other chip manufacturers are doing, which includes Intel.

Which is why we need a chip maker that already has a line of processors that can be pressed into service as PowerPC's with slight modifications. IBM with their POWER line fit this. So does AMD.

AMD is as far removed from having a design based on the PowerPC instruction set as you can get. It simply would not be financially feasible for AMD to make a PowerPC design for Apple's tiny market share. The only alternative that Apple has is converting over to the X86 platform.

Remember, this isn't insane. AMD does have a lot of PPC and RISC technology in their processors already. Would it really be that hard for them to produce a PPC?

It would take years for AMD to come up with a chip design for Apple. The 970 was worked on for three years before it was announced.

Wouldn't ANY chip maker be better than Motorola, at this point?

Apple doesn't have much choice at this point. There are very few chip manufacturers that design chips for computers. There is Sun, IBM, Intel, AMD, Fujitsu. Motorola has designs that are not exclusive to computers. Both Sun and AMD are losing buckets full of money and have a shaky future as chip designers or manufacturers. So, that leaves Fujitsu, Intel, Motorola or IBM. Since Fujitsu makes processors based on Sun's Ultrasparc, then it's doubtful Fujitsu has a big future in computer processors. Realistically it's either Motorola or IBM for PowerPC processors and Intel would be the obvious choice if Apple cannot increase market share significantly in the increasingly commodity based personal computers.

Snowy_River

Aug 1, 2003, 04:11 AM

Gee, this is getting to be fun... :D

No, wait, it's getting to be mundane, too... ;)

Originally posted by Phinius
No, the 8500 series are not even host processors. You can see the 8500 chip series listed on Motorola's website.

Well, not according to the PPC roadmap that they published... in '99 if memory serves. I don't have a copy of it any more, and I know it's long since been pulled from their website (like back in 2000). But I do recall that they showed the PPC8500 "G5" projected release in 2001.

The 750FX was announced in October of 2000 and it has yet to reach the stated maximum of 1 GHz. That's longer than 12 months and yet you have nothing but excuses for IBM letting down Apple by not stepping up and matching or outperforming the Motorola processors in the last few years. Clearly Motorola has created the top performing processors for Apple in the last few years and not IBM.

If IBM was so much better at making PowerPC processors than Motorola then why didn't the IBM PowerPC processors pass the Motorola processors in performance in the last few years? And why did Apple reduce the IBM processors to only the iBooks if IBM was clearly the superior manufacturer?

Okay, now I've offered an explanation to this before, but you've never bothered to say anything in response. Apple didn't want a faster G3 because they didn't have faster G4s! Now, I've heard that IBM has, over time, been working on their equivalent of Altivec, and trying to build it into the 750 line, but that's not a trivial task. Until they could do that, Apple wouldn't buy faster processors, so IBM wouldn't produce them. Now IBM has a chip with an Altivec equivalent, and Apple is moving to it quickly.

Motorola has done such a awful job of designing chips that IBM is using Altivec in the 970. It would seem that since IBM is so superior to Motorola that they could come up with something better than Altivec, now wouldn't they? Or could it be that there is so little money in designing chips exclusively for Apple that IBM had to tack on Altivec?

I, for one, never said that Moto wasn't any good at designing chips. They used to have a top flight PPC design team. But they've fired most of them. Many of them have even been hired by Apple. Now, Apple has made a substantial investment in Altivec, and would suffer greatly if they didn't maintain Altivec support in whatever chip they chose to use, both in PR and in performance. So, it's no great mystery as to why IBM is working with an Altivec equivalent vector unit.

Actually next year the 7457RM will support DDRII, have a onboard controller and a bus speed considerably faster than 250 MHz.

Ah, more vapor-ware. Weren't you the one who was putting people down for basing their opinions on rumors?

Apple doesn't have much choice at this point. There are very few chip manufacturers that design chips for computers...
<snip>
...Realistically it's either Motorola or IBM for PowerPC processors and Intel would be the obvious choice if Apple cannot increase market share significantly in the increasingly commodity based personal computers.

Okay, so by your own argument, Apple doesn't have a whole lot of choices here. Therefore, if it is Apple's judgement that Moto is a high-risk supplier, they will be forced to rely on IBM for their PPC chips. And, if IBM falls through, there's always the Marklar project as a 'Plan B'.

Snowy_River

Aug 1, 2003, 04:17 AM

Originally posted by Phinius
The G5 will not likely be used throughout Apple's computer models partially due to the higher costs of the chips.

Last words...

If what information we have at our disposal is to be believed, then the G5 is cheaper than the G4, and, therefore, your argument is fundamentally flawed.

You know, Phinius, I don't know why you're such a Motorola fan, but that's fine. We clearly have some areas where we disagree. I think that it's time to leave it at that. I'm tired of trading jabs with you. As I said in my previous post, it's gotten mundane. Until we meet in another thread to trade jabs there, it's been real! ;)

Phinius

Aug 1, 2003, 10:18 AM

Originally posted by Snowy_River
Well, not according to the PPC roadmap that they published... in '99 if memory serves. I don't have a copy of it any more, and I know it's long since been pulled from their website (like back in 2000). But I do recall that they showed the PPC8500 "G5" projected release in 2001.

You need to check Motorola's website. The 8500 PowerPC series are not host proessors.

Okay, now I've offered an explanation to this before, but you've never bothered to say anything in response. Apple didn't want a faster G3 because they didn't have faster G4s!

These are some of the most lame arguments I have ever heard. Motorola was doing a terrible job, but yet Apple didn't want IBM to pass the performance of the G4 chips. Geez...now that makes a lot of sense. Do you realize how incredibly stupid that argument is.

Now, I've heard that IBM has, over time, been working on their equivalent of Altivec, and trying to build it into the 750 line, but that's not a trivial task. Until they could do that, Apple wouldn't buy faster processors, so IBM wouldn't produce them. Now IBM has a chip with an Altivec equivalent, and Apple is moving to it quickly.

The 970 has Altivec tacked onto to it. It is not a IBM design. That got it from Motorola.

So, it's no great mystery as to why IBM is working with an Altivec equivalent vector unit.

Perhaps IBM is using Altivec because they were unwilling or unable to come up with something better.

Okay, so by your own argument, Apple doesn't have a whole lot of choices here. Therefore, if it is Apple's judgement that Moto is a high-risk supplier, they will be forced to rely on IBM for their PPC chips.

Motorola is coming out with a dual-core G4 chip and Apple will use it, period. There are no other major markets that would want a dual-core chip that uses a lot of watts.

Apple will not go with only one supplier. This is just what happened when Motorola supplied the G4 and it will occur again with the use of IBM's G5

macphoria

Aug 1, 2003, 10:48 AM

You need to check Motorola's website. The 8500 PowerPC series are not host proessors.
There was a speculation at one point.
http://www.theregister.co.uk/content/39/21692.html

These are some of the most lame arguments I have ever heard. Motorola was doing a terrible job, but yet Apple didn't want IBM to pass the performance of the G4 chips. Geez...now that makes a lot of sense. Do you realize how incredibly stupid that argument is.
Many believe IBM's 750FX G3 is already capable of running at 1GHz. People speculate the reason for it not appearing in Apple's product, iBook which is only Apple product with G3, is because they don't want iBook cannibalizing PowerBook sales which runs at 1Ghz also. Apple constantly made sure none of iBook's clock speed topped PowerBook's, although there were some instances like iBook G3 700Mhz vs PowerBook G4 667 or current iBook 900Mhz vs PowerBook G4 867.

The 970 has Altivec tacked onto to it. It is not a IBM design. That got it from Motorola.
AltiVec is Motorola's version of SIMD (Single Instruction Multiple Data). 970 uses its own version of SIMD.

Perhaps IBM is using Altivec because they were unwilling or unable to come up with something better.
IBM never approved of AltiVec.
http://www.macworld.com/2001/08/bc/buzzstandbycpu/

Motorola is coming out with a dual-core G4 chip and Apple will use it, period. There are no other major markets that would want a dual-core chip that uses a lot of watts.
If and when it does, Apple will probably use it. But their constant delay is not going to make anyone happy. Apple is trying hard to catch up to PC's in performance. G5 narrowed the gap. Other Apple products need to catch up as well, especially ones using G4.

Apple will not go with only one supplier. This is just what happened when Motorola supplied the G4 and it will occur again with the use of IBM's G5
They probably won't go with single supplier, for having only one option is a dangerous game to play. But you have to wonder how happy Apple is with Motorola and whether Apple is looking for other suppliers which unfortunately is none at the moment.

bcsimac

Aug 1, 2003, 02:19 PM

Back in 1998 and 1999, Motorola had roadmap that clearly showed G3 (750 series), G4 (7400 series), and G5 (7500) series. On that roadmap, the G5 was supposed to have a multi-core architecture, 64bit processing, multiple Altivec units, and Book E specs. Later on in late 1999 to early 2000, Motorola changed that roadmap by calling the G5 the 8500 with still the same specs and features previously listed for 7500 (G5). In 2002, the 8500 became listed as a non-host processors and became known as embedded processors for routers and such. We were also supposed to see 7460, 7470's and etc......and we still haven't. Motorola removed the references to 7460 and 7470. Now they have the 7457 as new high speed processor. From 1998-2002, all the talk was about the 7500 later changed to 8500 series G5's that would be able to go as fast as 5+ Ghz. As we got closer to 2002, the talk quickly changed to idea that either Apple or Motorola scrapped the G5 project.....the truth of which is unknown.

daveL

Aug 1, 2003, 02:32 PM

Snowy_River

Aug 1, 2003, 03:50 PM

Originally posted by bcsimac
Back in 1998 and 1999, Motorola had roadmap that clearly showed G3 (750 series), G4 (7400 series), and G5 (7500) series. On that roadmap, the G5 was supposed to have a multi-core architecture, 64bit processing, multiple Altivec units, and Book E specs. Later on in late 1999 to early 2000, Motorola changed that roadmap by calling the G5 the 8500 with still the same specs and features previously listed for 7500 (G5). In 2002, the 8500 became listed as a non-host processors and became known as embedded processors for routers and such. We were also supposed to see 7460, 7470's and etc......and we still haven't. Motorola removed the references to 7460 and 7470. Now they have the 7457 as new high speed processor. From 1998-2002, all the talk was about the 7500 later changed to 8500 series G5's that would be able to go as fast as 5+ Ghz. As we got closer to 2002, the talk quickly changed to idea that either Apple or Motorola scrapped the G5 project.....the truth of which is unknown.

Originally posted by daveL
You might find this interesting:

http://www.bayarea.net/~kins/AboutMe/CPUs.html

I'm glad to see that I'm not the only one with a memory of what Moto has said in the past. Of some note:

http://physics.arizona.edu/~wathen/picture.jpg

Now, I don't know that I'd believe some of the SPEC numbers that that site quotes, but then again, they were coming from rumor sites not of the highest repute. (According to those numbers, the 8500 would have been faster per cycle and at higher clock speeds two years ago than the 970 is now. Seems like a bit of an exageration to me.)

However, the fact that the 8500 was going to be a host processor, and likely the base for the Power Mac G5, before that branch of development was cancelled, is remembered by more people than just me.

bcsimac

Aug 1, 2003, 04:55 PM

Yeah, Same feeling here. In fact I remember Motorola saying the 7500 aka G5 would have 4 cores as the above mentioned multicore support. This thing was supposed to kick everybodies rear and make intel people wet their pants. Too bad it went by the way side. MacOS Rumors, the Register, and many others talked about this processor. There was posts in forums all over the place. It would have been funny to see intel people wet their pants though!

Originally posted by Snowy_River
I'm glad to see that I'm not the only one with a memory of what Moto has said in the past. Of some note:

http://physics.arizona.edu/~wathen/picture.jpg

Now, I don't know that I'd believe some of the SPEC numbers that that site quotes, but then again, they were coming from rumor sites not of the highest repute. (According to those numbers, the 8500 would have been faster per cycle and at higher clock speeds two years ago than the 970 is now. Seems like a bit of an exageration to me.)

However, the fact that the 8500 was going to be a host processor, and likely the base for the Power Mac G5, before that branch of development was cancelled, is remembered by more people than just me.

Phil Of Mac

Aug 1, 2003, 06:32 PM

Originally posted by the Motorola Information Minister
The 750FX was announced in October of 2000 and it has yet to reach the stated maximum of 1 GHz. That's longer than 12 months and yet you have nothing but excuses for IBM letting down Apple by not stepping up and matching or outperforming the Motorola processors in the last few years. Clearly Motorola has created the top performing processors for Apple in the last few years and not IBM.

Apple spent a good deal of time and money moving to Altivec. Moving back to the G3 wouldn't work, especially if they had the 970 to wait for.

IBM didn't produce G3's faster than Motorola's G4's because Apple didn't want them to. You don't bother producing a product no one's going to buy, especially when you're already working on the 970.

Originally posted by the Motorola Information Minister
AMD won't do that because they are having enough trouble just trying to keep with Intel. It would be a losing proposition for AMD to make a exclusive processor for Apple and tacking a PowerPC and x86 processor together would make a chip that would be a poor compromise for both markets and more expensive to boot due the larger size of it.

An AMD PPC would be a variant of the procesors they're already developing. AMD chip design allows that flexibility. Remember, AMD processors now are an AMD processor with an x86 frontend tacked on. If that AMD processor was a PowerPC, or if they replaced the x86 frontend with a PPC frontend or something, it would be very feasible.

Originally posted by the Motorola Information Minister
If IBM was so much better at making PowerPC processors than Motorola then why didn't the IBM PowerPC processors pass the Motorola processors in performance in the last few years?

Uh, that's what the 970 is, chief.

Originally posted by the Motorola Information Minister
The IBM PowerPC processors did not move to DDR memory either.

The 970 did.

Originally posted by the Motorola Information Minister
Now why do you suppose that Apple would have continued with Motorola if the company bungled just about everything as you claim? It's obvious that IBM has had every opportunity to step up to the plate in that time and come out with a PowerPC processor that could match or beat what Motorola was delivering.

Once again, they did. It's called the 970.

Originally posted by the Motorola Information Minister
Motorola has done such a awful job of designing chips that IBM is using Altivec in the 970. It would seem that since IBM is so superior to Motorola that they could come up with something better than Altivec, now wouldn't they? Or could it be that there is so little money in designing chips exclusively for Apple that IBM had to tack on Altivec?

Motorola developed AltiVec in the 90's and didn't do much since. If we're going by that criteria, then I guess IBM is superior because they invented the PowerPC in the early 90's :)

Originally posted by the Motorola Information Minister
What's hilarious about your arguments is you don't seem to have a clue to the fact that the G5 is supposed to be the top performer in Macs and Apple also needs lower cost processors for the low to mid range models also. The G5 will not likely be used throughout Apple's computer models partially due to the higher costs of the chips.

Are they really that much more expensive? Bought in the bulk quantities Apple's buying them? Have you even read about the Fishkill facility? IBM can mass produce the 970 or any other processor on the cheap, domestically. No foreign outsourcing. In fact, other chip makers are starting to outsource to IBM!

The 970 costs whatever IBM and Apple agree on. If they need to produce them in the millions to go into all product lines, they will.

Phil Of Mac

Aug 1, 2003, 06:42 PM

Originally posted by the Motorola Information Minister
IBM has had every opportunity to pass Motorola in host PowerPC processor performance and have failed up until now.

No one ever wanted to make a G3 that was faster than a G4. Even though it's been done at similar clock speeds, the G3 was never supposed to be faster than the G4, and Apple never ordered a G3 that was faster than the G4. They asked IBM for a G5 instead. And that's what they got.

Apple wasn't about to go back and lose Altivec.

Originally posted by the Motorola Information Minister
That's hilarious. Your blind to what the other chip manufacturers are doing, which includes Intel.

I don't doubt other chip makers will want to develop facilities similar to or better than Fishkill. But they haven't yet.

Originally posted by the Motorola Information Minister
AMD is as far removed from having a design based on the PowerPC instruction set as you can get. It simply would not be financially feasible for AMD to make a PowerPC design for Apple's tiny market share. The only alternative that Apple has is converting over to the X86 platform.

It would be a whole lot more feasible than relying on Motorola.

Originally posted by the Motorola Information Minister
It would take years for AMD to come up with a chip design for Apple. The 970 was worked on for three years before it was announced.

Perhaps they've already started. They already have a lot of PPC technology. It wouldn't be that unfeasible.

Originally posted by the Motorola Information Minister
Apple doesn't have much choice at this point. There are very few chip manufacturers that design chips for computers. There is Sun, IBM, Intel, AMD, Fujitsu. Motorola has designs that are not exclusive to computers. Both Sun and AMD are losing buckets full of money and have a shaky future as chip designers or manufacturers.

Motorola has shafted Apple more than enough times already. They are unreliable. So mark them off the list.

Sun wants nothing to do with Apple because of past history. AMD, however, might see Apple as a source of turnaround.

There is another option. Apple itself reportedly has many ex-Motorola chip designers. I don't consider it totally unfeasible that Apple themselves could design processors for IBM to manufacture at Fishkill.

Originally posted by the Motorola Information Minister
So, that leaves Fujitsu, Intel, Motorola or IBM. Since Fujitsu makes processors based on Sun's Ultrasparc, then it's doubtful Fujitsu has a big future in computer processors. Realistically it's either Motorola or IBM for PowerPC processors and Intel would be the obvious choice if Apple cannot increase market share significantly in the increasingly commodity based personal computers.

Who needs to increase market share? You don't see Mercedes-Benz fretting about their infinestimal market share.

Phil Of Mac

Aug 1, 2003, 06:49 PM

Originally posted by the Motorola Information Minister
These are some of the most lame arguments I have ever heard. Motorola was doing a terrible job, but yet Apple didn't want IBM to pass the performance of the G4 chips. Geez...now that makes a lot of sense. Do you realize how incredibly stupid that argument is.

They didn't want a non-Altivec chip that was faster than an Altivec chip. They chose instead for IBM to design an Altivec chip that was faster. The 970.

Originally posted by the Motorola Information Minister
The 970 has Altivec tacked onto to it. It is not a IBM design. That got it from Motorola.

It is Altivec compatible. But then again, Motorola has been using PowerPC designs ever since 1994. That they got from IBM.

Originally posted by the Motorola Information Minister
Perhaps IBM is using Altivec because they were unwilling or unable to come up with something better.

IBM is using an Altivec-compatible SIMD unit. It is not the same as Altivec.

Originally posted by the Motorola Information Minister
Motorola is coming out with a dual-core G4 chip and Apple will use it, period. There are no other major markets that would want a dual-core chip that uses a lot of watts.

Vaporware.

Originally posted by the Motorola Information Minister
Apple will not go with only one supplier. This is just what happened when Motorola supplied the G4 and it will occur again with the use of IBM's G5

The reason it happened with Motorola is because Motorola couldn't produce enough G4's to fill the entire product line. IBM has no problem producing enough chips at Fishkill.

Snowy_River

Aug 1, 2003, 07:37 PM

Originally posted by the Motorola Information Minister
http://smilies.crowd9.com/contrib/blackeye/lol.gif

LOL!!!

macphoria

Aug 1, 2003, 07:42 PM

LOL. That is hilarious. Didn't notice that.

bcsimac

Aug 1, 2003, 08:13 PM

Yeah Isn't it! Sort of a off-shoot of the Minister of Misinformation of Iraq...hee hee!....opps I forgot....He wasn't misinforming......oh no .......he was telling the truth.....he swears by Allah!

Originally posted by macphoria
LOL. That is hilarious. Didn't notice that.

tychay

Aug 1, 2003, 08:21 PM

Originally posted by Phil Of Mac
I don't doubt other chip makers will want to develop facilities similar to or better than Fishkill. But they haven't yet.

This is not quite true. Looking at Apple's description of FishKill (http://www.apple.com/g5processor/ibmprocess.html), I should note that all 300mm fabs are fully robotic and all chips with copper interconnects use the damascene process, and the AMD chips already use 9-layers, which, is not an inherent advantage.

For instance, on another thread someone pointed out to me that Intel has a 300mm fab in New Mexico (http://www.eetimes.com/semi/news/OEG20021023S0028). Of course, most people conveniently forget to mention that this fab has yet to ship a single Pentium out the door. :)

Companies like Intel, Texas Instruments and Motorola (yes, even them) have 300mm fabs in the works. If the market stays this bad, we may be waiting some time before they go on line.

RAM manufacturers have been producing on 90nm which is a step above what the G5 is fabbed at (The Intel Prescott and IBM 980 are rumored to using this process). The problem is it is much easier to make a work out lithography issues for a component with a lot of repeatability like RAM vs. something like a CPU or ASIC.

As for IBM taking on other clients (like Apple), that was the stated idea behind the building of Fishkill and was used to justify costs--they'd have been poorly put out if Fishkill didn't have other customers. Stealing away nVidia from some Taiwanese company, was a major coup and does speak volumes as to how well done Fishkill is.

I won't comment on your other points made against the Motorola Information Minister. :)

tychay

Aug 1, 2003, 08:29 PM

Originally posted by Phil Of Mac
IBM is using an Altivec-compatible SIMD unit. It is not the same as Altivec.

Whoa! The only reason IBM's SIMD unit isn't "Altivec" is because "Altivec" is a trademark of Motorola--just like VMX is of IBM and "Velocity Engine" is for Apple--three names for the same thing. The actual implementation may be slightly different, of course, but the instruction set is the same.

I seem to remember that Altivec technology was developed at the joint development center by all three companies together (IBM was later to pull out but I think this was after the work was done). Rumor has it that IBM and Apple pressured a reluctant Motorola to put it in the G4 because IBM was moving out of the desktop processor market into servers and Apple wanted it for their PowerMacintosh line. Not sure if this is true.

Phil Of Mac

Aug 1, 2003, 09:26 PM

Motorola's Altivec and IBM's Altivec are still slightly different. The instruction counts, at least, are different. Minor difference, though.

So Intel has a fully automated facility as well. I think Fishkill has other advantages--producing chips from a larger wafer size, for instance. But the fact is, IBM already has the facility, Motorola doesn't yet. And if Motorola builds production facilities the same way they build processors, it won't be here for awhile :)

Phinius

Aug 2, 2003, 01:47 PM

Originally posted by macphoria

There was a speculation at one point.
http://www.theregister.co.uk/content/39/21692.html

Motorola has a PowerPC roadmap (http://e-www.motorola.com/files/archives/doc/roadmap/PPCRMAP.pdf), dated 11/30/2001, that clearly mentions the G5 as a 85XX processor. If you look at the bottom of the roadmap there is an explanation of each processor number. Motorola states that the 85XX (G5) is a 'integrated processor targeting the communication and consumer markets'. The 7XXX (G4) processor is a 'high performance microprocessor targeting computing and high-end embedded'. The processors that Apple uses are not integrated processors and so therefore the G5 on the roadmap is not intended for Apple's use in computers.

If you then look at the PowerPC processors products listed on Motorola's website you will see a category for 8XXX integrated host processors and 85XX integrated communications and 82XX processors.

The 8500 series of G5 chips was never intended for Apple's market. A 8500 processor for Apple is based on wild assed speculation and rumors that are completely unfounded.

Although much of the features on the G5 roadmap will be brought over to the 7457RM G4 processors in 2004.

Many believe IBM's 750FX G3 is already capable of running at 1GHz. People speculate the reason for it not appearing in Apple's product, iBook which is only Apple product with G3, is because they don't want iBook cannibalizing PowerBook sales which runs at 1Ghz also. Apple constantly made sure none of iBook's clock speed topped PowerBook's, although there were some instances like iBook G3 700Mhz vs PowerBook G4 667 or current iBook 900Mhz vs PowerBook G4 867.

Do you realize that both the eMac and iMac use a G4 processor that runs at 1 GHz? The idea that IBM didn't release a G3 running at 1 GHz because of having to keep it below the frequency of the G4 is ludicrous. The G4 would obviously have a higher performance than than G3 when both are running at the same frequency due to the G4 having Altivec and also the availability of a L3 cache for the G4.

AltiVec is Motorola's version of SIMD (Single Instruction Multiple Data). 970 uses its own version of SIMD.

The 970 (http://www-3.ibm.com/chips/techlib/techlib.nsf/techdocs/FBFA164f824370f987256d6a006f424d) has Altivec.

IBM never approved of AltiVec.
http://www.macworld.com/2001/08/bc/buzzstandbycpu/

Really! Well, Altivec is used on the 970. So much for IBM disapproving of it. Seems Motorola just came up with a feature that was better than what IBM had to offer.

If and when it does, Apple will probably use it. But their constant delay is not going to make anyone happy. Apple is trying hard to catch up to PC's in performance. G5 narrowed the gap. Other Apple products need to catch up as well, especially ones using G4.

How about the G3 iBooks catching up with the PC in performance. Or is it too much to ask for IBM to do the same thing with the G3 that is expected of Motorola's G4?

Apple has fallen behind in performance compared to PCs due to a falling market share. If you believe that is mainly due to Motorola, then ask yourself why is it that Apple's CFO Fred Anderson stated in the last fiscal quarterly report with analysts (webcast available on Apple's website) that he did not believe the G5 PowerMacs would ever obtain the peak quarterly sales rate of the G4 Power Mac. In fact he stated that it would probably be over 200,000 units per quarter and never reach 300,000 (G4 Power Mac sales reached 370,000 units in one quarter).

They probably won't go with single supplier, for having only one option is a dangerous game to play. But you have to wonder how happy Apple is with Motorola and whether Apple is looking for other suppliers which unfortunately is none at the moment.

If Apple goes with a single supplier it will more than likely be Intel. AMD is mentioned as a possible savior for Apple, but AMD has had months of delays getting Opteron out and the company has been running deeply into the red for several quarters.

It is very unlikely that IBM would be Apple's sole supplier with 9XX processors anytime soon because to do that IBM would have to be willing to sell some 9XX processors to Apple for less than $100 apiece in order for Apple to compete against the Celeron processors. IBM is undoubtedly selling Apple the 970s for several hundred dollars each at this time. In order for Apple to compete in the sub $1,000 computer market the company needs low priced processors and the G4 is likely to be one of those choices. Or do you believe that Apple should abandon the sub $1,000 computer market and simply have a much smaller market share with only elite performing computers?

The G3 seems destined to still be used by Apple at least through the first 6 months of 2004 with the 750GX just recently announced. What product the 750GX will be used in is difficult to tell since it would be a long dry spell for the iBooks to be bumped up to 1.1 GHz in January 2004 and the current top speed of the iBooks tops out at 900 MHz.

Phinius

Aug 2, 2003, 02:42 PM

Originally posted by Phil Of Mac
Apple spent a good deal of time and money moving to Altivec. Moving back to the G3 wouldn't work, especially if they had the 970 to wait for.

[quote]IBM didn't produce G3's faster than Motorola's G4's because Apple didn't want them to. You don't bother producing a product no one's going to buy, especially when you're already working on the 970.

Right, no one would be interested in buying G3 iBooks running at a higher frequency than 900 Mhz, that would make it much less competitive. Geez....do you even stop to think your statements through before posting?

An AMD PPC would be a variant of the procesors they're already developing. AMD chip design allows that flexibility. Remember, AMD processors now are an AMD processor with an x86 frontend tacked on. If that AMD processor was a PowerPC, or if they replaced the x86 frontend with a PPC frontend or something, it would be very feasible.

You mean to say go with the chip manufacturer that had repeated delays getting the Opteron processor to market? Isn't getting processors out on time the main reason you believe that Apple should dump Motorola? Now you advocate going with a chip manufacturer that has a recent history of delays that are far worse than what Motorola is exhibiting.

Uh, that's what the 970 is, chief.

No Mac has been sold yet with with a 970 inside. Apple at this point is taking preorders for G5 Power Macs.

The 970 did.

The 970 is the future and not the present or past. There are no Macs sold yet with a 970 inside.

Once again, they did. It's called the 970.

Once again, the 970 is not available yet in a Mac. Apple states a G5 Power Mac will be available starting later this month.

Motorola developed AltiVec in the 90's and didn't do much since. If we're going by that criteria, then I guess IBM is superior because they invented the PowerPC in the early 90's :)

Motorola added SOI and a low k-dielectric process since then.

Also, can you name a SIMD process by IBM or Intel that is superior in performance to Altivec? Quite simply Altivec has been Apple's saving grace in performance comparison to Wintel computers.

The PowerPC was not invented by IBM alone. It was a joint collaboration by Motorola, IBM and Apple.

Are they really that much more expensive? Bought in the bulk quantities Apple's buying them? Have you even read about the Fishkill facility? IBM can mass produce the 970 or any other processor on the cheap, domestically. No foreign outsourcing. In fact, other chip makers are starting to outsource to IBM!

The average price of a desktop PC currently sold is about $749 according to a recent market research survey. Quite a lot of those computers are using processors that have a list price of under $70 from Intel.

It's quite possible that the 970 costs Apple less than the 1.42 GHz 7455 processors. However, the 970 is almost certainly costing Apple several hundred dollars apiece. That is a long ways from the under $70 that Intel gets for Celeron processors.

IBM is not likely to match or beat the prices of the lowend 7457 G4 processors in the next year or so. That's due to IBM targeting a higher performance market and in doing so the company can ask higher prices. If IBM lowered the price of some 970 processors to $70 or so, then that would undermine the demand for the much higher priced 970 processors.

The 970 costs whatever IBM and Apple agree on. If they need to produce them in the millions to go into all product lines, they will.

A bigger processor with a larger die size usually costs more to produce than a smaller die size. By virtue of the G3 being around 33 mm^2 in size it will naturally cost less to produce than the 970 which is about 117 mm^2. The 7457 will be a little bit smaller than the 970, so it could quite possibly be cheaper to manufacture than the 970.

For IBM to make G5 processors that are sold in every Mac that would mean there must be some sold to Apple for well under $100. Either that or you are advocating that Apple abandon the sub $1,000 computer market. It's not likely that IBM will be selling G5 processors to Apple for anywhere close to lowend Celeron prices anytime soon. Because of that alone Apple needs some other PowerPC processors to fill in on the low priced Macs and that is likely to be the G3 and G4. If you believe that the G3 running at about 1 GHz can do the trick for desktop and notebook Macs competing against Celeron chips running at well over 2 GHz, then you are in a fantasy world.

Apple needs Motorola and the G4 processors for quite some time to come. Judging by what Motorola has planned for the G4 it looks like Apple intends to use it through at least 2004.

macphoria

Aug 2, 2003, 03:35 PM

Motorola has a PowerPC roadmap, dated 11/30/2001, that clearly mentions the G5 as a 85XX processor. If you look at the bottom of the roadmap there is an explanation of each processor number. Motorola states that the 85XX (G5) is a 'integrated processor targeting the communication and consumer markets'. The 7XXX (G4) processor is a 'high performance microprocessor targeting computing and high-end embedded'. The processors that Apple uses are not integrated processors and so therefore the G5 on the roadmap is not intended for Apple's use in computers.

If you then look at the PowerPC processors products listed on Motorola's website you will see a category for 8XXX integrated host processors and 85XX integrated communications and 82XX processors.

The 8500 series of G5 chips was never intended for Apple's market. A 8500 processor for Apple is based on wild assed speculation and rumors that are completely unfounded.

Although much of the features on the G5 roadmap will be brought over to the 7457RM G4 processors in 2004.

I did say "There was speculation at one point." And you agreed with enthusiasm, "based on wild assed speculation." If you do agree, no need to bring up the same point over and over.

Do you realize that both the eMac and iMac use a G4 processor that runs at 1 GHz? The idea that IBM didn't release a G3 running at 1 GHz because of having to keep it below the frequency of the G4 is ludicrous. The G4 would obviously have a higher performance than than G3 when both are running at the same frequency due to the G4 having Altivec and also the availability of a L3 cache for the G4.
A lot of consumers, especially "switchers" who are not familiar with difference between G3 and G4, often rely on clock speed for judging performance. Therefor they did not want iBook with identical clock speed as PowerBook to prevent iBook sales cannibalizing PowerBook sales. They want PowerBook to be definitively and clearly better than iBook in all specs. Also, G3 is capable of out performing G4 with similar clock speed in some non-AltiVec tasks. Not every software benefits from AltiVec.

The 970 has Altivec.
970 has AltiVec compatible SIMD. It only makes sense 970 provides AltiVec guidelines.
http://www-3.ibm.com/chips/products/powerpc/newsletter/dec2002/newproductfocus2.html

How about the G3 iBooks catching up with the PC in performance. Or is it too much to ask for IBM to do the same thing with the G3 that is expected of Motorola's G4?

Apple has fallen behind in performance compared to PCs due to a falling market share. If you believe that is mainly due to Motorola, then ask yourself why is it that Apple's CFO Fred Anderson stated in the last fiscal quarterly report with analysts (webcast available on Apple's website) that he did not believe the G5 PowerMacs would ever obtain the peak quarterly sales rate of the G4 Power Mac. In fact he stated that it would probably be over 200,000 units per quarter and never reach 300,000 (G4 Power Mac sales reached 370,000 units in one quarter).
IBM has constantly upgraded G3 and made it just as good as G4 in non-AltiVec tasks. And now there are speculations about even more advanced G3 that includes SIMD functionality, 750vx. And at the same time, IBM also came up with G4 replacement, 970. IBM is doing better job at living up to expectations than Motorola.

The part of reason why Apple lost market share is due to Motorola's slow development on G4. G4 was stuck around 500MHz for a long time, during which AMD and Intel went ahead and hit 1GHz long before G4. Apple had to introduce Dual Processor Power Macs just to keep up.

If Apple's CFO said he doesn't think G5 PowerMac sales will ever reach that level, then there are several reasons for that. One, when G4 PowerMacs were selling well, computer market in general was doing well. It is completely the opposite now. And to add to that problem Apple has less market share.

If Apple goes with a single supplier it will more than likely be Intel. AMD is mentioned as a possible savior for Apple, but AMD has had months of delays getting Opteron out and the company has been running deeply into the red for several quarters.

It is very unlikely that IBM would be Apple's sole supplier with 9XX processors anytime soon because to do that IBM would have to be willing to sell some 9XX processors to Apple for less than $100 apiece in order for Apple to compete against the Celeron processors. IBM is undoubtedly selling Apple the 970s for several hundred dollars each at this time. In order for Apple to compete in the sub $1,000 computer market the company needs low priced processors and the G4 is likely to be one of those choices. Or do you believe that Apple should abandon the sub $1,000 computer market and simply have a much smaller market share with only elite performing computers?

The G3 seems destined to still be used by Apple at least through the first 6 months of 2004 with the 750GX just recently announced. What product the 750GX will be used in is difficult to tell since it would be a long dry spell for the iBooks to be bumped up to 1.1 GHz in January 2004 and the current top speed of the iBooks tops out at 900 MHz.
Why would Apple go with Intel when it is not interested in consumer level 64bit desktop processor, which Apple is trying to push? Intel's 64bit processor solution is not even backward compatible, whereas AMD's is 64 and 32bit compatible. AMD would make a better candidate, not Intel.

Phinius

Aug 2, 2003, 03:45 PM

Originally posted by Phil Of Mac

No one ever wanted to make a G3 that was faster than a G4. Even though it's been done at similar clock speeds, the G3 was never supposed to be faster than the G4, and Apple never ordered a G3 that was faster than the G4. They asked IBM for a G5 instead. And that's what they got.

The 970 is based on the Power4 core architecture and was designed from the start to be a highend server processor. Since Apple's first G5 Mac will be sold starting this month it's obvious that the 970 was thought up quite a bit later than the Power4.

Intel has simply increased the pace of development for chip manufacturing and the other manufactures have found that there is little financial benefit to try and keep up with them. If you believe that AMD has been able to do that, well AMD has lost buckets of money attempting it recently.

The reason the G3 or G4 has not kept pace with what Intel offers is simply due to economics. For every $1 that IBM or Motorola takes in from Apple, Intel rakes in $40 or $50 selling chips to the Windows world. So it takes IBM or Motorola much longer to invest in building a new multi-billion dollar chip facility for manufacturing Apple processors than it takes Intel to make enough for a new plant for Pentium chips. Also, the amount that Intel can devout to designing processors creates a faster turnaround time than what competitors can muster.

IBM spent at least 2 billion dollars creating the new Fishkill facility with the vew that demand for their chips would increase dramatically. Well, it turns out that it hasn't and IBM is losing money on chip manufacturing.

IBM created the Power4 knowing that they could hide the several thousand dollars cost of designing and manucturing each chip in high priced servers. Making a cut down version for use in Apple and IBM computers is simply a extension of the Power4 core sales. IBM would not have had the financial incentive to create a PowerPC processor for Apple from scratch.

Likewise Motorola does not have the financial incentive to design a PowerPC processor that would only be used for Macs. Motorola would not have the means to sell a Apple PowerPC derivitive for several thousand dollars each in a high priced computer. So, essentially they are stuck with having to make a low priced PowerPC processor for Apple's use than can also be used in other markets.

IBM will have a increasingly difficult time trying to keep up with Intel in server processors. IBM will be coming out with the Power5 which will run upt to 4 times faster than a Power4 computer. Intel intends to put two Itanium processors on one chip which will nearly halve the price of each processor. Then in mid decade Intel will be producing Itanium processors that have about 10 times the performance of the current Itanium chips. As a IBM engineer stated in a forum a few months back, 'how much longer can IBM stay ahead of Intel, when Intel has such deep pockets?'

Apple wasn't about to go back and lose Altivec.

Why not, according to you Motorola's Altivec technology has not advanced since 1999. It would seem that IBM would have come up with something better by now since you state that IBM is so superior to Motorola.

I don't doubt other chip makers will want to develop facilities similar to or better than Fishkill. But they haven't yet.

Intel will be rolling out Prescott processors on a 90-nm process in the fourth quarter of 2003. Where is IBM's 90-nm process for the 970?

Also IBM intends to come out with strained silicon on the 65-nm process. Intel will introduce strained silicon with Prescott at the 90-nm process level which will boost its performance by about 10%.

It would be a whole lot more feasible than relying on Motorola.

Apple has added the G5 to the Mac lineup, its not necessary at this time to have it replace the G4 or G3. Now Apple has more to processors to choose from and yet you advocate that Apple reduce their choices and simply go with a single design in the G5. Having the G3, G4 and G5 enables Apple to target each market segment much better than if they relied on the G5 for everything. The G5 will never match the low power use of the G3, nor will it come close to the low power use that the G4 will be made at. The G5 has advantages, but its power use is not a advantage when it comes to portables.

Perhaps they've already started. They already have a lot of PPC technology. It wouldn't be that unfeasible.

Your confusing instruction set with chip architecture. PPC is an instruction set and AMD does not use it whatsoever.

Motorola has shafted Apple more than enough times already. They are unreliable. So mark them off the list.

Sun wants nothing to do with Apple because of past history. AMD, however, might see Apple as a source of turnaround.

There is another option. Apple itself reportedly has many ex-Motorola chip designers. I don't consider it totally unfeasible that Apple themselves could design processors for IBM to manufacture at Fishkill.

Why would a lot of Motorola chip designers come to work for a company that only designed motherboards?

What financial advantage would Apple have to design processors? The processors would only be used in Macs and that is a very small segment of the 133 million+ PC sales per year. Designing their own processors would only add to the costs of making each Mac. Apple needs off the shelf hardware to compete on price and performance with the Wintel crowd. It's becoming increasingly difficult for Apple to keep using PowerPC processors and the G5 processors are probably the last hope for Apple to continue using a PowerPC processor instead of X86 chips.

Who needs to increase market share? You don't see Mercedes-Benz fretting about their infinestimal market share.

The personal computer market and the car market are completely different. Personal computers have become commodities much like cell phones, tvs and cd players. Apple needs to have market share in order to have developers continue creating applications and also to lower costs of building each computer so that their product can compete on price/performance. Apple has been losing not only market share in the last four years, but also Mac sales have been falling overall. Meanwhile, except for one year, personal computers have been increasing in sales over the last four years. So, Apple has been shrinking in sales while the overall personal market has grown. That cannot continue with any hope of Apple remaining a personal computer manufacturer when each Mac would be slower, have much less software, and sold at a much higher price.

Phinius

Aug 2, 2003, 04:21 PM

Originally posted by macphoria
[I did say "There was speculation at one point." And you agreed with enthusiasm, "based on wild assed speculation." If you do agree, no need to bring up the same point over and over.

I clearly pointed out that the G5 was not intended for the Apple market on Motorola's roadmap way back in 2001.

Also, G3 is capable of out performing G4 with similar clock speed in some non-AltiVec tasks.

The G3 is made on a smaller process size than the current G4. On the same process size the 1.3 GHz 7457 will clearly outperform the 900 MHz 750FX.

970 has AltiVec compatible SIMD. It only makes sense 970 provides AltiVec guidelines.
http://www-3.ibm.com/chips/products/powerpc/newsletter/dec2002/newproductfocus2.html

970's Altivec has the same amount of instructions as the G4. IBM lists the 970 as having Altivec and Altivec is a trademarked name by Motorola. Therefore the 970 is using Motorola's Altivec.

IBM has constantly upgraded G3 and made it just as good as G4 in non-AltiVec tasks.

Really? Does the G3 have Altivec or a L3 cache? The G3 has 4 pipeline stages and the G4 has 7. The G3 can use up to a 200 MHz bus and the G4 can use a 167 MHz bus. With the use of a L3 cache, the G4 overcomes the G3s speed advantage of a faster bus. The G3 does have double the L2 cache of the G4 and that is a advantage, but the 7457 will have the same L2 cache size as the 750FX, so the 750FX will no longer have an advantage there.

And now there are speculations about even more advanced G3 that includes SIMD functionality, 750vx.

IBM just recently announced the 750GX that will move the 750 design up to 1.1 GHz and a doubling of L2 cache to 1 MB. That still won't overcome the speed advantage of the 7457 which lists up to 1.3 GHz and can reach speeds of 1.8 GHz on higher voltage chips intended for Apple's market.

And at the same time, IBM also came up with G4 replacement, 970. IBM is doing better job at living up to expectations than Motorola.

If you'll notice Apple will still be selling G4 Macs when the G5 is released. There are no indications yet from Apple that the G5 will displace either the G3 or the G4. It looks like the G5 is simply an additional choice for Apple.

The part of reason why Apple lost market share is due to Motorola's slow development on G4. G4 was stuck around 500MHz for a long time, during which AMD and Intel went ahead and hit 1GHz long before G4. Apple had to introduce Dual Processor Power Macs just to keep up.

If Apple's CFO said he doesn't think G5 PowerMac sales will ever reach that level, then there are several reasons for that. One, when G4 PowerMacs were selling well, computer market in general was doing well. It is completely the opposite now. And to add to that problem Apple has less market share.

Your claim is that part of the reason Apple lost market share was due to Motorola's slow development of the G4. Then you go on to state that a reason G5 Power Macs will not sell as well as G4 Power Macs once did is due to the overall personal computer market not doing well. That is simply not the case. The overall computer market is growing and in only one of the past four years did the overall market fall in sales. Fred Anderson obviously believes that the G5 Power Mac will not reach the high sales point of the G4 Power Mac even with a bigger overall personal computer market.

Why would Apple go with Intel when it is not interested in consumer level 64bit desktop processor, which Apple is trying to push?

Apple as of yet is not pushing 64 bits. One of Apple's executives stated that Apple at this point sees very little advantage to 64 bits for their market. Its the speed increase that the G5 brings which interests Apple at this point in time.

Intel's 64bit processor solution is not even backward compatible, whereas AMD's is 64 and 32bit compatible. AMD would make a better candidate, not Intel.

You mean AMD which delayed the release of Opteron for several months? So this struggling company would spend several years designing and building a processor exclusively for Apple's small and shrinking market? That makes perfect sense to me.

Phinius

Aug 2, 2003, 04:53 PM

Originally posted by Snowy_River
I'm glad to see that I'm not the only one with a memory of what Moto has said in the past.

Where did Motorola every state that a G5 PowerPC processor was intended for the Apple market? All of this has been conjecture based on rumors. You jump to the conclusion that Motorola has let Apple down by not delivering on what was projected in these rumors? Geez...

If you'll take a look at Motorola's PowerPC roadmap dated November of 2001 you can see that the G5 8500 chips are intended for markets that Apple is not in. But then maybe you don't want to believe the actual information from the company that manufacturers the chips, but instead rely on rumors from unknown sources.

(According to those numbers, the 8500 would have been faster per cycle and at higher clock speeds two years ago than the 970 is now. Seems like a bit of an exageration to me.)

That should tip you off that these rumors of a 8500 G5 Motorola processor made for Apple could be a complete fabrication.

However, the fact that the 8500 was going to be a host processor, and likely the base for the Power Mac G5, before that branch of development was cancelled, is remembered by more people than just me.

How do these rumors become facts to you when there are no known written documents where Motorola states that the G5 was ever intended for Apple's market? However, there have been internal Motorola documents posted on the internet that mention a future dual-core G4, the 7457 and a later 7457-RM.

Phinius

Aug 2, 2003, 05:46 PM

Originally posted by Phil Of Mac
They didn't want a non-Altivec chip that was faster than an Altivec chip. They chose instead for IBM to design an Altivec chip that was faster. The 970.

Apple clearly knew that IBM was going to make a Power4 processors designed for highend servers long before it went into production. It simply makes sense that Apple would make every effort to try and convince IBM to make a cut down version for the PowerPC market. Apple's Jon Rubenstein stated that Apple worked with IBM for three years on the 970. IBM did not design a PowerPC processor exclusively for Apple's market, its a derivitive of a exisiting chip design.

It wouldn't matter whether Motorola came out with a G5 processor for Apple two years ago and Motorola was the most reliable chip manufacturer in the world . Apple would still want to get a PowerPC version of the Power4 processor. The 970 is not intended as a replacement for what Motorola offers. The 970 allows Apple to expand into markets that desktop processors are not designed for. The 970 is derived from the 64-bit server market that the Power4 and Itanium are made for. Even the Pentium 4 is restricted to 1-2 processor servers and the desktop or notebook markets. The 970 can be configured up to 16 processors.

It is Altivec compatible.

It is Altivec and IBM clearly states it in literature about the 970.

But then again, Motorola has been using PowerPC designs ever since 1994. That they got from IBM.

The G4 is not a IBM design.

IBM is using an Altivec-compatible SIMD unit. It is not the same as Altivec.

Oh yes it is.

Vaporware.

You mean like the 980 that you claim will come out next year? Where are ANY IBM literature or statements published on the internet that confirm that rumor? Or for that matter where does IBM claim that a 970 or 980 will reach speeds of 4 GHz or 5 GHz next year as you have claimed? IBM statements and documents have clearly shown that the 970 is expected to reach 3 Ghz a year from now and also a German IBM document stated that IBM intends to come out with a server that uses 970 processors that run up to 2.4 or 2.5 GHz. Everything beyond that seems to be based simply on unconfirmed rumors and wild speculation.

The 7457 however was clearly mentioned and shown on a roadmap posted from a internal Motorola document which stated that the 7457 will reach speeds of 1.8 GHz. This document also mentions the expected 7457-RM version.

Also, this Register article (http://www.theregister.co.uk/content/39/31026.html) states that their staff has seen a Motorola document which indicated the company would give a presentation on the upcoming dual-core G4.

The reason it happened with Motorola is because Motorola couldn't produce enough G4's to fill the entire product line. IBM has no problem producing enough chips at Fishkill.

More evidence that you simply make a lot of this up. Your stating that Motorola could not produce enough processors to eliminate the G3 from the iBooks? That's simply not plausible. Apple probably held on to the G3 due to it's low cost and reasonable price/performance for low priced notebook computers. The G3 is almost 1/3 the size of the G4 and so therefore should be considerably cheaper. For the desktop market the G3 was simply outclassed by the G4. In fact Apple has stated that one of the demands by schools for a iMac replacement was to have the G4. So Apple included the G4 when the eMac was created.

You have little or no information about the production of IBM's Fishkill fabrication facility and yet you jump to the conclusion that IBM has no problem producing enough chips there.

macphoria

Aug 2, 2003, 06:15 PM

The G3 is made on a smaller process size than the current G4. On the same process size the 1.3 GHz 7457 will clearly outperform the 900 MHz 750FX.
1.3 GHz G4 vs 900MHz G3? I certainly hope G4 will win. It better. If it ever comes out. But my comment was on G3 and G4 of equal clock speed.

970's Altivec has the same amount of instructions as the G4. IBM lists the 970 as having Altivec and Altivec is a trademarked name by Motorola. Therefore the 970 is using Motorola's Altivec.
You did not even read that official IBM press release. It clearly states, "With the introduction of the PowerPC 970, IBM has taken PowerPC performance to new heights. At up to 1.8 GHz, the PowerPC 970 is the fastest PowerPC yet introduced. But the 970 employs much more than frequency to answer the demands of high-performance computing customers. The 970's multiple execution units including an AltiVec™ COMPATIBLE vector processor are fed by an up to 900-MHz processor interface bus, which can deliver data at a rate of up to 6.4 GBps."

Really? Does the G3 have Altivec or a L3 cache? The G3 has 4 pipeline stages and the G4 has 7. The G3 can use up to a 200 MHz bus and the G4 can use a 167 MHz bus. With the use of a L3 cache, the G4 overcomes the G3s speed advantage of a faster bus. The G3 does have double the L2 cache of the G4 and that is a advantage, but the 7457 will have the same L2 cache size as the 750FX, so the 750FX will no longer have an advantage there.
Yes, what you say is true. But irrelevant to my comment. You said "How about the G3 iBooks catching up with the PC in performance." So I replied and suggested it might be doing just that, G3's non-AltiVec performance is challenging G4 and there is possibility of G3 with SIMD, 750vx. But you are replying along the lines "How about the G3 iBooks catching up with the PC in performance" and going around in circles as if I didn't reply.

IBM just recently announced the 750GX that will move the 750 design up to 1.1 GHz and a doubling of L2 cache to 1 MB. That still won't overcome the speed advantage of the 7457 which lists up to 1.3 GHz and can reach speeds of 1.8 GHz on higher voltage chips intended for Apple's market.
I certainly hope so. If it ever comes out.

If you'll notice Apple will still be selling G4 Macs when the G5 is released. There are no indications yet from Apple that the G5 will displace either the G3 or the G4. It looks like the G5 is simply an additional choice for Apple.
Of course not. Apple has two distinct product categories, Professional and Consumer. They need various processors for different categories. And indeed G5 IS "and additional choice for Apple." But you are again missing the point and arguing for reasons I can't understand. My comment was about IBM living up to expectations, how IBM now provides G3 and G5, which Motorola ended up not making. Nothing to do with G5 replacing anything. Your reply has nothing to do with my comment about living up to expectations. Had you said something like "Motorola's product will still be used in more Mac products (eMac, iMac, PowerBooks) than IBM's products (iBook and Power Mac) then yes you do have some point about living up to expectations.

Your claim is that part of the reason Apple lost market share was due to Motorola's slow development of the G4. Then you go on to state that a reason G5 Power Macs will not sell as well as G4 Power Macs once did is due to the overall personal computer market not doing well. That is simply not the case. The overall computer market is growing and in only one of the past four years did the overall market fall in sales. Fred Anderson obviously believes that the G5 Power Mac will not reach the high sales point of the G4 Power Mac even with a bigger overall personal computer market.
First, it is known fact that Power Mac sales slowed down when G4's clock speed remain unchanged for a long time. Second, I said "One, when G4 PowerMacs were selling well, computer market in general was doing well. It is completely the opposite now." My point was that computer market was far more active then with internet boom and all that, resulting in high sales. It will be hard for latest sales to catch up to that level. And I don't even know how you came to conclusion about "Fred Anderson obviously believes..."

Apple as of yet is not pushing 64 bits. One of Apple's executives stated that Apple at this point sees very little advantage to 64 bits for their market. Its the speed increase that the G5 brings which interests Apple at this point in time.
This is partially true. I do believe their current goal is to increase performance, not necesarilly 64 bit processing right away. However, they are committed to 64 bit processing, and probably same could be said for AMD, whereas Intel showed lack of interest. And this was the context of my comment about Apple "pushing" 64 bit, not necessarily the way you took it.

You mean AMD which delayed the release of Opteron for several months? So this struggling company would spend several years designing and building a processor exclusively for Apple's small and shrinking market? That makes perfect sense to me.
This is true. But my previous comment still stands. AMD has made more practical 64bit solution, which is backward compatible, therefore better partner for Apple which is also pushing 64 bit consumer level desktop processing. And if "delayed" release of a processor is such and issue for you, you should take up the issue with Motorola as well.

Now, I'm getting tired of you going off in tangents and arguing for the sake of arguing. I made effort to read your comments and grasp where you are coming from and make counter argument. But you do not fully read other people's comments nor make effort to understand different views. You have yet to acknowledge a single fault on your part, even the most blatantly obvious ones. You simply argue for the sake of arguing and go around in circles. It is prefectly normal to have different points of views and I enjoy heated discussions. But I've grown tired of your half hearted effort to communicate and interact. If a place to chase your own tail and go around in circles is what you want, then this forum is yours. I'm done.

Phil Of Mac

Aug 2, 2003, 06:18 PM

Originally posted by the Motorola Information Minister
Motorola has a PowerPC roadmap (http://e-www.motorola.com/files/archives/doc/roadmap/PPCRMAP.pdf), dated 11/30/2001, that clearly mentions the G5 as a 85XX processor. If you look at the bottom of the roadmap there is an explanation of each processor number. Motorola states that the 85XX (G5) is a 'integrated processor targeting the communication and consumer markets'. The 7XXX (G4) processor is a 'high performance microprocessor targeting computing and high-end embedded'. The processors that Apple uses are not integrated processors and so therefore the G5 on the roadmap is not intended for Apple's use in computers.

That roadmap is from late 2001. Earlier versions, circa 1999, definitely showed a G5 intended for personal computing applications.

Originally posted by the Motorola Information Minister
Do you realize that both the eMac and iMac use a G4 processor that runs at 1 GHz? The idea that IBM didn't release a G3 running at 1 GHz because of having to keep it below the frequency of the G4 is ludicrous. The G4 would obviously have a higher performance than than G3 when both are running at the same frequency due to the G4 having Altivec and also the availability of a L3 cache for the G4.

Actually, on non-Altivec related tasks, the G3 often outperforms the G4.

Originally posted by the Motorola Information Minister
How about the G3 iBooks catching up with the PC in performance. Or is it too much to ask for IBM to do the same thing with the G3 that is expected of Motorola's G4?

Considering that the G4 is a newer and superior design than the G3, it is. That would be like asking the G4 to outperform the G5.

Originally posted by the Motorola Information Minister
Apple has fallen behind in performance compared to PCs due to a falling market share.

What? Sales figures magically alter a computer's performance, more than the CPU?

Originally posted by the Motorola Information Minister
If you believe that is mainly due to Motorola, then ask yourself why is it that Apple's CFO Fred Anderson stated in the last fiscal quarterly report with analysts (webcast available on Apple's website) that he did not believe the G5 PowerMacs would ever obtain the peak quarterly sales rate of the G4 Power Mac. In fact he stated that it would probably be over 200,000 units per quarter and never reach 300,000 (G4 Power Mac sales reached 370,000 units in one quarter).

It's called a recession. That's when the economy goes down.

Originally posted by the Motorola Information Minister
It is very unlikely that IBM would be Apple's sole supplier with 9XX processors anytime soon because to do that IBM would have to be willing to sell some 9XX processors to Apple for less than $100 apiece in order for Apple to compete against the Celeron processors. IBM is undoubtedly selling Apple the 970s for several hundred dollars each at this time.

How do you know this? I thought you worked for Motorola!

Originally posted by the Motorola Information Minister
In order for Apple to compete in the sub $1,000 computer market the company needs low priced processors and the G4 is likely to be one of those choices. Or do you believe that Apple should abandon the sub $1,000 computer market and simply have a much smaller market share with only elite performing computers?

This isn't like Intel. IBM doesn't produce processors, put a list price on them, and put them on sale. They sign a contract with Apple to provide them in bulk. I don't know the terms of that contract, but it's concievable that if Apple wants millions upon millions of parts, IBM will make them and lower the price. It's also possible that Apple will subsidize some of the R&D for the 970 so they can buy the parts for just above what it costs to manufacture them, which is probably not prohibitively expensive.

Phinius

Aug 2, 2003, 06:19 PM

Originally posted by Phil Of Mac
Motorola's Altivec and IBM's Altivec are still slightly different. The instruction counts, at least, are different. Minor difference, though.

The 970 uses the same amount of SIMD instructions as the 7450. That's due to IBM using Altivec for the 970.

So Intel has a fully automated facility as well. I think Fishkill has other advantages--producing chips from a larger wafer size, for instance.

Intel has several chip manufacturing facilities that use 300-mm silicon wafers, IBM has just one.

But the fact is, IBM already has the facility, Motorola doesn't yet.

Wrong again.

Motorola is working in a collaborative effort with Phillips and STMicroelectronics to produce microprocessors. In this Silicon Strategies article (http://www.siliconstrategies.com/article/printableArticle.jhtml?articleID=10802061) Motorola's Claudine Simson states that her company will have PowerPC test chips running this summer on a 90-nm process size and 300-mm wafers in STMicroelectronics Crolles fabrication facility in France.

She also states in this June interview that a PowerPC processor made on a 130-mm process was 'now moving out'. So why doesn't Apple have a Mac computer yet that uses this processor? Quite simply because Motorola announced that the 7457 will not be in production until the fourth quarter of this year. The interview does clearly indicate that a PowerPC is already on the 130-nm process and since Apple always gets the first Motorola PowerPC processors on a new process, then its obvious that she was talking about the 7457 chips.

And if Motorola builds production facilities the same way they build processors, it won't be here for awhile :)

Motorola intends to have partners for most new fabrication facilities due to the multi-billion dollar costs of building them. Intel recently stated that only a handful of chip manufactuers with the highest revenue can now fiancially justify building a fabrication facility. Unfortunately IBM is not one of those top five revenue producers, in fact IBM is not even in the top ten. That makes it much less likely that IBM will want to create a next generation chip facility in the next 2-3 years like Intel will. Which means IBM could fall further and further behind in state of the art chip fabrication without a joint venture with other chip manufacturers.

Phil Of Mac

Aug 2, 2003, 06:29 PM

Originally posted by the Motorola Information Minister
Right, no one would be interested in buying G3 iBooks running at a higher frequency than 900 Mhz, that would make it much less competitive. Geez....do you even stop to think your statements through before posting?

It's called product differentiation. If IBM could pump up the older G3 design to truly compete with the G4 in terms of speed, who would buy a G4? Then we lose Altivec, which Apple spent good time and money coding Mac OS X for.

Originally posted by the Motorola Information Minister
You mean to say go with the chip manufacturer that had repeated delays getting the Opteron processor to market? Isn't getting processors out on time the main reason you believe that Apple should dump Motorola? Now you advocate going with a chip manufacturer that has a recent history of delays that are far worse than what Motorola is exhibiting.

So buy AMD, cancel everything, and use the engineers to design PPC's. If Apple had the personnel, they could surely work together with IBM at IBM facilities. Perhaps Apple and IBM could work out an arrangement much like Somerset, where Apple and IBM chip designers work together on new designs.

Originally posted by the Motorola Information Minister
The 970 is the future and not the present or past. There are no Macs sold yet with a 970 inside.

Many have been sold. They are simply not yet delivered. And, once again, you are missing the point. Apple wanted IBM to beat the G4 in performance. They asked them to do it with the 970, not with the G3.

Originally posted by the Motorola Information Minister
Motorola added SOI and a low k-dielectric process since then.

SOI was an IBM invention. So were copper interconnects, a few years ago.

Originally posted by the Motorola Information Minister
Also, can you name a SIMD process by IBM or Intel that is superior in performance to Altivec? Quite simply Altivec has been Apple's saving grace in performance comparison to Wintel computers.

That's nice. But this is business, not academia. You don't get tenure for what you did 4 years ago, especially if you've slacked off since then.

Originally posted by the Motorola Information Minister
The PowerPC was not invented by IBM alone. It was a joint collaboration by Motorola, IBM and Apple.

The PowerPC 601, the first PowerPC processor, was designed exclusively by IBM. PowerPC is an IBM trademark. It is based on IBM's POWER technologies. Motorola was there simply to help, and to provide a friendly face to Mac users who still saw IBM as the enemy.

Originally posted by the Motorola Information Minister
The average price of a desktop PC currently sold is about $749 according to a recent market research survey. Quite a lot of those computers are using processors that have a list price of under $70 from Intel.

Hooray.

Originally posted by the Motorola Information Minister
It's quite possible that the 970 costs Apple less than the 1.42 GHz 7455 processors. However, the 970 is almost certainly costing Apple several hundred dollars apiece. That is a long ways from the under $70 that Intel gets for Celeron processors.

Wait, you get angry at us for using rumors, while you're just making that up!

Originally posted by the Motorola Information Minister
A bigger processor with a larger die size usually costs more to produce than a smaller die size. By virtue of the G3 being around 33 mm^2 in size it will naturally cost less to produce than the 970 which is about 117 mm^2. The 7457 will be a little bit smaller than the 970, so it could quite possibly be cheaper to manufacture than the 970.

It's manufactured by Motorola in facilities that aren't nearly as economic as Fishkill.

Phil Of Mac

Aug 2, 2003, 07:52 PM

Originally posted by the Motorola Information Minister
Also, the amount that Intel can devout to designing processors creates a faster turnaround time than what competitors can muster.

I didn't know it was a religious experience for Intel.

Originally posted by the Motorola Information Minister
IBM will have a increasingly difficult time trying to keep up with Intel in server processors. IBM will be coming out with the Power5 which will run upt to 4 times faster than a Power4 computer. Intel intends to put two Itanium processors on one chip which will nearly halve the price of each processor. Then in mid decade Intel will be producing Itanium processors that have about 10 times the performance of the current Itanium chips. As a IBM engineer stated in a forum a few months back, 'how much longer can IBM stay ahead of Intel, when Intel has such deep pockets?'

Itanium has been a spectacular failure so far. Regardless, the PowerPC has other advantages, stemming from its superiority as a chip architecture to x86. While x86 is attempting to overcome its disadvantages, IBM and Apple can spend time and money working on the advantages of PPC.

Originally posted by the Motorola Information Minister
Why not, according to you Motorola's Altivec technology has not advanced since 1999. It would seem that IBM would have come up with something better by now since you state that IBM is so superior to Motorola.

Altivec has indeed not advanced since 1999. That's because it's in many ways a supplemental instruction set. If they wanted to advance it, they'd have to change that instruction set, and everyone would have to rewrite everything.

IBM's not interested in adding supplemental instruction sets. They're interested in making a better processor. All the better for them.

Originally posted by the Motorola Information Minister
Intel will be rolling out Prescott processors on a 90-nm process in the fourth quarter of 2003. Where is IBM's 90-nm process for the 970?

IBM has said nothing about it that I know of, but reportedly it'll be here around January or so.

Originally posted by the Motorola Information Minister
Apple has added the G5 to the Mac lineup, its not necessary at this time to have it replace the G4 or G3. Now Apple has more to processors to choose from and yet you advocate that Apple reduce their choices and simply go with a single design in the G5. Having the G3, G4 and G5 enables Apple to target each market segment much better than if they relied on the G5 for everything. The G5 will never match the low power use of the G3, nor will it come close to the low power use that the G4 will be made at. The G5 has advantages, but its power use is not a advantage when it comes to portables.

The G5 can be changed so it has better power use. That'll probably happen at 90 nm. As for reasons to abandon the G3 and G4, well, the G3 is too far behind in terms of technology, while the G4 is made by Motorola, who has been proven to be unreliable. Additionally, with the G3 we saw Apple move to a "one processor design for all computers" model. This was so even the lowly iMac had the same type of ass-kicking processor that the most powerful Power Mac had (albeit at a lower frequency). Given Motorola's production problems with the G4 (they've never been able to produce enough), Apple couldn't do that with the G4. The G5 will undoubtedly be plentiful enough to install across the line.

Originally posted by the Motorola Information Minister
Your confusing instruction set with chip architecture. PPC is an instruction set and AMD does not use it whatsoever.

Regardless, AMD processors actually use a proprietary AMD instruction set. There's an x86 frontend attached to the processor which translates from AMD to x86. I'm willing to bet that with the technology sharing AMD has done with PowerPC chip makers, the internal AMD instruction set is not that dissimilar from PPC. An AMD to PPC translation frontend, much like the AMD to x86 frontend, may not be unfeasible.

If the AMD instruction set is more similar to PPC than we thought, then AMD might just switch to the PPC instruction set, and design a PPC to x86 frontend.

Originally posted by the Motorola Information Minister
Why would a lot of Motorola chip designers come to work for a company that only designed motherboards?

Apple's already consulted with IBM and Motorola on the design of the G3. Without a doubt, Apple already has some people who know something about chip design. If the reports of Motorola chip designers going to Apple is true, then perhaps Apple is going to design processors.

Originally posted by the Motorola Information Minister
What financial advantage would Apple have to design processors? The processors would only be used in Macs and that is a very small segment of the 133 million+ PC sales per year. Designing their own processors would only add to the costs of making each Mac. Apple needs off the shelf hardware to compete on price and performance with the Wintel crowd. It's becoming increasingly difficult for Apple to keep using PowerPC processors and the G5 processors are probably the last hope for Apple to continue using a PowerPC processor instead of X86 chips.

Apple already pays the cost for IBM or Motorola to design their processors. Neither IBM or Motorola wants to produce processors just for Apple. Apple does. And if Apple can move the costs in-house for chip design, then they only have to rely on IBM or whoever for production. There's also a possibility of Apple and IBM doing joint processor design, much like IBM and Motorola once did.

Originally posted by the Motorola Information Minister
The personal computer market and the car market are completely different. Personal computers have become commodities much like cell phones, tvs and cd players. Apple needs to have market share in order to have developers continue creating applications and also to lower costs of building each computer so that their product can compete on price/performance. Apple has been losing not only market share in the last four years, but also Mac sales have been falling overall. Meanwhile, except for one year, personal computers have been increasing in sales over the last four years. So, Apple has been shrinking in sales while the overall personal market has grown. That cannot continue with any hope of Apple remaining a personal computer manufacturer when each Mac would be slower, have much less software, and sold at a much higher price.

Apple is dying, eh?

You know what, though, you've missed the number one point. In this recession, only two PC makers are profitable: Dell and Apple. If Apple was profitable in spite of Motorola, then with the G5, Apple can be even more profitable. With G5's across the line, Apple will have the price-performance advantage over almost every other PC. Within a short time, I think IBM can surpass Intel in performance in a clear and significant way. The PowerPC has almost always outperformed Intel. The one clear exception--the G4--was the fault of Motorola. With the G5 and the partnership with IBM, Apple's future is bright indeed.

Phil Of Mac

Aug 2, 2003, 08:00 PM

Originally posted by the Motorola Information Minister
I clearly pointed out that the G5 was not intended for the Apple market on Motorola's roadmap way back in 2001.

In 2001, Motorola gave up on developing the G5 for the Mac because of a serious design flaw. Pre-2001 roadmaps show the 8500 G5 as a personal computing CPU.

Originally posted by the Motorola Information Minister
970's Altivec has the same amount of instructions as the G4. IBM lists the 970 as having Altivec and Altivec is a trademarked name by Motorola. Therefore the 970 is using Motorola's Altivec.

IIRC, the 970's implementation of Altivec has 160 instructions. The 74xx's implementation has 162 instructions.

Originally posted by the Motorola Information Minister
If you'll notice Apple will still be selling G4 Macs when the G5 is released. There are no indications yet from Apple that the G5 will displace either the G3 or the G4. It looks like the G5 is simply an additional choice for Apple.

Apple can't redesign all their products at once. Check again in 18 months. The G3 and G4 will be dead.

Originally posted by the Motorola Information Minister
Your claim is that part of the reason Apple lost market share was due to Motorola's slow development of the G4.

Staying at 500 MHz for a year does give you a price/performance disadvantage.

Originally posted by the Motorola Information Minister
Then you go on to state that a reason G5 Power Macs will not sell as well as G4 Power Macs once did is due to the overall personal computer market not doing well. That is simply not the case. The overall computer market is growing and in only one of the past four years did the overall market fall in sales.

We're in a recession. The PC market is in a slump.

I've decided you're stuck in the 90's, Phinius. You think the PC market is growing, Motorola is still a reliable chip maker, and Apple is dying. All three were true in the 90's. None are true today.

Phil Of Mac

Aug 2, 2003, 08:08 PM

Originally posted by the Motorola Information Minister
It wouldn't matter whether Motorola came out with a G5 processor for Apple two years ago and Motorola was the most reliable chip manufacturer in the world . Apple would still want to get a PowerPC version of the Power4 processor. The 970 is not intended as a replacement for what Motorola offers. The 970 allows Apple to expand into markets that desktop processors are not designed for. The 970 is derived from the 64-bit server market that the Power4 and Itanium are made for. Even the Pentium 4 is restricted to 1-2 processor servers and the desktop or notebook markets. The 970 can be configured up to 16 processors.

By that reasoning the 970 would be in the Xserve. It's in the Power Mac, which is intended for content creation professionals. Thus, the 970 is intended for general personal computing because it blows the doors off the G4.

Originally posted by the Motorola Information Minister
You mean like the 980 that you claim will come out next year? Where are ANY IBM literature or statements published on the internet that confirm that rumor?

It's been stated that IBM and Apple are working on the next generation after the G5. It pretty much follows from that that they're working on a Power5 derivative which will come out around the same time as the Power5.

Originally posted by the Motorola Information Minister
Or for that matter where does IBM claim that a 970 or 980 will reach speeds of 4 GHz or 5 GHz next year as you have claimed? IBM statements and documents have clearly shown that the 970 is expected to reach 3 Ghz a year from now and also a German IBM document stated that IBM intends to come out with a server that uses 970 processors that run up to 2.4 or 2.5 GHz. Everything beyond that seems to be based simply on unconfirmed rumors and wild speculation.

I was exaggerating when I said we'd be at 4 or 5 GHz next year. Obviously you didn't look up the meaning of the word "hyperbole" like I told you to.

Originally posted by the Motorola Information Minister
The 7457 however was clearly mentioned and shown on a roadmap posted from a internal Motorola document which stated that the 7457 will reach speeds of 1.8 GHz. This document also mentions the expected 7457-RM version.

Motorola is less reliable than most rumors sites.

Originally posted by the Motorola Information Minister
More evidence that you simply make a lot of this up. Your stating that Motorola could not produce enough processors to eliminate the G3 from the iBooks? That's simply not plausible.

Given Motorola's production problems ever since the release of the G4, it's very plausible.

Phinius

Aug 2, 2003, 08:10 PM

Originally posted by Phil Of Mac

It's called product differentiation. If IBM could pump up the older G3 design to truly compete with the G4 in terms of speed, who would buy a G4? Then we lose Altivec, which Apple spent good time and money coding Mac OS X for.

IBM is using Altivec in the 970 design, so its stands to reason that IBM could have added Altivec to the G3. Oh wait!, IBM is rumored to be being just that with a 750VX version. So since IBM added Altivec to the 970 design and still intends to use the G4 and G3, then Apple would not have lost the G4 is IBM matched its performance with the G3.

As it now stands IBM intends to move the 750 performance up a few notches by doubling its L2 cache and bumping up the frequency by 200 MHz. That will still put its performance below the G4 due to Altivec. If IBM does add Altivec to a 750 chip design, by that time Motorola will have put two G4 processors on one chip.

Are you beginning to see how Apple is coordinating the chip development from both IBM and Motorola? Instead of duplicate chip performances from the two companies, Apple is making sure that each chip fits in a different Mac product. But that still doesn't justify IBM staying below the performance of Motorola's chips if indeed Motorola was doing such a lousy job for Apple. It's more like Intel was simply running away from the competition and Motorola was tagged as the bad guy in all this since they were the preemmenent chip maker for Apple. If Apple would have kept a 30-50% market share, then Motorola would have little trouble keeping up with Intel. Instead Apple chose a road that started in the early 80s that lead to a falling market share, which in turn led to them falling behind in performance and prices compared to the much larger Windows computer makers.

So buy AMD, cancel everything, and use the engineers to design PPC's.

So Apple should buy AMD and take on the companies debt? What in the world would be the advantage to Apple for them to buy AMD? That would just increase Apple's overhead in making Macs.

If Apple had the personnel, they could surely work together with IBM at IBM facilities. Perhaps Apple and IBM could work out an arrangement much like Somerset, where Apple and IBM chip designers work together on new designs.

Apple is already working closely with IBM as I am sure they are doing with Motorola. Jon Rubenstein even stated that Apple worked with IBM for three years to get the 970 in production.

Many have been sold. They are simply not yet delivered. And, once again, you are missing the point. Apple wanted IBM to beat the G4 in performance. They asked them to do it with the 970, not with the G3.

Apple needed to step up in performance in order to keep up with Intel's Pentium 4 processors and neither of the PowerPC chip manufacturers that would sink money into designing a chip architecture just for Apple's personal computer market. Motorola had just come out with the G4 in mid year of 1999 and a processor core design has a life cycle of usually 4-5 years.

Apple's discussions with IBM probably were more along the lines of trying to design the 970 to keep pace with Intel's Pentium chips rather than trying to beat Motorola's G4. In order to accomplish that Apple probably insisted that Altivec be included so that the 970 would have a speed advantage in vector processing if not in integer or frequency performance.

SOI was an IBM invention. So were copper interconnects, a few years ago.

IBM did get there first. However, Motorola seems to have a different SOI method than IBM. Intel however does not feel the costs of SOI justify the gains in performance so they have yet to implement it and just recently Intel started using copper wiring.

Also IBM announced a strained silicon process well ahead of Intel and yet Intel will be introducing it a process generation ahead of IBM.

That's nice. But this is business, not academia. You don't get tenure for what you did 4 years ago, especially if you've slacked off since then.

If Altivec has grown outmoded, then IBM should have come up with a different vector unit for the 970 instead of using Altivec.

The truth is Intel has simply stepped up the pace to the degree that few chip makers can even hope to keep up without losing money at it. IBM has been able to design a Power4 processor that would take on Sun's UltraSparc, but IBM is also losing money on their hardware division for some time due to Intel cutting prices and Wintel servers encroaching on Power server sales.

The PowerPC 601, the first PowerPC processor, was designed exclusively by IBM. PowerPC is an IBM trademark. It is based on IBM's POWER technologies. Motorola was there simply to help, and to provide a friendly face to Mac users who still saw IBM as the enemy.

PowerPC is a instruction set and the 'G' moniker is designating what generation the chip is for Apple or Motorola. IBM does not use the G designation. Likewise the chip designs are independently coming from Motorola and IBM, the 'G' or PowerPC designation does not indicate similiarities in chip design.

Wait, you get angry at us for using rumors, while you're just making that up!

The dual 2 Ghz G5 Power Mac is priced three hundred dollars more than the dual 1.42 GHz Power Mac that it replaced. That alone should give an indication that Apple is paying about the same or slightly more for the 2 Ghz G5 than for the 1.42 GHz G4. Motorola listed the 1 GHz 7455 at either two or three hundred dollars when it was first announced. The 1.42 GHz 7455 running at a higher voltage and the addition of a low k-dielectric should raise the price of each chip considerably beyond what the 1 Ghz chips sell for. In fact I believe the Register speculated that Apple may have payed over four hundred dollars for each 1.42 Ghz G4 processor based on what the going rate for the processors are on the open market. Seeing how Intel charges up to six hundred dollars for topend Pentium 4 processors, it's more than likely that Apple is paying several hundred dollars per each G5 or topend G4 processor.

So no, I'm not exactly making that up, it's just a rather educated guess as to what Apple is paying for processors. But, there is abosolutely no way that Apple is paying under a hundred dollars per chip for G5 processors. IBM would not be stupid enough to sell it to them for that price when Intel gets up to $600 for each topend Pentium 4 or Xeon processor.

It's manufactured by Motorola in facilities that aren't nearly as economic as Fishkill.

The 7457 maybe produced in Motorola's older chip facilities, but they intend to make some 90-nm PowerPC processors at the Crolles facility on 300-mm wafers. Since Apple gets Motorola's topend processors, then it's obvious that a G4 designed for Apple's use will be made at the Crolles fabrication facility.

Phil Of Mac

Aug 2, 2003, 08:15 PM

Originally posted by the Motorola Information Minister
Motorola's Claudine Simson states that her company will have PowerPC test chips running this summer on a 90-nm process size and 300-mm wafers in STMicroelectronics Crolles fabrication facility in France.

Motorola isn't quite reliable, now is it? So what if they do? IBM already has PowerPC 970's running on their fabrication facility in New Jersey!

So Motorola now has 90 nm *this summer*? When Intel won't have it until the end of the year? I thought you said Motorola couldn't keep up with Intel!

Originally posted by the Motorola Information Minister
Quite simply because Motorola announced that the 7457 will not be in production until the fourth quarter of this year.

They promised it would be bumped up to May. They lied.

Originally posted by the Motorola Information Minister
Intel recently stated that only a handful of chip manufactuers with the highest revenue can now fiancially justify building a fabrication facility. Unfortunately IBM is not one of those top five revenue producers, in fact IBM is not even in the top ten. That makes it much less likely that IBM will want to create a next generation chip facility in the next 2-3 years like Intel will. Which means IBM could fall further and further behind in state of the art chip fabrication without a joint venture with other chip manufacturers.

Intel's just spewing out some FUD. They have a vested interest in derailing other chip makers.

IBM right now can make money from other chip designers going to IBM for fabrication. If IBM does that, then they'll have a profitable business.

Phil Of Mac

Aug 2, 2003, 08:39 PM

Originally posted by the Motorola Information Minister
IBM is using Altivec in the 970 design, so its stands to reason that IBM could have added Altivec to the G3. Oh wait!, IBM is rumored to be being just that with a 750VX version.

I don't buy those rumors and neither do you.

Originally posted by the Motorola Information Minister
So since IBM added Altivec to the 970 design and still intends to use the G4 and G3, then Apple would not have lost the G4 is IBM matched its performance with the G3.

IBM doesn't use the G4. In fact, IBM doesn't use the G3 either. In fact, I have no damn clue what you're trying to say.

Originally posted by the Motorola Information Minister
As it now stands IBM intends to move the 750 performance up a few notches by doubling its L2 cache and bumping up the frequency by 200 MHz. That will still put its performance below the G4 due to Altivec. If IBM does add Altivec to a 750 chip design, by that time Motorola will have put two G4 processors on one chip.

And in the meantime, the G5 will be down to 90 nm and across the entire product line, and the G6 will be released. So it's useless to bother with the G3 and G4.

Originally posted by the Motorola Information Minister
Are you beginning to see how Apple is coordinating the chip development from both IBM and Motorola? Instead of duplicate chip performances from the two companies, Apple is making sure that each chip fits in a different Mac product. But that still doesn't justify IBM staying below the performance of Motorola's chips if indeed Motorola was doing such a lousy job for Apple.

If Apple was already dedicated to Altivec, then they couldn't move away from it with a shift back to the G3. Besides, the G3 was still an inferior design to the G4. Apple didn't *want* IBM to outperform the G4.

Originally posted by the Motorola Information Minister
It's more like Intel was simply running away from the competition and Motorola was tagged as the bad guy in all this since they were the preemmenent chip maker for Apple.

Intel was running away because Motorola was stuck at 500 MHz.

Originally posted by the Motorola Information Minister
If Apple would have kept a 30-50% market share, then Motorola would have little trouble keeping up with Intel. Instead Apple chose a road that started in the early 80s that lead to a falling market share, which in turn led to them falling behind in performance and prices compared to the much larger Windows computer makers.

Funny then how the PowerPC always outperformed the competition, until the G4 at least.

Originally posted by the Motorola Information Minister
Apple is already working closely with IBM as I am sure they are doing with Motorola. Jon Rubenstein even stated that Apple worked with IBM for three years to get the 970 in production.

Yes, but I'm saying that, just like when Motorola and IBM collaborated on chip design, Apple and IBM should collaborate the same way, with Apple doing some actual chip design work.

Originally posted by the Motorola Information Minister
Apple needed to step up in performance in order to keep up with Intel's Pentium 4 processors and neither of the PowerPC chip manufacturers that would sink money into designing a chip architecture just for Apple's personal computer market.

Which is why it's necessary to use Power deriviatives, like Apple's doing with IBM.

Originally posted by the Motorola Information Minister
Apple's discussions with IBM probably were more along the lines of trying to design the 970 to keep pace with Intel's Pentium chips rather than trying to beat Motorola's G4. In order to accomplish that Apple probably insisted that Altivec be included so that the 970 would have a speed advantage in vector processing if not in integer or frequency performance.

Or floating point for that matter.

Of course they didn't approach it in terms of beating the G4. It wasn't hard to beat the G4. But Apple still wanted the G3 kept below the G4 for business reasons. And they wanted the 970 to compete with the Pentium IV.

Originally posted by the Motorola Information Minister
If Altivec has grown outmoded, then IBM should have come up with a different vector unit for the 970 instead of using Altivec.

So everyone has to rewrite their apps again? Altivec isn't a bad technology. But then again, it's not proof that Motorola is a good chip designer either. It's proof that they were a good designer four years ago, but today, they're not the same company. Pointing to Altivec to say that they are is a fallacy.

Originally posted by the Motorola Information Minister
The truth is Intel has simply stepped up the pace to the degree that few chip makers can even hope to keep up without losing money at it.

It's hard to keep up when you're stuck at 500 MHz for a year.

Originally posted by the Motorola Information Minister
The dual 2 Ghz G5 Power Mac is priced three hundred dollars more than the dual 1.42 GHz Power Mac that it replaced. That alone should give an indication that Apple is paying about the same or slightly more for the 2 Ghz G5 than for the 1.42 GHz G4.

You don't think the cost of developing a completely new motherboard and using Serial ATA, PCI-X, and AGP 8x had anything to do with it?

Originally posted by the Motorola Information Minister
So no, I'm not exactly making that up, it's just a rather educated guess as to what Apple is paying for processors. But, there is abosolutely no way that Apple is paying under a hundred dollars per chip for G5 processors. IBM would not be stupid enough to sell it to them for that price when Intel gets up to $600 for each topend Pentium 4 or Xeon processor.

Fair enough. When it was new, how much did the G3 processor cost? How about competing Pentium II's? It was perfectly plausible to use the G3 across all product lines, remember.

Phinius

Aug 2, 2003, 09:24 PM

Originally posted by Phil Of Mac

Itanium has been a spectacular failure so far. Regardless, the PowerPC has other advantages, stemming from its superiority as a chip architecture to x86.

For all the superiority that the PowerPC has in comparison to the X86, Intel is still able to get the X86 chips to a higher performance level than even the 970.

Intel is willing to spend ten years creating a market for the Itanium processors with a belief that they can use the same off-the-shelf sales idea for the server market. Recent announcements that IBM and Dell will use Itanium indicates a growing market.

While x86 is attempting to overcome its disadvantages, IBM and Apple can spend time and money working on the advantages of PPC.

Unfortunately IBM and Apple cannot spend more time and money than Intel, because Intel is bigger than the next three or four chip makers combined and IBM is not even in the top ten chip producers. Intel simply has much deeper pockets than IBM or Apple can spend on chip design and manufacture.

Altivec has indeed not advanced since 1999. That's because it's in many ways a supplemental instruction set. If they wanted to advance it, they'd have to change that instruction set, and everyone would have to rewrite everything.

Boy, I've never heard so many excuses for why a company cannot or will not pass a competitor in performance. You seem to look at this as simply a good guy, bad guy scenerio. IBM is totally good and Motorola is totally bad, even though IBM has years of opportunity to pass Motorola in desktop processors performance. Ultimately IBM did pass Motorola, but only by make a version of a chip designed for the highend server market down to the PowerPC desktop market.

IBM's not interested in adding supplemental instruction sets. They're interested in making a better processor. All the better for them.

Why should IBM even try to come up with a better design than Altivec when Altivec is already faster than any other vector design for personal computers?

IBM has said nothing about it that I know of, but reportedly it'll be here around January or so.

It takes time and money to move to a updated processor technology. It's more than likely that IBM will move the 970 to a smaller process size before a Power5 derived version comes out. If IBM moved the Power5 design over to the G5 a few months after the 970 was released, then IBM would have had little time to recoup the investment that was made in the Power4 derived 970 chips.

The G5 can be changed so it has better power use. That'll probably happen at 90 nm.

The frequencies that Apple now uses for the G5 will uses a lot less power on a smaller 90-nm process. There might be a 1.8 GHz 970 that can run at about 20 watts on the smaller process. Still the Pentium M should be well beyond 2 Ghz by that time.

As for reasons to abandon the G3 and G4, well, the G3 is too far behind in terms of technology, while the G4 is made by Motorola, who has been proven to be unreliable.

I haven't seen much proof that Motorola has been unreliable in recent years other than that they cannot match the speeds that Intel is putting in their chips. But that's not an indication of being unreliable. That could simply be due to a lack of funds to keep up with the competition, which is more likely the case for much of the disappointment.

Additionally, with the G3 we saw Apple move to a "one processor design for all computers" model. This was so even the lowly iMac had the same type of ass-kicking processor that the most powerful Power Mac had (albeit at a lower frequency).

It's difficult now to have one processor that works extremely well in several different markets. The server market has different demands for chip performance than the desktop and the notebook market has the requirement of low power use which is not needed in the desktop market.

Given Motorola's production problems with the G4 (they've never been able to produce enough), Apple couldn't do that with the G4. The G5 will undoubtedly be plentiful enough to install across the line.

If Motorola has not been able to produce enough G4 processors, then it necessarily follows that Apple is always running behind on G4 equipped Mac orders. The reality is that Apple usually only falls behind in filling G4 Mac orders when a update or new model is introduced.

Regardless, AMD processors actually use a proprietary AMD instruction set. There's an x86 frontend attached to the processor which translates from AMD to x86.

AMD pays Intel a license fee to use the X86 instruction set. It's difficult just from that extra financial burden to keep up with Intel.

I'm willing to bet that with the technology sharing AMD has done with PowerPC chip makers, the internal AMD instruction set is not that dissimilar from PPC. An AMD to PPC translation frontend, much like the AMD to x86 frontend, may not be unfeasible.

What technology sharing? IBM and AMD have entered into a collaboration where IBM shares chip making technology with AMD and Motorola belongs to the HyperTransport consortum. But that does not mean that the chip designs share a whole lot in common.

If the AMD instruction set is more similar to PPC than we thought, then AMD might just switch to the PPC instruction set, and design a PPC to x86 frontend.

I don't believe that AMD would be too interested in switching to the PPC instruction set when the X86 market is much bigger and more lucrative financially.

Apple's already consulted with IBM and Motorola on the design of the G3. Without a doubt, Apple already has some people who know something about chip design. If the reports of Motorola chip designers going to Apple is true, then perhaps Apple is going to design processors.

Have you noticed how Apple doesn't design any hardware other than the box and motherboard? It's just much less expensive to leave hardware design and manufacturing to companies that can make them in much larger quantities and Apple desperately needs to lower costs for production, not increase them.

Apple already pays the cost for IBM or Motorola to design their processors. Neither IBM or Motorola wants to produce processors just for Apple. Apple does. And if Apple can move the costs in-house for chip design, then they only have to rely on IBM or whoever for production. There's also a possibility of Apple and IBM doing joint processor design, much like IBM and Motorola once did.

It's cheaper for Apple if they can get a processor design that is made for a larger market than just their use. The 970 will also be used by IBM and it is derived from the Power4. That spreads the cost of developing the core design over a much larger sales than if Apple went it alone. Also, Motorola's designs that Apple uses are also sold to other manufacturers, which in turn lowers the cost per each chip for Apple. Unfortunately for Motorola there isn't to much demand for a high performance PowerPC processor other than IBM or Apple. Since its unlikely that IBM would use anything that Motorola comes up with, that leaves only Apple for a computer customer of the chip.

You know what, though, you've missed the number one point. In this recession, only two PC makers are profitable: Dell and Apple.

Apple has been able to make a profit in spite of falling sales and a lack of profit on computer sales (Apple has been making most of their profit off of interest on savings for two years in a row)

Also Dell grew by about 28% in personal computer sales this past quarter in the United States while Apple's sales shrunk.

With G5's across the line, Apple will have the price-performance advantage over almost every other PC.

If anything, in the foreseeable future Apple will only have the possibility of a price/performance advantage against the higher priced Intel chip equipped computers. It would be extremely difficult at this point for Apple to have a better price/performance than the low priced Wintel computers without resorting to using the extra income from software and services to offset the losses from lowering hardware prices.

Within a short time, I think IBM can surpass Intel in performance in a clear and significant way. The PowerPC has almost always outperformed Intel. The one clear exception--the G4--was the fault of Motorola. With the G5 and the partnership with IBM, Apple's future is bright indeed.

If a companies sales have fallen significantly in comparison to the overall market through the years it doesn't make sense to believe that the hardware designed for their products will outperform the product that has the majority of sales. Motorola simply does not have the money to keep up with Intel on the desktop and Apple has very little chance of supplying them with enough market share to do that.

The G5 has a very good chance of outperforming the Pentium chips in the next year. However, even with that, Apple's Fred Anderson stated that the G5 Power Mac would probably never reach the sales level that the G4 Power Mac had years ago. One of his reasons for that happening is a shift in sales to notebook computers. However the overall personal computer market is growing and not shrinking. Dell and HP are increasing sales at a rate beyond the average for the overall personal computer market. So if Apple's sales are depressed mainly due to the lack of speed from Motorola, then having the G5 should allevitate the sales slide and even gain a significant market share back. Unfortunately it doesn't look like Fred Anderson believes that will happen.

Phinius

Aug 2, 2003, 09:53 PM

Originally posted by Phil Of Mac

Motorola isn't quite reliable, now is it? So what if they do? IBM already has PowerPC 970's running on their fabrication facility in New Jersey!

Apple announced G5 Power Macs well ahead of actually shipping them. It could very well be that Apple could announce Macs with 7457 processors about the same time that the G5 Power Macs are being shipped. Which would be in the next two weeks or so. That could coincide with the official announcement that some Best Buy stores will be carrying Mac computers. Hmmm...do you think that Apple would put computers into Best Buy stores that would be outmoded in a handful of weeks. It's more likely that Apple coincided the introduction in Best Buy stores with much more competitive computers being made.

So Motorola now has 90 nm *this summer*? When Intel won't have it until the end of the year? I thought you said Motorola couldn't keep up with Intel!

Those are test chips, PowerPC production is not scheduled to begin until next year.

They promised it would be bumped up to May. They lied.

Where did Motorola state this? You are saying they went from stating March samples to full production in May? That simply is totally unrealistic to expect that and its more than likely to be just a rumor. It takes months and not a few weeks to get a sample chip into full production.

Intel's just spewing out some FUD. They have a vested interest in derailing other chip makers.

Intel is the leader in getting new process sizes into production. They are months ahead of other manufacturers.

IBM right now can make money from other chip designers going to IBM for fabrication. If IBM does that, then they'll have a profitable business.

It's tough to say at what point IBM would make money on the Fishkill facility. There is a glut of chip manufacturing capability right now due to smaller process sizes creating a higher production limit in each facility and over expectations in chip sales.

Phinius

Aug 2, 2003, 10:42 PM

Originally posted by Phil Of Mac

I don't buy those rumors and neither do you.

If these rumors are true, then its likely that a Altivec equipped G3 maybe up to a year away, so its a little to early to get expectations up.

IBM doesn't use the G4. In fact, IBM doesn't use the G3 either. In fact, I have no damn clue what you're trying to say.

I meant to say Apple's use of the G3 and G4.

And in the meantime, the G5 will be down to 90 nm and across the entire product line, and the G6 will be released. So it's useless to bother with the G3 and G4.

It would be useless to bother with the G3 and G4 if Apple was making all three chips. The G3 and G4 still have some uses for Apple competing some of the Intel designs. There are low and ultra low watt Pentium chips that the G3 would stack up well against and the G4 can be quite competitive with a dual-core design. You have to remember that Apple is up against the Pentium M and Celeron chips also. The G3 and G4 maybe a better match against these chips in some cases than the 9XX series of chips.

Intel was running away because Motorola was stuck at 500 MHz.

Intel started running away from Apple after Microsoft came out with Windows 95. With a graphical user interface much like Apple's there was much less reason to buy a Mac. The money from Windows 95 sales created a huge advantage for Intel and the rate of momentum just kept increasing with Windows 98.

Motorola had to have started on the G4 design well over a year before it came out. The G4 started out at a top speed of 500 MHz, the Pentium 4 started out at a top speed of 1.5 GHz. That's a tripling of frequency over the G4 design that Motorola could not overcome without coming out with a whole new chip architecture. That wasn't about to happen since chip architectures typically last about 4-5 years. Motorola was stuck running behind the Pentium 4 yet even still they were able to almost triple the frequency of the G4 since its introduction. Apple knew that they had to take a giant step up in performance to try and catch up to the overwhelming advantage that the Wintel side had with much bigger sales. The only realistic way to do that in a PowerPC chip design was to convince IBM to create a smaller version of the Power4 processor.

Funny then how the PowerPC always outperformed the competition, until the G4 at least.

The pace of development on the Wintel side picked up as the sales increased. It was due in no small part with Microsoft closing the gap with the Mac user interface. Motorola simply wasn't getting enough money flowing in from sales to Apple to keep up with the increasingly deep pockets of Intel.

Apple also knew that they were in trouble on the operating system side and that is why they spent a large sum of money to buy up an existing OS to replace what they had. It wasn't just in processors that Apple was falling behind, it was the whole ball of wax.

Yes, but I'm saying that, just like when Motorola and IBM collaborated on chip design, Apple and IBM should collaborate the same way, with Apple doing some actual chip design work.

Which is why it's necessary to use Power deriviatives, like Apple's doing with IBM.

That certainly spread out the cost of development over a larger volume of processors both for the PowerPC and Power chips. In other words IBM's server platform is joining forces with Apple's desktop platform. Now IBM can spend more money to speed up development for the Power processors and the PowerPC side can take advantage of some of the strides made for large servers. Luckily for Apple that IBM decided to make the Power4 a dual processor design from the beginning. If they had decided on a uniprocessor design then it would have been much more difficult to shrink it enough to get a low cost design for Apple's use. Apple would have been stuck with no place to turn too except perhaps Intel.

So everyone has to rewrite their apps again? Altivec isn't a bad technology. But then again, it's not proof that Motorola is a good chip designer either. It's proof that they were a good designer four years ago, but today, they're not the same company. Pointing to Altivec to say that they are is a fallacy.

Motorola is a leader in chip manufacturing and embedded processors sales. Unfortunately Apple doesn't have the sales for Motorola to be the leader in desktop processors. If Apple would have chosen a different route in the early 80s it might be a different story today with Motorola being top dog.

It's hard to keep up when you're stuck at 500 MHz for a year.

It's also hard to keep up when Windows 95 and 98 came out.

Fair enough. When it was new, how much did the G3 processor cost? How about competing Pentium II's? It was perfectly plausible to use the G3 across all product lines, remember.

It's perfectly plausible that IBM could find ways to reduce the costs of 9XX production, but its tough to do when they must also try and keep up the highend Pentium processors. Dramatically shrinking the size of the chip would reduce costs, but it would also limit performance. Intel reduces costs on the Pentium design by reducing cache size and bus performance. That product is the Celeron.

avus

Aug 2, 2003, 11:02 PM

Get a life, people. :rolleyes:

Phil Of Mac

Aug 2, 2003, 11:38 PM

Originally posted by the Motorola Information Minister
For all the superiority that the PowerPC has in comparison to the X86, Intel is still able to get the X86 chips to a higher performance level than even the 970.

At this point in time, yes. In the near future, due to advantages inherent to PowerPC, IBM will surpass them. The PowerPC is what gives the Mac an advantage, because despite Intel's money, they're working with ancient technology

Originally posted by the Motorola Information Minister
Boy, I've never heard so many excuses for why a company cannot or will not pass a competitor in performance. You seem to look at this as simply a good guy, bad guy scenerio. IBM is totally good and Motorola is totally bad, even though IBM has years of opportunity to pass Motorola in desktop processors performance. Ultimately IBM did pass Motorola, but only by make a version of a chip designed for the highend server market down to the PowerPC desktop market.

And that's cheating?

Listen to me: THE G3 WAS NEVER MEANT TO OUTPERFORM THE G4. NO ONE EVER ASKED IBM TO MAKE A G3 THAT WOULD OUTPERFORM THE G4, BECAUSE THAT WOULD BE REDUNDANT. THE 970 WAS ALREADY IN DEVELOPMENT.

Originally posted by the Motorola Information Minister
Why should IBM even try to come up with a better design than Altivec when Altivec is already faster than any other vector design for personal computers?

You were the one saying that IBM would. I don't have any problem with Altivec.

Originally posted by the Motorola Information Minister
It takes time and money to move to a updated processor technology. It's more than likely that IBM will move the 970 to a smaller process size before a Power5 derived version comes out. If IBM moved the Power5 design over to the G5 a few months after the 970 was released, then IBM would have had little time to recoup the investment that was made in the Power4 derived 970 chips.

I'm predicting 18 months until the Power5 derivative comes out.

Originally posted by the Motorola Information Minister
The frequencies that Apple now uses for the G5 will uses a lot less power on a smaller 90-nm process. There might be a 1.8 GHz 970 that can run at about 20 watts on the smaller process. Still the Pentium M should be well beyond 2 Ghz by that time.

So? The PPC is faster at any given clockspeed than the Pentium. Obviously that doesn't mean that a 500 MHz G4 will outperform a 1 GHz Pentium 4, but at 1.8 GHz vs. 2 GHz, the PPC would be very competitive.

Originally posted by the Motorola Information Minister
I haven't seen much proof that Motorola has been unreliable in recent years other than that they cannot match the speeds that Intel is putting in their chips. But that's not an indication of being unreliable. That could simply be due to a lack of funds to keep up with the competition, which is more likely the case for much of the disappointment.

I already listed the proof, your blindness to it is not my failt.

Originally posted by the Motorola Information Minister
It's difficult now to have one processor that works extremely well in several different markets. The server market has different demands for chip performance than the desktop and the notebook market has the requirement of low power use which is not needed in the desktop market.

At 90nm, no one really worries about it anymore.

Originally posted by the Motorola Information Minister
If Motorola has not been able to produce enough G4 processors, then it necessarily follows that Apple is always running behind on G4 equipped Mac orders. The reality is that Apple usually only falls behind in filling G4 Mac orders when a update or new model is introduced.

Which is where a good part of the demand is. Once they have the chips fabricating at a good rate, it can keep up with sales. But whenever the chip is upgraded and they start a new run through fabrications, they always have problems.

Originally posted by the Motorola Information Minister
What technology sharing? IBM and AMD have entered into a collaboration where IBM shares chip making technology with AMD and Motorola belongs to the HyperTransport consortum. But that does not mean that the chip designs share a whole lot in common.

Motorola and AMD have shared some technology in the past.

Originally posted by the Motorola Information Minister
I don't believe that AMD would be too interested in switching to the PPC instruction set when the X86 market is much bigger and more lucrative financially.

If you were paying attention you'd know that I was suggesting that AMD could design a chip that would be compatible with either instruction set after simply changing the frontend.

Originally posted by the Motorola Information Minister
Apple has been able to make a profit in spite of falling sales and a lack of profit on computer sales (Apple has been making most of their profit off of interest on savings for two years in a row)

Apple's losing money on computer sales? Where do you get this?

Originally posted by the Motorola Information Minister
If anything, in the foreseeable future Apple will only have the possibility of a price/performance advantage against the higher priced Intel chip equipped computers. It would be extremely difficult at this point for Apple to have a better price/performance than the low priced Wintel computers without resorting to using the extra income from software and services to offset the losses from lowering hardware prices.

I don't know, an eMac with a G5 at a low clockspeed 9 months from now would probably still be belew $1000. While it wouldn't be $600, the ratio of price to performance would still be better.

Apple's here to make full-featured computers. We aren't competing with eMachines.

Originally posted by the Motorola Information Minister
The G5 has a very good chance of outperforming the Pentium chips in the next year. However, even with that, Apple's Fred Anderson stated that the G5 Power Mac would probably never reach the sales level that the G4 Power Mac had years ago. One of his reasons for that happening is a shift in sales to notebook computers.

The shift in sales to notebook computers is the exact strategy Apple is seeking to capitalize on, if only the 7457's would arrive on time.

Phil Of Mac

Aug 2, 2003, 11:54 PM

Originally posted by the Motorola Information Minister
If these rumors are true, then its likely that a Altivec equipped G3 maybe up to a year away, so its a little to early to get expectations up.

A year away from now, Apple could migrate entirely to the G5.

Originally posted by the Motorola Information Minister
The G3 and G4 still have some uses for Apple competing some of the Intel designs. There are low and ultra low watt Pentium chips that the G3 would stack up well against and the G4 can be quite competitive with a dual-core design. You have to remember that Apple is up against the Pentium M and Celeron chips also. The G3 and G4 maybe a better match against these chips in some cases than the 9XX series of chips.

A 90nm G5 would be power-efficient enough to slay the Pentium-M. As for the Celeron, Apple has never competed in the "cheap crap to check your email with" market.

Originally posted by the Motorola Information Minister
Intel started running away from Apple after Microsoft came out with Windows 95. With a graphical user interface much like Apple's there was much less reason to buy a Mac. The money from Windows 95 sales created a huge advantage for Intel and the rate of momentum just kept increasing with Windows 98.

Uh, you do realize I'm talking about processor performance and not market share? Windows 95 didn't give Intel an edge in processor performance. Keeping the G4 at 500 MHz did.

Originally posted by the Motorola Information Minister
Motorola had to have started on the G4 design well over a year before it came out. The G4 started out at a top speed of 500 MHz

450 MHz. It was supposed to come out at 500 MHz, but an "errata" in the chip design forced a ceiling of 450 MHz.

Originally posted by the Motorola Information Minister
the Pentium 4 started out at a top speed of 1.5 GHz.

The Pentium 4 was released a year after the G4. The fact that the G4 went an entire year at the same speed was Motorola's fault.

Originally posted by the Motorola Information Minister
The pace of development on the Wintel side picked up as the sales increased. It was due in no small part with Microsoft closing the gap with the Mac user interface.

Microsoft didn't close the gap.

Originally posted by the Motorola Information Minister
Apple also knew that they were in trouble on the operating system side and that is why they spent a large sum of money to buy up an existing OS to replace what they had. It wasn't just in processors that Apple was falling behind, it was the whole ball of wax.

At the release of the PowerPC in 1994, it outperformed the Pentium. This lasted for several years. In 1996, Apple almost died as a company. In 1997, the turnaround began with the G3, which outperformed the Pentium II. In 1998, the iMac started Apple's recovery. In 1999, the G4 was released at 450 MHz, still competitive with the faster-clocked yet still comparable in speed Pentium III of the time. Then the G4 stayed at 500 MHz for a year. That is when the Pentium III and 4 gained a significant performance advantage.

Originally posted by the Motorola Information Minister
It's also hard to keep up when Windows 95 and 98 came out.

Well, since Windows 95 has such an effect on processor speed :rolleyes:

Phinius

Aug 3, 2003, 03:16 AM

Originally posted by Phil Of Mac

At this point in time, yes. In the near future, due to advantages inherent to PowerPC, IBM will surpass them. The PowerPC is what gives the Mac an advantage, because despite Intel's money, they're working with ancient technology

How many years now has Intel stayed ahead of the PPC using so called ancient technology?

Listen to me: THE G3 WAS NEVER MEANT TO OUTPERFORM THE G4. NO ONE EVER ASKED IBM TO MAKE A G3 THAT WOULD OUTPERFORM THE G4, BECAUSE THAT WOULD BE REDUNDANT. THE 970 WAS ALREADY IN DEVELOPMENT.

Let me get this straight. Apple let their sales slide for three years because Motorola did not supply them with an adequate amount of fast processors in a timely manner. Also Apple did not want to get faster processors from IBM because that would surpass those inadequate G4 processors from that awful manufacturer Motorola. Well now that sure involves a lot of business sense.

The truth is that Microsoft and Intel picked up the pace and left Apple behind with dropping sales and a ever smaller market share. You can lay most of the blame on Motorola if you want, but the fact is that Microsoft and Intel played a much better hand than Apple and won.

I'm predicting 18 months until the Power5 derivative comes out.

That's a reasonable estimate and if it arrives sooner then you will be pleasantly surprised.

So? The PPC is faster at any given clockspeed than the Pentium. Obviously that doesn't mean that a 500 MHz G4 will outperform a 1 GHz Pentium 4, but at 1.8 GHz vs. 2 GHz, the PPC would be very competitive.

The trouble with that is the Pentium 4 is at 3.2 GHz and the fastest PowerPC processor is running at 2 GHz.

At 90nm, no one really worries about it anymore.

Heat and power use is becoming one of the biggest concerns of chip designers. Making a smaller chip concentrates the power use into a smaller area. The G5 is probably limited to about 1.2 GHz right now without resorting to a heatsink and a heatsink would make it too thick to fit into a slim cased PowerBook. That's the reason the Pentium 4s are only in thicker notebook designs.

Which is where a good part of the demand is. Once they have the chips fabricating at a good rate, it can keep up with sales. But whenever the chip is upgraded and they start a new run through fabrications, they always have problems.

It's tough to get production up to meet short spikes in demand. When sales fall a couple of months later then Apple would be stuck with too much capacity if they quickly met the demand of new products.

If you were paying attention you'd know that I was suggesting that AMD could design a chip that would be compatible with either instruction set after simply changing the frontend.

Chip design is not that simple. It took three years of working with IBM to get a PowerPC version of the Power4. At first glance it would seem fairly easy to just split the Power4 in two and come up with a PowerPC version in a few short easy steps. If that is true then why did it take three years?

Apple's losing money on computer sales? Where do you get this?

From Apple's own quarterly reports. In fact of the last four fiscal quarters Apple lost money on overall sales minus expenses in three of those quarters. Of the last eleven quarters Apple lost money in six of them on sales minus operating expenses. The losses started around the time when the Pentium 4 was introduced. In fact Apple's sales plunged so far and fast in the Christmas quarter of 2000 that they lost 420 million dollars from sales minus expenses in that quarter. Apple's sales have never even come close to the level that they were at before that quarter.

I don't know, an eMac with a G5 at a low clockspeed 9 months from now would probably still be belew $1000. While it wouldn't be $600, the ratio of price to performance would still be better.

Apple needs to work on getting Macs made that sell at a much lower entry level price. Only using a G5 processor that costs two hundred or more per chip won't get them there. That cost to Apple could be at least 2/3 of the retail selling price of the computer.

Apple's here to make full-featured computers. We aren't competing with eMachines.

You may believe that Apple can survive by only serving the classes and not the masses, but Apple losing money on sales in three of the last four quarters tells me that Apple does not believe the same thing you do. If Apple did believe what you seem to then they would have kept prices at the same level or even raised them to make sure they kept a good profit on sales. Obviously Apple saw trying to keep sales up to be more important than short term profits.

The shift in sales to notebook computers is the exact strategy Apple is seeking to capitalize on, if only the 7457's would arrive on time.

Apple has been using the excuse of shifting to notebook computers because their desktop computers have been way off the mark in price/performance. Now Apple's thin and light notebook designs are being attacked by Wintel competitors using Intel's Banias chip.

tychay

Aug 3, 2003, 03:48 AM

Originally posted by Phinius
The 970 uses the same amount of SIMD instructions as the 7450. That's due to IBM using Altivec for the 970.

Both of you are right (or wrong). IBM does not use "Altivec" because "Altivec" is a registered trademark of Motorola. IBM uses "VMX" which is a 100% Altivec-compatible instruction set. Apple calls it "Velocity Engine" no matter what the vendor calls it. The three of them developed this technology together, but IBM pulled out of production for the reasons you mention.

Personally, I don't think Altivec's presence in the G5 makes a case for either side. This was a strategic error on Apple's part (pumping up Velocity for the G4) that was further compounded by developing Velocity-specific optimizations into the OS to make up for a chip that was lagging behind others (not just Intel, AMD too).

What Motorola did really bothers me but I can't bring myself to place the whole blame on them. Intel was cherry-picking their best designers and redefining the term "predatory hiring practices" and Apple had just wiped out a lucrative market just before the G4 was taped out (by shutting down the clone market). So who else buys such a chip: BeOS? LinuxPPC? Amiga?

Intel has several chip manufacturing facilities that use 300-mm silicon wafers, IBM has just one.

Intel has two 300mm fabs and were planning to build two more (one has since been cancelled, the other will go online in 2005 or 2006). One of the current fabs is development-only and the other has yet to ship a single chip out of it's doors (Fab11X). The confusion (deception?) lies in that the latter's sister-fab (Fab11) which is a 9" fab and produces the current Pentium 4.

There is a lot of industry evidence that Intel is having serious issues with their 12" fabs. Prescott (aka Pentium 5) debuting at 90nm with 11X producing all the world's P5s is not going to happen this year (it's already been delayed from 3Q to 4Q). If Prescott debuts, I'd guess that they'll use 130nm design rules while the expect to ramp it up to "real" 90nm as time goes on (a trick they did very successfully with the P4, I might add). (11 and 11X combined could satisfy all the demand for the thing for quite some time.)

With the industry demand being so depressed, I seriously question whether Intel needs a 12" fab when their 9" fabs keep up with all current demands for the P4 with capacity to spare. As you noted, they seem to be the only CPU manufacturer actually making a profit on microcomputer CPUs. (I think this might change as IBM seems to have a good setup here of migrating their POWER stuff to PowerPC at low cost.)

I'm taking you to task here because your comment is misleading. Fishkill is a superior facility because it has a proven design and production record as a 12" Fab. It is not making money most-likely because demand is so low right now, certainly lower than anyone would have predicted in 2000 when they broke ground on Fishkill. It is not, however, as unique as a lot of Mac pundits would have you believe, so I understand your frustration.

Phinius

Aug 3, 2003, 03:50 AM

Originally posted by Phil Of Mac

A year away from now, Apple could migrate entirely to the G5.

Apple could but they would probably have to raise the prices on most of those products, just like they did with the Power Mac when it moved from the G4 to the G5. Raising the minimum price level will reduce overall sales and market share. Which in turn lowers the amount of developers writing applications and raises the cost of parts and manufacturing per each unit sold.

A 90nm G5 would be power-efficient enough to slay the Pentium-M.

The Pentium-M is up to 1.7 GHz on a 130-nm process. The 970 could probably fit into a PowerBook at the 1.2 GHz frequency. That's almost a 50% frequency advantage that the Pentium-M has over the 970 for thin and light notebook use.

Moving both the Pentium-M and the 970 to a smaller 90-nm process size reduces the watt use for both designs. For notebooks the 970 would move up in frequency and also the Pentium M. I see no advantage for the 970 compared to the Pentium-M if both are moved to a smaller process size.

As for the Celeron, Apple has never competed in the "cheap crap to check your email with" market.

Maybe that's one of the reasons Apple is losing sales and market share since the average selling price of a desktop PC now is about $749, which is below the minimum price that Apple offers in a desktop computer.

Or is it that more people won't buy a product when prices are lowered significantly and the competitors do not have to reduce prices also in order to keep sales from falling? That would seem to run counter to what most economists teach.

So Sony reduced the prices for the Playstation 2, then the Xbox and Gamecube followed shortly thereafter simply because it looked like a fun thing to do at the time? Microsoft and Nintendo were not the least bit worried that Sony could take away sales from them by reducing the price of a competing product?

Uh, you do realize I'm talking about processor performance and not market share? Windows 95 didn't give Intel an edge in processor performance. Keeping the G4 at 500 MHz did.

You mean Windows 95 did not boost Microsoft OS sales considerably, which in turn lead to more sales and profits for Intel? And Windows 98 did not give Intel a big boost in sales and cash flow when more people bought Wintel computers?

450 MHz. It was supposed to come out at 500 MHz, but an "errata" in the chip design forced a ceiling of 450 MHz.

It seems that Intel had a problem similar to that when they had to recall chips. Also AMD had delayed the Opteron processor for several months and IBM kept the Power4 at about the same frequency for months and months. In fact the progression rate for frequency improvements in the Power4 chip has crept along with the Power4 going from about 1.1 Ghz at its introduction to a top speed of 1.7 Ghz now.

The Pentium 4 was released a year after the G4. The fact that the G4 went an entire year at the same speed was Motorola's fault.

You mean to say that due to Motorola keeping the G4 at 500 Mhz for a year it has had an adverse effect on Apple's computer sales every since?

Apple announced the 500 MHz Power Mac on August 31 of 1999. In the next 41 months Motorola increased the frequency by 2.84 times that.

Intel introduced the first Pentium 4 that ran up to 1.5 GHz on November 20th in the year 2000. Intel increased it's frequency in the next 31 months by 2.1 times that.

Motorola may not have gained any ground in frequency increases with the G4 against the Pentium 4, but they were able to stay at about the same pace of frequency improvements that Intel made with the Pentium 4.

Microsoft didn't close the gap.

You mean DOS was just as close to the Mac OS in performance as Windows is compared to OSX today? That gap has not narrowed considerably and made the Microsoft operating system much more appealing compared to the Mac?

At the release of the PowerPC in 1994, it outperformed the Pentium. This lasted for several years. In 1996, Apple almost died as a company.

Notice how that is about one year after Microsoft introduced Windows 95.

In 1997, the turnaround began with the G3, which outperformed the Pentium II. In 1998, the iMac started Apple's recovery. In 1999, the G4 was released at 450 MHz, still competitive with the faster-clocked yet still comparable in speed Pentium III of the time. Then the G4 stayed at 500 MHz for a year. That is when the Pentium III and 4 gained a significant performance advantage.

Microsoft also turned it up another notch with the introduction of Windows 98, which in turn increased computer sales and that gave more money for Intel to develop chips.

Well, since Windows 95 has such an effect on processor speed :rolleyes:

Do you believe that the amount of R&D for chip development and capital investment in fabrication facilities stays the same when the revenue from sales increases by several fold? You honestly believe that Motorola should have been able to keep pace with Intel with a increasingly smaller fraction of the money that Intel was getting from personal computer sales?

tychay

Aug 3, 2003, 04:24 AM

Originally posted by Phinius
AMD pays Intel a license fee to use the X86 instruction set. It's difficult just from that extra financial burden to keep up with Intel.

Do you have evidence of this? I'd like to see it, if you can dig it up.

Back in the old days, I believe AMD licensed the 386 and 486 designs, but the K6, Athlon, Duron, Opteron, and Athlon64 have no licensing fee paid to Intel.

There is an "extra financial burden" but it isn't in licensing fees. They have to reverse engineer the CPU every time. There is a company that will do this. I think both Intel and AMD pay this burden--yes, even Chipzilla would like to know what AMD is up to, strange as it may sound.

That's part of what makes the chip market so cool--it's really competitive. I jog by both AMD's and Intel's world HQs, which I find amusing given the rest of the world's view that they are each other's throats. My roommate from graduate school works at the AMD HQ and my current roommate works at the Intel HQ and they seem to get along fine. I tease them about their company's foibles, they tease me because I like Macs but my last two purchases had an AMD Athlon and Intel Pentium4 in them. :)

tychay

Aug 3, 2003, 04:40 AM

Originally posted by Phinius
Have you noticed how Apple doesn't design any hardware other than the box and motherboard?

Not quite true. Apple designs the ASIC controller in the G5 PowerMac. It is produced in Fishkill by IBM along with the 970s which may be the reason for your confusion.

It makes sense too since the ASIC has a lot of stuff in it that would be overkill for a blade--IBMs targeted use of the 970.

I'm sure that it is much easier to design an ASIC than a CPU. But then again, an ASIC is not exactly RAM either.

They also design a lot of peripherals like the iSight, some of the iPod (based on PortalPlayer's reference design), mouse, keyboard, wireless networking cards, etc., and various controller chips now and then. It can be argued that Apple doesn't really "design" the motherboard since these are tweaked from an IBM reference design.

Apple certainly does more hardware design than Microsoft and Dell combined. That doesn't mean they're ready for CPUs and wasn't that the point of bedding with IBM and Motorola?

IBM is extremely well positioned in the desktop CPU market. The only company better positioned right now is Intel. They're the only two companies that actually seem to cover the gamut from notebooks to 64-bit servers, but certainly not the only two companies in the market.

panphage

Aug 3, 2003, 05:04 AM

Jesus, can someone shut these two up? Or give them their own thread? Not you, tychay. Phinius and Phil of Mac. I'm glad macforia took hirself out of the spam-fest.

Phinius

Aug 3, 2003, 12:19 PM

Originally posted by tychay

Do you have evidence of this? I'd like to see it, if you can dig it up.

Here it is from a 2001 news.com article (http://news.com.com/2100-1040-257059.html). AMD signed a ten year licensing deal with Intel.

Back in the old days, I believe AMD licensed the 386 and 486 designs, but the K6, Athlon, Duron, Opteron, and Athlon64 have no licensing fee paid to Intel.

AMD made copies of the 486 and now they have their own chip design but pay a royalty fee to Intel for the right to use the X86 instruction set.

There is an "extra financial burden" but it isn't in licensing fees. They have to reverse engineer the CPU every time.

I don't see why AMD would want to reverse engineer a Intel processor since they are paying a license fee for the use of X86.

Phinius

Aug 3, 2003, 12:35 PM

Originally posted by tychay

It can be argued that Apple doesn't really "design" the motherboard since these are tweaked from an IBM reference design.

In this Arstechnica interview (http://www.arstechnica.com/cpu/03q2/ppc970-interview/ppc970-interview-2.html) IBM's Peter Sandon states Apple designed their own chipset for the 970 and IBM manufactures it. You will find the question and answer under the the 'Miscellany: bus ratios compilers, power management' heading.

IBM is extremely well positioned in the desktop CPU market. The only company better positioned right now is Intel. They're the only two companies that actually seem to cover the gamut from notebooks to 64-bit servers, but certainly not the only two companies in the market.

AMD seems much more likely to produce a higher volume of processors for notebooks, servers and the desktop than IBM will since AMD makes chips for the large Windows market.

As of yet IBM does not makes processors that are used in notebooks, servers and the desktop. Although that could change if Apple introduces a 970 equipped PowerBook and Xserve.

Phil Of Mac

Aug 3, 2003, 07:28 PM

Originally posted by Phinius
How many years now has Intel stayed ahead of the PPC using so called ancient technology?

Ever since shortly after the introduction of the G4 in 1999.

Originally posted by Phinius
Let me get this straight. Apple let their sales slide for three years because Motorola did not supply them with an adequate amount of fast processors in a timely manner. Also Apple did not want to get faster processors from IBM because that would surpass those inadequate G4 processors from that awful manufacturer Motorola. Well now that sure involves a lot of business sense.

Apple was already too devoted to Velocity Engine in their codebase to use a non-Velocity Engine chip. An IBM G3 with VE would take just about as long as the 970 to develop, as far as I can guess, so Apple found it more reasonable to go ahead and get the 970.

Originally posted by Phinius
The truth is that Microsoft and Intel picked up the pace and left Apple behind with dropping sales and a ever smaller market share. You can lay most of the blame on Motorola if you want, but the fact is that Microsoft and Intel played a much better hand than Apple and won.

If Apple's cards don't improve for a full year, then Intel and Microsoft are eventually going to play better cards.

Originally posted by Phinius
Heat and power use is becoming one of the biggest concerns of chip designers. Making a smaller chip concentrates the power use into a smaller area. The G5 is probably limited to about 1.2 GHz right now without resorting to a heatsink and a heatsink would make it too thick to fit into a slim cased PowerBook. That's the reason the Pentium 4s are only in thicker notebook designs.

Exactly. The heat signature of a 1.2 GHz 970 is identical to current G4's. It will definitely be more economical for Apple to wait for the 90nm 970 to move the G5 into PowerBooks.

Originally posted by Phinius
Of the last eleven quarters Apple lost money in six of them on sales minus operating expenses. The losses started around the time when the Pentium 4 was introduced. In fact Apple's sales plunged so far and fast in the Christmas quarter of 2000 that they lost 420 million dollars from sales minus expenses in that quarter. Apple's sales have never even come close to the level that they were at before that quarter.

Incidentally, the G4 was stuck at 500 MHz during that period. Apple had no problem slaying Pentium 1 through 3.

Originally posted by Phinius
Apple needs to work on getting Macs made that sell at a much lower entry level price. Only using a G5 processor that costs two hundred or more per chip won't get them there. That cost to Apple could be at least 2/3 of the retail selling price of the computer.

In terms of being a high-end processor, why would the G5 cost more than the G3 cost in its day? The G3 was in the entry-level iMac.

Originally posted by Phinius
You may believe that Apple can survive by only serving the classes and not the masses, but Apple losing money on sales in three of the last four quarters tells me that Apple does not believe the same thing you do. If Apple did believe what you seem to then they would have kept prices at the same level or even raised them to make sure they kept a good profit on sales. Obviously Apple saw trying to keep sales up to be more important than short term profits.

Apple's prices are not at a bad level now. They're on the high end of computer makers, but as long as Apple's low-end prices can straddle the $1000 mark, Apple will be just fine, especially if they can maintain the quality advantage.

Phil Of Mac

Aug 3, 2003, 07:46 PM

Originally posted by panphage
Jesus, can someone shut these two up? Or give them their own thread? Not you, tychay. Phinius and Phil of Mac. I'm glad macforia took hirself out of the spam-fest.

Discussing processors is spam? Then what would you call posting just to complain about it?

Originally posted by Phinius
Apple could but they would probably have to raise the prices on most of those products, just like they did with the Power Mac when it moved from the G4 to the G5.

That probably had more to do with the R&D they had to do on the architecture of the motherboard to support PCI-X, AGP 8x, Serial ATA, etc. The Power Mac G5 is unique not only because of its processor, but because as a workstation, it's a fully integrated, high-tech solution using the latest standards. Apple did a lot of work on the G5, and the higher price is to pay for all that. It probably has very little to do with the processor itself.

Originally posted by Phinius
Moving both the Pentium-M and the 970 to a smaller 90-nm process size reduces the watt use for both designs. For notebooks the 970 would move up in frequency and also the Pentium M. I see no advantage for the 970 compared to the Pentium-M if both are moved to a smaller process size.

No, but at that point at least IBM is in the race. The PPC is very capable of staying ahead of Intel in a drag race if both are actually developed at a constant rate (as opposed to being stuck at 500 MHz.)

Originally posted by Phinius
So Sony reduced the prices for the Playstation 2, then the Xbox and Gamecube followed shortly thereafter simply because it looked like a fun thing to do at the time? Microsoft and Nintendo were not the least bit worried that Sony could take away sales from them by reducing the price of a competing product?

Game consoles are an entirely different market. No game console has any definite advantage over any other console in quality, so they have to be sold at relatively the same prices. The Mac, on the other hand, has a definite quality advantage in terms of the operating system.

Originally posted by Phinius
You mean to say that due to Motorola keeping the G4 at 500 Mhz for a year it has had an adverse effect on Apple's computer sales every since?

It was the most direct cause of Apple's perfomance disadvantage between the introduction of the Pentium 4 and the G5.

Originally posted by Phinius
Apple announced the 500 MHz Power Mac on August 31 of 1999. In the next 41 months Motorola increased the frequency by 2.84 times that.

Apple introduced the 500 MHz Power Mac on August 31, 1999. Within a couple months, Motorola decreased the frequency by 10% of that. After increasing it back to 500 MHz, Motorola increased the frequency by 0% for the next year.

Originally posted by Phinius
Intel introduced the first Pentium 4 that ran up to 1.5 GHz on November 20th in the year 2000. Intel increased it's frequency in the next 31 months by 2.1 times that.

Motorola may not have gained any ground in frequency increases with the G4 against the Pentium 4, but they were able to stay at about the same pace of frequency improvements that Intel made with the Pentium 4.

And that's what they were able to do after taking an entire year off of development.

Originally posted by Phinius
You mean DOS was just as close to the Mac OS in performance as Windows is compared to OSX today? That gap has not narrowed considerably and made the Microsoft operating system much more appealing compared to the Mac?

Mac had more of an advantage over DOS than over Windows 3.1, and Windows 95 narrowed the gap, but didn't really catch up to System 7.5. Windows 98 was hardly an upgrade at all. Windows ME was a downgrade. In the meantime, Mac OS 8 gained significant ground over Windows, and a couple years later, Mac OS X put even more distance between Mac and Windows. Windows XP gained on Mac OS X slightly, but the Mac OS has always been the superior operating system.

Originally posted by Phinius
Notice how that is about one year after Microsoft introduced Windows 95.

Apple's 1996 crisis was caused by poor management more than anything else.

Originally posted by Phinius
Do you believe that the amount of R&D for chip development and capital investment in fabrication facilities stays the same when the revenue from sales increases by several fold? You honestly believe that Motorola should have been able to keep pace with Intel with a increasingly smaller fraction of the money that Intel was getting from personal computer sales?

It's not so much that Motorola hasn't been keeping up. It's their constant production shortages and complete lack of development between 1999 and 2000. If Intel were able to widen the gap through designing faster and better, then that would be fine, but that's not what happened. What happened is that Motorola stopped and after that didn't really go again.

Intel didn't outrun Motorola. Motorola collapsed and laid down on the track for a year before getting up and jogging off.

Phinius

Aug 3, 2003, 10:04 PM

Originally posted by Phil Of Mac

Ever since shortly after the introduction of the G4 in 1999.

By your own admission Intel has stayed ahead of the performance of the PowerPC processors since 1999 and the introduction of the G5 has not reversed that trend yet. It will be difficult even for IBM to have a faster processor for the desktop market than Intel. That's simply due to the economic forces at work here, Intel does a much larger volume of business on the desktop and so therefore can create a processor that is specifically designed for that market. IBM does not have the financial means to due that and so therefore they have to use a processor designed for the server market for the desktop and notebooks. That involves compromises that Intel does not have to make. Which also increases the likelihood that Apple will need another PowerPC processor in addition to the G5 in order to meet the demands of each product segment. That's probably why the G3 and G4 are still being updated by IBM and Motorola.

Notice also that Intel has yet to eliminate the Pentium III from production? Now why is that when the Pentium 4 would have seemed to be its immediate replacement? It's simply due to the fact that the Pentium III could meet some market demands where the Pentium 4 was ill equipped. Like low power use products. With the introduction of Banias the Pentium III is now scheduled to cease production.

In much the same way Apple will probably not eliminate the G4 or perhaps the G3 anytime soon. Both of those chips can fill product demands where the G5 is less capable. Like low power use and cost.

Apple was already too devoted to Velocity Engine in their codebase to use a non-Velocity Engine chip.

Apple has never stopped using non-Velocity engine chips since the G4 was first introduced. It would make sense though that if Apple intends to use the G3 for at least another year that they would want IBM to move Altivec over to it. Don't believe that Apple would need a ultra low watt PowerPC processor running at a top speed of 1.1 Ghz in January of 2004? Well, some of the tablet personal computers use Pentium-M processors that run at 900 MHz. It could be that Apple will produce a new product in January that could use a low watt G3 processor. The G3 should be much cheaper to produce than the G4 and G5 to boot.

An IBM G3 with VE would take just about as long as the 970 to develop, as far as I can guess, so Apple found it more reasonable to go ahead and get the 970.

Adding Altivec onto the G3 would not take as long to develop as the 970. IBM had to make much more extensive changes to the Power4 architecture in order to create the 970 than it would be to add a Velocity engine onto a G3 chip.

If Apple's cards don't improve for a full year, then Intel and Microsoft are eventually going to play better cards.

A one year lack of progress does not therefore become a excuse for falling behind in sales for three consecutive years. Nor did Apple maintain its market share before the G4.

The Pentium 4 did play a major role in the decline of Apple's sales in the past three years though. You can literally see Apple's sales fall in one fiscal quarter like they went off a cliff, which was also the same quarter that the Pentium 4 was first available in personal computers. Apple lost 420 million dollars from sales minus expenses in that quarter.

Intel was able to overcome the advantages of the PPC with a much higher revenue coming from a greater market share than what Motorola or IBM processors had with Macs. Intel was simply out spending on R&D and capital improvements which lead to the Pentium 4. Intel's market share advantage over AMD also was the major reason why Intel jumped in front of AMD with the Pentium 4. Intel was simply setting a pace of improvement for microprocessors that was extremely difficult for any other company to match.

Exactly. The heat signature of a 1.2 GHz 970 is identical to current G4's.

IBM's published watt usage for the G5 ranges from about 17 watts for the 1.2 GHz to 42 watts for the 1.8 Ghz version.

If you look at Motorola's website which lists the 7XXX host processors it states that the 7455 has a typical power use of 21.3 watts at 1 GHz. With a slight amount of thinking its obvious that the 32 and 35.5 watts listed below the 1GHz 7455 are for the 1.25 GHz and 1.42 Ghz processors that were used in the Power Macs. Motorola was able to boost the listed 1 GHz topend perfomance of the 7455 for Apple by raising the voltage from 1.6 v to 1.85 v and then again by adding a low-k dielectric material that can boost frequencies by up to 20%.

The 7457 running at 1.3 GHz has about a 2 watt less average power use than the low volt 970 at 1.2 GHz. It would seem plausible that Apple would put up to a 1.3 GHz 7457 in the topend PowerBooks in the next few weeks. That would make the transition to 1.2 GHz G5 PowerBooks go over pretty well with even non Mac users, since the G5 would only have a drop of 100 Mhz from the G4.

It will definitely be more economical for Apple to wait for the 90nm 970 to move the G5 into PowerBooks.

I'd bet that Apple will come out with a G5 PowerBook around January of 2004. It might be limited to 1.2 GHz, but that could tide them over until about July when the 90-nm process size might become available.

Incidentally, the G4 was stuck at 500 MHz during that period. Apple had no problem slaying Pentium 1 through 3.

That maybe true but Apple's marketshare kept falling.

In terms of being a high-end processor, why would the G5 cost more than the G3 cost in its day? The G3 was in the entry-level iMac.

Times have changed. Intel has increased it's sales volume and lowered the average price of its processors.

Apple's prices are not at a bad level now. They're on the high end of computer makers, but as long as Apple's low-end prices can straddle the $1000 mark, Apple will be just fine, especially if they can maintain the quality advantage.

If PC manufacturers lower the average price of computers then Apple must respond by lowering prices in order to not lose sales and market share. Right now Apple would have a difficult time lowering prices much since the company is just about breaking even on sales minus expenses.

Phinius

Aug 3, 2003, 11:44 PM

Originally posted by Phil Of Mac

Apple did a lot of work on the G5, and the higher price is to pay for all that. It probably has very little to do with the processor itself.

Well if higher prices are the result, then how is Apple going to displace the G4 used in the consumer Macs by the end of 2004 as you imagine they will? If Apple did that with higher costs due to a changeover to the G5 then the price of the computers would probably have to increase also. That would get Apple's average Mac prices moving in the opposite direction of the overall personal computer market, which is falling. That's a sure fire way to insure that Apple's market share falls.

It makes more sense to see that Apple will likely use the G4 and possibly the G3 through at least 2004, due to the time needed and higher costs of implementing the G5. If it is so easy to transition to the G5, then where is the announcement for the G5 Xserve or the G5 PowerBooks? Obviously if it takes time to come up with the needed motherboards etc. for a G5 computer, then it would take time for Apple to make a G5 iMac or a G5 eMac. It would also take time for IBM to ramp up the production of the G5 enough to fill every need for processors in every Mac category for Apple.

With IBM announcing the 750GX and Motorola giving the green light to the dual-core G4, it's very unlikely that Apple would move completely over to just using a G5 for every Mac by the end of 2004. Obviously though the G3 and G4 are about at the end of their life cycles and if Apple intends to keep using another PowerPC chip architecture in addition to the G5, then there should be a announcement by IBM and or Motorola sometime in 2004 about the generation of processors that will replace the G3 and or G4. The G4 and G3 are unlikely to go beyond a 90-nm process. The 750 series may even end on the 130-nm process, but a dual-core G4 chip has to be made on a process size at least as small as 90-nm. Otherwise two G4 processors would not fit on a reasonable die size.

If you believe that Apple would have no use for a dual-core G4 in 2004, you must realize that Dell still uses the Pentium III in a 1U server and blade server. In 2004 a G4 processor made on a 90-nm process size could be very competitive in a server box that is small and compact.

No, but at that point at least IBM is in the race. The PPC is very capable of staying ahead of Intel in a drag race if both are actually developed at a constant rate (as opposed to being stuck at 500 MHz.)

The problem is that the PPC and Intel X86 processors will not both be developed at the same rate. Intel has much deeper pockets to devote to increasing the performance of desktop or notebook processors than IBM can muster. As a matter of fact Intel can sink much more money into developing server processors than IBM can afford to.

No game console has any definite advantage over any other console in quality, so they have to be sold at relatively the same prices. The Mac, on the other hand, has a definite quality advantage in terms of the operating system.

If Apple does not have to compete has much on price as the other companies making personal computers, then why is Apple reducing prices to a point that they are making little if any profit off of sales? Your argument does not hold water. Apple lost money in 6 of the last 11 fiscal quarters and their computer sales have fallen for three years in a row. Whereas the overall personal computer market had one year of falling sales in the last four years. Other than that the overall personal computer market has been growing.

It was the most direct cause of Apple's perfomance disadvantage between the introduction of the Pentium 4 and the G5.

You mean Intel getting a ever increasing share of the money going to processors for personal computers had no influence on their ability to make major capital improvements or pick up the pace of R&D for chip development? Microsoft played a major role in getting a bigger share of the personal computer market for Intel.

And that's what they were able to do after taking an entire year off of development.

Motorola was still able to keep pace with Intel's frequency advances for the Pentium 4, even though the G4 stayed at 500 MHz for one year. I'd give Motorola a lot of credit for finding ways to make significant advances in frequency for the G4 when it went from 500Mhz to 1.42 GHz.

Mac had more of an advantage over DOS than over Windows 3.1, and Windows 95 narrowed the gap, but didn't really catch up to System 7.5.

The mere fact that Microsoft was able to narrow the gap against the Mac made for a lot less reasons to buy a Mac. That undoubtedly played a major role in Apple losing market share.

Windows XP gained on Mac OS X slightly, but the Mac OS has always been the superior operating system.

Windows XP is more than a slight upgrade from Windows 98 or ME. The fact remains that Windows created a lot less differentiation between the Mac and Microsoft's OS and so therefore there was much less need for people to buy a Mac over the many PC alternatives.

Apple's 1996 crisis was caused by poor management more than anything else.

So its not now poor management when Apple's sales and market share have fallen for three straight years in a row?

And are you still willing to blame almost the entire Mac sales decline since the G4 on Motorola? Even though Apple's Fred Anderson states he does not believe that the G5 Power Mac will every obtain the peak sales level of the G4 Power Mac? If the fall in Apple's sales was mostly attributable to a under performing processor, then it stands to reason that a much more competitive processor will reverse that. Fred is stating something entirely different. It's more like the G5 Power Mac will not be competitive enough in price/performance to increase sales to a level of three or four years ago. Meanwhile, Dell's PC sales are increasing far beyond what they were three or four years ago.

It's not so much that Motorola hasn't been keeping up. It's their constant production shortages and complete lack of development between 1999 and 2000. If Intel were able to widen the gap through designing faster and better, then that would be fine, but that's not what happened. What happened is that Motorola stopped and after that didn't really go again.

You are ignoring the facts. Motorola moved the G4 from 500 Mhz to 1.42 GHz since it was first introduced. That's it hardly stopping and not moving again. Intel has not been able to advance the Pentium 4 at a greater frequency % increase that what Motorola has done in that amount of time.

Intel didn't outrun Motorola. Motorola collapsed and laid down on the track for a year before getting up and jogging off.

If Apple did work with IBM for three years getting a PowerPC version of the Power4 as Jon Rubenstein states, then Motorola almost certainly got wind of it somehow. It would not make much economic sense for Motorola to sink loads of money into creating the next generation of processor for Apple's use if Apple intended to use IBM's spinoff of the Power4. It would be very evident to Motorola that their precentage of the Mac market would shrink with a cut down version of the Power4 in use by Apple. Motorola would be limited to probably no more than half of Apple's shrinking market share and they could see that their processors would be used only in the lower priced consumer Macs. Which means Motorola would have to sell processors at the lower prices to Apple and IBM would get the higher prices.

If I were running Motorola at the time of hearing about Apple working with IBM on a derivitive of the Power4, then I would have likely wanted to kill any G5 host processor project right then. It simply would not make much economic sense to spend a lot of money on developing a new processor when the main company you intend to sell it to is likely to buy most of their processors from another company. Motorola simply did not have the financial means to match the performance of a Power4 derived PowerPC processor. The 970 limited Motorola to being second banana to what IBM could offer Apple. So with smaller sales would come less financial incentive to spend on making processor improvements to that market and that is just what Motorola did.

It would make much more sense for Motorola to spend money updating the G4 processor with new technologies that could be transfered throughout the entire PowerPC line. And that's just what they have been doing, instead of sinking money into a project that would have limited use outside of Apple's market.

Oh, and here's a link to a G4 Power Mac review (http://www.aceshardware.com/read.jsp?id=50000333) in Aceshardware, where the author has some quite favorable comments about the G4. So is the G4 as fast as the fastest Pentium chips? No, but then it doesn't have to be, that's the role of the G5. The G4 can be used in products that compete against the Pentium-M and Celeron chips. The G5 is simply another choice for Apple that ranks above the G4 and the G4 ranks above the G3, but the G5 does not necessarily have to replace either of those chips immediately.

Phil Of Mac

Aug 4, 2003, 12:20 AM

Originally posted by Phinius
Well if higher prices are the result, then how is Apple going to displace the G4 used in the consumer Macs by the end of 2004 as you imagine they will? If Apple did that with higher costs due to a changeover to the G5 then the price of the computers would probably have to increase also. That would get Apple's average Mac prices moving in the opposite direction of the overall personal computer market, which is falling. That's a sure fire way to insure that Apple's market share falls.

The higher price of the Power Mac G5 was due to a revamped architecture. Applying that architecture to other products won't necessarily mean a price rise of any substance, because the PMG5 price rise was intended to be what pays for the development of G5 architecture. Additionally, Apple may not revamp the architecture of the consumer products, i.e. no price rise.

Originally posted by Phinius
It makes more sense to see that Apple will likely use the G4 and possibly the G3 through at least 2004, due to the time needed and higher costs of implementing the G5. If it is so easy to transition to the G5, then where is the announcement for the G5 Xserve or the G5 PowerBooks? Obviously if it takes time to come up with the needed motherboards etc. for a G5 computer, then it would take time for Apple to make a G5 iMac or a G5 eMac. It would also take time for IBM to ramp up the production of the G5 enough to fill every need for processors in every Mac category for Apple.

Time, yes. But eventually, Apple probably wants to move to G5's across the board.

Originally posted by Phinius
If you believe that Apple would have no use for a dual-core G4 in 2004, you must realize that Dell still uses the Pentium III in a 1U server and blade server. In 2004 a G4 processor made on a 90-nm process size could be very competitive in a server box that is small and compact.

Low-end Xserve perhaps? Well, considering that the 970 is being used in IBM blade servers, I think it's also a good choice for the Xserve.

Originally posted by Phinius
The problem is that the PPC and Intel X86 processors will not both be developed at the same rate. Intel has much deeper pockets to devote to increasing the performance of desktop or notebook processors than IBM can muster. As a matter of fact Intel can sink much more money into developing server processors than IBM can afford to.

Intel may have more money, but IBM has a better architecture to develop. The PPC has kept pace with x86 and even surpassed it before, despite Apple's small market share and near-death experience. Well, Apple's coming back now, and so is the Mac.

Originally posted by Phinius
If Apple does not have to compete has much on price as the other companies making personal computers, then why is Apple reducing prices to a point that they are making little if any profit off of sales? Your argument does not hold water. Apple lost money in 6 of the last 11 fiscal quarters and their computer sales have fallen for three years in a row. Whereas the overall personal computer market had one year of falling sales in the last four years. Other than that the overall personal computer market has been growing.

I'm not saying Apple doesn't have to compete on price at all. I am saying that Apple has a quality advantage over the competition that justifies the slightly higher prices in place now.

Originally posted by Phinius
You mean Intel getting a ever increasing share of the money going to processors for personal computers had no influence on their ability to make major capital improvements or pick up the pace of R&D for chip development? Microsoft played a major role in getting a bigger share of the personal computer market for Intel.

Considering that from 1994 to 1999 the PPC was still ahead, I seriously doubt how much that was a factor. Intel's rate of chip development didn't suddenly increase in 1999, it kept going at the same constant rate, while Motorola's stopped and flatlined for a year. Even if Motorola got back into pace with Intel, Intel got ahead during that year and stayed ahead.

Originally posted by Phinius
Windows XP is more than a slight upgrade from Windows 98 or ME. The fact remains that Windows created a lot less differentiation between the Mac and Microsoft's OS and so therefore there was much less need for people to buy a Mac over the many PC alternatives.

I would say there's more differentiation between XP and Mac OS X than there was between System 7.5 and Windows 95. The gap between Mac OS 8 and Windows 98 was even wider, as was the gap between Mac OS 9 and Windows ME. Considering that Apple is coming out with a significant upgrade to Mac OS X every year while Microsoft has upgraded every three years (with some updates, notoriously Windows 98, adding very little at all), the differentiation will only grow over time.

Originally posted by Phinius
So its not now poor management when Apple's sales and market share have fallen for three straight years in a row?

Apple is currently a profitable and innovative company that was able to overcome their performance disadvantage (courtesy of Motorola) and will now begin to increase sales as the performance gap closes.

Originally posted by Phinius
And are you still willing to blame almost the entire Mac sales decline since the G4 on Motorola? Even though Apple's Fred Anderson states he does not believe that the G5 Power Mac will every obtain the peak sales level of the G4 Power Mac?

Why do you bring this up when you've already been refuted?

Originally posted by Phinius
If Apple did work with IBM for three years getting a PowerPC version of the Power4 as Jon Rubenstein states, then Motorola almost certainly got wind of it somehow.

What makes you think so? Three years ago was when Motorola left the G4 untouched for a year. It makes sense that Apple would at that point start to line up a Plan B. But how would Motorola find out when all of the rumors sites didn't? Apple can keep a secret, when they must. I am sure this was one of those secrets.

Originally posted by Phinius
The 970 limited Motorola to being second banana to what IBM could offer Apple. So with smaller sales would come less financial incentive to spend on making processor improvements to that market and that is just what Motorola did.

So you admit that Motorola can't develop as well as IBM? Phinius, you have made the most confusing arguments. IBM can't keep up with Intel, because no one can keep up with Intel, and Motorola can't keep up with IBM, but Motorola gained on Intel during the time period that Intel passed Motorola?

Originally posted by Phinius
It would make much more sense for Motorola to spend money updating the G4 processor with new technologies that could be transfered throughout the entire PowerPC line. And that's just what they have been doing, instead of sinking money into a project that would have limited use outside of Apple's market.

If that's what they were doing then they aren't a very good chip maker, because they haven't done a good job.

Phinius

Aug 4, 2003, 02:25 AM

Originally posted by Phil Of Mac

Additionally, Apple may not revamp the architecture of the consumer products, i.e. no price rise.

If Apple is going to use the G5 in iMacs and eMacs, then there would have to be new motherboards made and other design changes as well. It's also quite likely that IBM would want to charge a overall higher price for the 970 than what Motorola sells the G4 for. Why is that? Because the 970 has a much higher performance level than the G4, therefore higher prices can be demanded.

Time, yes. But eventually, Apple probably wants to move to G5's across the board.

Intel is still making Pentium IIIs even though its been years since the Pentium 4 first went into production. Although Banias pretty much signals the end of production for the Pentium III. In the same vein its likely that Apple will continue to use the G4 and perhaps the G3 for quite some time to come. The G5 does not necessarily replace the G4 or G3, it could simply enable Apple to move into markets that neither the G3 or G4 can reach. In other words it allows expanded opportunties for Apple.

Low-end Xserve perhaps? Well, considering that the 970 is being used in IBM blade servers, I think it's also a good choice for the Xserve.

Low end servers is probably a major reason why Motorola has given the green light to a dual-core G4.

Intel may have more money, but IBM has a better architecture to develop. The PPC has kept pace with x86 and even surpassed it before, despite Apple's small market share and near-death experience.

Apple has never had as small a market share as they do now. Apple has fallen out of the top five personal computer computers in quarterly sales surveys, except in the U.S. consumer market.

I'm not saying Apple doesn't have to compete on price at all. I am saying that Apple has a quality advantage over the competition that justifies the slightly higher prices in place now.

The average price of desktop PC sales is about $749. Apple's average desktop price is over $2000. I'd hardly call that a slightly higher price for Macs. Price is a major contributing factor in desktop Mac sales falling steadily over the last couple of years. It's also a major reason why G5 Power Macs will probably never reach the peak sales level that the G4 Power Mac obtained. A G5 Power Mac is simply way more money than what the vast majority of personal computer purchasers are willing to pay now.

Intel's rate of chip development didn't suddenly increase in 1999, it kept going at the same constant rate,

Intel shortened the amount of time for smaller process technology introductions from three years to two. That had a major financial impact on competitors. Intel's rate of development did not stay constant, it picked up as revenue increased. Whereas with Apple's chip suppliers, Apple's market share and therefore revenue compared to competitors declined over many years.

while Motorola's stopped and flatlined for a year. Even if Motorola got back into pace with Intel, Intel got ahead during that year and stayed ahead.

Intel also got well ahead of AMD when the Pentium 4 went into production and they have stayed well ahead of AMD until the Opteron. In fact AMD has lost buckets of money trying to keep up with Intel's falling prices and increased chip performance in the last two years.

I would say there's more differentiation between XP and Mac OS X than there was between System 7.5 and Windows 95. The gap between Mac OS 8 and Windows 98 was even wider, as was the gap between Mac OS 9 and Windows ME. Considering that Apple is coming out with a significant upgrade to Mac OS X every year while Microsoft has upgraded every three years (with some updates, notoriously Windows 98, adding very little at all), the differentiation will only grow over time.

You seem to be almost completely blind to the market forces at work here. Apple has simply been outmanuevered by Intel and Microsoft. This is something that has happened over a period of many years. Even Steve Jobs admitted that Microsoft had won before his return to Apple. He recommended that Apple milk the Mac for all its worth and then go on to something else.

Apple is currently a profitable and innovative company that was able to overcome their performance disadvantage (courtesy of Motorola) and will now begin to increase sales as the performance gap closes.

Apple hardly had much of a performance advantage, if at all, with the old OS compared to what Microsoft was offering with Windows. Apple had gotten behind in OS technology and that is why Next was purchased at a high sum of some six hundred million dollars. Microsoft had moved ahead of Apple with more modern features and Apple needed to catch up.

The reason Apple's sales and market share declined over the years was a combination of factors. Apple's own mistakes and miscues were a major part of that. Steve Jobs even admitted that what happened under John Sculley was a major contributing factor to Microsoft becoming the standard in personal computers. Jobs stated that he believed that Apple should have gone for low prices sold to a wide customer base. Instead he stated Apple went for selling Macs at prices that were far greater than what he thought they should sell for. In his view Apple went for short term profits, instead of market share. Because of this and other mistakes Apple is in a market share position that is extremely difficult to grow out of.

Why do you bring this up when you've already been refuted?

Don't let the facts stand in the way of your personal beliefs and prejudices. You have a very narrow focus of blaming most if not all of Apple's ills on one supplier. When in fact Apple made a lot of mistakes over the years. Which when compounded has led them into the market share that they have now. One supplier could not have been the sole reason for Apple's current situation.

What makes you think so? Three years ago was when Motorola left the G4 untouched for a year. It makes sense that Apple would at that point start to line up a Plan B.

When Apple had their asses kicked by the Pentium 4 is when Apple knew that they had to step up the pace with processors. Using a processor design that was geared mainly to Apple's small market was not going to keep up with the pace of development from Intel. Motorola coming out with a G5 processor would not have been enough to combat Intel. Apple needed a much more ambititious design.

But how would Motorola find out when all of the rumors sites didn't? Apple can keep a secret, when they must. I am sure this was one of those secrets.

It's more than likely that Motorola found out from Apple. Motorola and Apple work very closely together, as does Apple and IBM. There is a lot of collaborative planning that goes into future products, it isn't a situation of Motorola coming up with a product and then seeing if Apple would be interested. Apple has to show a commitment to a product design before Motorola would invest millions in developing it. Apple probably would not commit to purchasing a Motorola G5 processor in the volumes necessary to make the venture worth investing in for Motorola.

So you admit that Motorola can't develop as well as IBM?

There literally is much less revenue to be made for Motorola making a high performance chip for Apple than what IBM can make off of one. That's due to IBM using the core design in highend computers and also their interest in using using the chip themselves in low cost servers and workstations. That in turn increases chip volume and revenue tremendously. What markets are open for Motorola to sell a desktop processor in? Who the heck would be interested in buying it besides Apple?

Phinius, you have made the most confusing arguments. IBM can't keep up with Intel, because no one can keep up with Intel, and Motorola can't keep up with IBM, but Motorola gained on Intel during the time period that Intel passed Motorola?

You jumping to conclusions that simply cannot be made from the information that I have presented. You seem to be limited to a narrow few of all bad or all good. IBM has to be all good to you and Motorola is all bad. That's third grade thinking you are exhibiting. You seem to get totally confused when the person arguing does not stay down those narrow confines of all good or all bad.

If that's what they were doing then they aren't a very good chip maker, because they haven't done a good job.

You have this child like fantasy that Motorola should have been able to keep up with Intel even though Motorola was restricted to selling to a ever smaller percentage of the overall personal computer market. Apple's market share did not go down just in one, two or even three years. It fell in most years dating back to the early 80s. Motorola was being hopelessly outgunned by Intel's market share that has now reached about a fifty-to-one ratio over what Motorola is limited to with Apple's market share. If Motorola spent the money to meet or beat what Intel is offering, the return on investment would still be restricted to Apple's puny market share. Apple is simply not going to grow fast enough to recoup that investment.

If you believe that IBM will be able to get back their investment in the Power4 core and plant facility investments from sales to Apple you are sadly mistaken. The only way that IBM could possibly have a competitive PowerPC processor for Apple's market is to use a design that is intended for other markets where IBM can recoup more of the costs of development. Motorola simply does not have that opportunity.

jaedreth

Aug 4, 2003, 09:06 AM

IBM is selling its own servers with Power4, and it's own workstations running Linux (often SuSE). It can easily sell these workstations using the 970, and thus guarantee a return on their investment.

Jaedreth

vrapan

Aug 4, 2003, 09:37 AM

I would just like to comment on a couple of things:

Intel's superiority: Intel P4 is not a particularly godd processor. If you go back to its introduction a P4 at 1.5GHz could easily compare to a P3 at 1.2GHz and at some tests losing. The only advantage of P4 is the 3GHz. AMDs processors run in 1GHz less and still keep up if not surpass P4 performance. On the same point IBM does not have the huge profits from the production of chips but they have a large pool of cash and tons of other operations. They can subsidise their chip operations if they care about gaining traction and keep developing a superior product. Given P4s inefficiencies I wouldn't surprised if a Dual 2GHz is much faster than a 3GHz P4 in the same way that a 2GHz Athlon can keep up with a P4.

As about Apples market share Moto has played a part but I don't know that it is as large as many seem to believe. There are several factors that have prevented Apple from gaining market share. One thing is that Intel got a rival. AMD. After that prices plummeted and GHz skyrocketed. Moto couldn't do much about that. If Intel was the only serious player on the PC side PCs wouldn't be at 3.X GHz right now. On top of that XP came soon after closing the gap between the Mac OS and Windows. Apple was still marketing it self as an education and arts platform rather than an every day computer. On top of that games didn't and dont play well on a 1000$ Mac but do so extremely well on a 1000$ PC. Games is a HUGE PC seller. And I come to the last point. Apple has sub-1000$ computers BUT they cannot compete with sub-1000$ PCs. Especially in crucial markets like gaming and every day use. The G5 which seems to be the only serious competitor to a PC box is TOOOOO EXPENSIVE.

Many PC users upgradetheir computers to death. eMac (the sub 1000) and the cheap iMac the expensive but close to 1000$ are NON UPGRADABLE. You are spending 1000$ for something that you can change nothing on. The PC user buys a 600$ computer and knows that in a year from now he can spend another 400$ and make it a much better computer and again the year after that. If you buy a Mac and you fill the 60GB of HD what are you gonna do? Buy a new one with 80GB? or spend 2200$ (I include tax) on a 1.6GHz G5 that is actually upgradable? The second choice seems reasonable but guess what. The more reasonable to too many people is to buy a 600$ an UPGRADABLE PC today rather than an non upgradable ~1000$ iMac or eMac or a 2200$ PowerMac simple as that.

From my part I moved from PC notebooks to Mac notebooks last Feb when I bought my first 12" PB. Awesome machine awesome operating system. I am not going back to PCs for my laptops. 1800$ for a 12" quality notebook is a good deal, this price compares well to 12" PC notebooks. This is why PBs and iBooks sell better than PowerMacs and iMacs. However as long as decently fast desktop Macs cost more than double than a decently fast PC I will keep buying PCs and new PBs every other year.

Pricing is what has kept Apple back and pricing is what will be Macs doom. I have said it elsewhere. A stripped down 1.6GHz (60GB, lower spec graphics card, not DVDR) in a bundle with a 17" Apple for around 1500+tax gives many more reasons to switch than a 1300$ i mac or a 1000$ eMac. And gives the 1.6GHz a reason of existance. People are not hooked to GHz as they used to mainly because of AMD. Apple should give reasons to people to switch and offer that an enticing price not on Mac pricing standards. Mac fans are willign to pay the hefty premium for a premium computing experience but before you can hook a PC user to the Mac experience you have to entice them with PC like prices for PC like configurations.....

Phinius

Aug 4, 2003, 09:55 AM

Originally posted by jaedreth

IBM is selling its own servers with Power4, and it's own workstations running Linux (often SuSE). It can easily sell these workstations using the 970, and thus guarantee a return on their investment.

IBM seems to have big plans for the 9XX series of PowerPC processors. Some IBM executives have already stated that they believe the companies Linux sales will increase 10 or 20 fold with the help of these chips.

The 9XX chips will help IBM combat Intel's encroachment on their server sales. At this point in time it looks like the Power core based chips will outsell the Itanium type procesors for awahile yet. That will in turn help justify the R&D needed to keep abreast of Intel in server processor development.

Phinius

Aug 4, 2003, 10:18 AM

Originally posted by vrapan

Intel's superiority: Intel P4 is not a particularly godd processor.

Good or bad, the Pentium 4 is outperforming the competition both in sales and in SPECint benchmark numbers.

The only advantage of P4 is the 3GHz.

The P4 also has a fast bus. Having a high frequency has been a key part of keeping the Pentium chips performance above the competition.

Given P4s inefficiencies I wouldn't surprised if a Dual 2GHz is much faster than a 3GHz P4 in the same way that a 2GHz Athlon can keep up with a P4.

The Pentium processors are king of the hill in desktop performance in the last two years and it looks to remain that way at least through the end of the year. The inefficiencies of the Pentium design do not effect its ability to perform in the desktop market. In fact it also does very well in the notebook market, even with a high amount of power use. Maybe Apple should also make a large and heavy PowerBook line with the top G5 speeds. It seems that most Mac users on this website are mainly concerned about topend speed for notebooks and not not size, weight or battery life. Just give them a high MHz notebook with lots of fans inside, they'll be more than happy and the product will probably be a big hit.

Intel was the only serious player on the PC side PCs wouldn't be at 3.X GHz right now.

Intel has 82% of the Windows box market. That's a very serious market share. That dominance has enabled the company to invest heavily in new fabrication facilities and R&D, which in turn keeps them ahead of the competition.

And I come to the last point. Apple has sub-1000$ computers BUT they cannot compete with sub-1000$ PCs. [/b]

Apple has one big advantage that other personal computer manufacturers do not and that is they can also create much of the software and services for the Mac. The much higher profit margin potential from software and services could make it possible for Apple to lower the prices of Macs considerably in the future. Afterall Dell has about a 18% overall gross margin on sales and Microsoft has about 82% margins on overall sales. Apple is starting to get very aggressive new software and services taking up a bigger precentage of overall sales revenue. The release of Panther and Windows iMusic store in the fourth quarter could justify lowering prices on Macs considerably for the Christmas season.

vrapan

Aug 4, 2003, 11:12 AM

OK I just tried to make a couple of point not to punch any others points. Trying to take a phrase of someone and not even bother to understand the context in which it was said is not a good start to make a point. Especially when your arguments do not have a point. In any case these were my two points if you care to criticise them please make sure that you are trying to prove something. Dont do it for criticising shakes.

Intel has a small margin on performance and that comes only from the fact that they use a higher GHz processor. What I was trying to say I guess is that a 3GHz 970 will be a faster less power hungry processor and Moore himself said that P4 cannot go much further in GHz. He said that they cannot keep doubling the number of GHz every 18 months on that processor. So all I am saying is that P4 is relying in raw power rather than an ingenious design. 82 % is good but there is another 18%. There is pressure. Without pressure why would intel invest in 3GHz processors production? Even while having the fastest processors they are losing slowly market share on AMDs which have become from a nothing to a serious competitor.

I dont know about others but I dont want yet ANOTHER 10pound 30minutes lasting desktop replacement. For that I have a desktop.

Most copies of Panther will ship with the hardware and with about 20 million users Panther cannot be considered a serious money making product. As far as the iTMS you can go check but Apple is losing money on it still and it is not going to be profitable for several more months. When it will be profitable it will not be making serious money to subsidise cheaper hardware simply because most of the 99c go to the record companies rather tha to Apple.Especially given that in order for the Apple computers to become competitive price wise they need at least a 20% discount (I am talkign about the lower G5s and iMacs). iTMS is iPod seller. So although it is true that software is more important for Apple it seems to be so not because they make money from software as such but because it supports hardware sales. So Apple is not a Dell but it is closer to a Dell than to an MS and in any case it is a very unique software resembling a bit of MS(OS + essential software), Sony(computers + digital media software) and Dell (hardware producer).
And in any case all I was comparing is todays and yesterdays prices. If by christmas the 1.6GHz sells "as little as" 50% more than a comparatively specced PC I would be the first to buy it.

Their laptops are reasonably well placed to face competition but their desktop machines are WAY over priced this is my second point.

nextdayflight5

Aug 4, 2003, 03:34 PM

okay lets stop this crud. cant we just take it back old school

i miss these posts:

"i cant wait til them new powerbooks come in"

or

"i heard some little shop in alabamas gots a new 15" powerd book"

short-simple-no one has to think about then afterward

come on guys, lets take it back old school.:cool:

Cubeboy

Aug 4, 2003, 04:27 PM

Originally posted by vrapan
Intel has a small margin on performance and that comes only from the fact that they use a higher GHz processor. What I was trying to say I guess is that a 3GHz 970 will be a faster less power hungry processor and Moore himself said that P4 cannot go much further in GHz. He said that they cannot keep doubling the number of GHz every 18 months on that processor. So all I am saying is that P4 is relying in raw power rather than an ingenious design. 82 % is good but there is another 18%. There is pressure. Without pressure why would intel invest in 3GHz processors production? Even while having the fastest processors they are losing slowly market share on AMDs which have become from a nothing to a serious competitor.

Currently in the x86 market, the fastest Pentium 4 does tend to do significantly better than the fastest Athlon in most real world tests. I don't recall AMD ever gaining marketshare since the days when the Athlon was at it's prime. I have heard that they've been losing money and marketshare for quite a while now.

You assume that cpus can only increase clockspeed based on core design and architectural improvements. This is only partly true, the other method of increasing clockspeed is to move to a more advanced manufacturing process. Most current cpus (the Pentium 4, Athlon, G5, but not the G4) use .13 micron process, along with various enhancements like SOI. Using .13 micron process, the Pentium 4 is able to scale to 3+ GHz while the Athlon and G5 are both able to scale to 2+ GHz. It is highly unlikely that the G5 will be able to scale to 3 GHz without moving to a smaller micron process. When the Pentium 4 (or Pentium 5 I should say) moves to a smaller micron process (.09 micron process), it will be able to scale significantly higher (4-5 GHz is official although it will be closer to the latter).

I don't ever recall Gordon Moore ever saying anything about the Pentium 4 being unable to scale any further, it would be plainly wrong considering any chip will be able to scale to significantly higher clockspeeds as newer and more advanced manufacturing processes arrive. After all, the general rule is a 50% increase in clockspeed for every shrink in micron process.

If you want to talk about raw power, the Athlon, with it's triple FPU, triple ALUs, and triple decoders all backed up by a huge L1 cache is going to have alot more raw power than the Pentium 4 in many cases, however due to it's higher latency, smaller (72)instruction buffer, generally weaker memory subsystem, and the very nature of x86 code (low ILP, lots of load-stores), there are very few instances where it can surpass the Pentium 4 and in many instances it can barely keep up. I find the Pentium 4's architecture, with it's low latency caches, large (126) instruction buffer, excellent memory subsystem, and very good branch prediction, tends to be very efficient with all it's available "raw power".

I don't see where your going with the ingenious design thing, exactly what makes the Pentium 4 any less of a ingenious design than a G5 or Athlon? Clock speed, isn't raw power or brute force, it's just another aspect of performance. It simply isn't fair to call a processor uninnovative (or uningenious) just because it's method of achieving necessary performance is by reaching higher clockspeeds.

tychay

Aug 4, 2003, 08:38 PM

Originally posted by Phinius
Here it is from a 2001 news.com article (http://news.com.com/2100-1040-257059.html). AMD signed a ten year licensing deal with Intel.

Ahh, the ten-ual cross-licensing renewal, I had forgotten, thanks for pointing it out. The actual sum of the royalties is secret, but I'm willing to bet it is a pittance since all the game theory results show that AMD would sue if the pricing were predatory as you imply. Besides, nobody ever mentions it besides the C|net article--the whole thing reeks of Intel spin.

I don't see why AMD would want to reverse engineer a Intel processor since they are paying a license fee for the use of X86.

What does licensing the instruction set have anything to do with the technology underneath? BTW, the cross-licensing also gives them rights to the patents, but that doesn't hold water either. Intel and AMD need access to the design and manufacturing decisions that went into the chip, which isn't available in a patent application. In any case, I was alluding to companies such as this one (http://www.chipworks.com/), which definitely get paid by 3rd parties to do competitive analysis of competing platforms.

In this Arstechnica interview, Peter Sandon states Apple designed their own chipset for the 970 and IBM manufactures it. You will find the question and answer under the the 'Miscellany: bus ratios compilers, power management' heading.

I don't need to since I already posted in that same post that the ASIC was an Apple design under IBM manufacture. In my comment, which you've pointlessly taken out of context because your combative nature, I was alluding to the Common Hardware Reference Platform (CHRP) for PowerPC published by IBM. Apple used it as a base to do a design shortcut on all motherboards since the iMac.

AMD seems much more likely to produce a higher volume of processors for notebooks, servers and the desktop than IBM will since AMD makes chips for the large Windows market.

As of yet IBM does not makes processors that are used in notebooks, servers and the desktop. Although that could change if Apple introduces a 970 equipped PowerBook and Xserve.

I say a fact, and you talk about something completely different. I never said AMD isn't going to produce a higher volume of chips, I said that IBM is better positioned. Since AMD is running to Fishkill for manufacturing and design center work, your example just enforces IBM as the #2 player in CPUs.

IBM does make processors that are used in notebooks (G3 (http://www.apple.com/ibook/)), SMP servers (970 (http://www.osnews.com/story.php?news_id=2932) and Power4 (http://www-132.ibm.com/content/home/store_IBMPublicUSA/en_US/eServer/pSeries/mid_range/pSeries_midrange.html)) and desktop workstations (Power4 (http://www-1.ibm.com/servers/eserver/pseries/news/features/2003/annc_527.html) and 970 (http://www.apple.com/g5processor/)). Also, their entire Intel x86 server line (http://www.pc.ibm.com/us/eserver/xseries/) uses a PowerPC as a system controller chip.

You are correct to point out that AMD will have an offering in all three categories. I had forgotten about the Athlon XP-M (http://www.amd.com/us-en/Processors/ProductInformation/0,,30_118_756_807,00.html) because notebooks with it haven't come out.

Phinius

Aug 5, 2003, 01:21 AM

Originally posted by vrapan

Most copies of Panther will ship with the hardware and with about 20 million users Panther cannot be considered a serious money making product.

Apple now sells computers at a rate of 3 million or less per year. Your estimate of 20 million Mac users would be the equivilant of all the Macs sold in the last 8 years. It's very unlikely that people keep their computers on average that long.

If people keep their computers an average of three years, then that would be about 9 million in Mac sales. If only 2 million people buy the Panther upgrade in the year after its release, then thats close to 200 million dollars in gross revenue for Apple. It's not much revenue compared to what Mac computer sales bring in, but it's got a very high gross margin and so therefore a big potential money maker. It's tough for Apple to make money directly on computer sales since Dell is selling a much higher volume at low prices and a low gross margin. Apple can counter some of that with annual revenue from OSX upgrades, services such as .Mac and iTunes music store.

As far as the iTMS you can go check but Apple is losing money on it still and it is not going to be profitable for several more months. When it will be profitable it will not be making serious money to subsidise cheaper hardware simply because most of the 99c go to the record companies rather tha to Apple.

The iTunes music store has yet to break even according to Apple, but it has such low overhead that it is bound to be profitable when Apple releases the Windows version to a much bigger potential base of customers. The iTunes music store will not bring in enough revenue to drop Mac prices $100 across the board. However it does bring in a revenue stream that could help to lower prices on Macs.

Especially given that in order for the Apple computers to become competitive price wise they need at least a 20% discount (I am talkign about the lower G5s and iMacs).

Apple could get some of that needed 20% price reduction on computers through the increasing software and services revenue that continues to pour from Mac users long after the initial purchase of the computer. The surge in revenue from the introduction of Panther and Windows music store should enable Apple to decrease prices on consumer computers somewhat in the fourth quarter and yet still maintain gross margins.

Their laptops are reasonably well placed to face competition but their desktop machines are WAY over priced this is my second point.

Apple's competitiveness in notebook computers is under siege from HP's and Dell's heated battle in lowering prices and also from the introduction of Intel's Banias. This enables PC notebook manufacturers to offer notebooks that are comparable in size and weight to what Apple offers, yet at a much lower price.