<strong>G4's will be at 1.5GHz in September, and 2Ghz by the end of the 1st quarter of 2003. I think that will be the last iteration of G4 that Apple uses - so we're talking at the earliest Summer 2003 for G5.</strong><hr></blockquote>
I'm no engineer, but obviously you aren't either. If you talk to sources who understand the design and technologies currently being put into play by MOT, there is agreement that there is NO WAY the G4 design will reach 2 GHz. Perhaps 1.2 - 1.4 GHz, but not more without a significant redesign.
Of course if you can cite a logical reason why MOT would sink millions into redesinging an existing chip when they've already toiled on the G5, I'm all ears.
We may end up using G4's for another year yet, but that's NOT because the G4 will magically scale another 1000MHz in one year when it basically took three to go the first 1000.
It doesn't really matter what MOTO's timetable is or what IBM or anyone else are planning technology wise. The only way we will ever see a G5 is if his steveness decides to use it. I don't think that this will happen in a hurry. Jobs is currently obsessed with the consumer market and sees the G4 as ideal for it. It is also apple's way of holding back on any gamble during lean times for the computer industry as a whole.
I think pro users are far far down apple's list - something that the likes of Adobe, Macromedia and Quark have now realised (Alias must be weeping - all that R&D).
Bu the time we get a G5 the wintel world will be at 2GHz Itanium, DDR2 et al.
Every mac user I know (and I know hundreds - we've had our own poll over email) has resisited the urge to buy the new imac or the G4s as we are waiting for the G5. Even if it doesn't come out - the rumours are so big (fuelled in many ways by apple themselves) that they will never get the kind of sales they might have expected from the current crop of QS towers.
lets face it - its the G5, not the Mhz, that are the myth.
[quote]If the G5 will not be available until midyear, whatever that means(June/July). Will that be enough time for Apple to recieve the chips, do any final testing, and manufacture computers and announce it @ MWNY?
<hr></blockquote>
I think what this means is that the G5 won't be available in quantities for sale by Motorola. But Moto can still supply Apple with prototypes of the chips so that they can design Powermacs around them. Otherwise Moto would have new CPUs all ready for sale but Apple wouldn't buy them for 6-9 months because they have to design new mobos around them...not a very efficient design strategy.
Moto can supply Apple with plenty G5s for testing purposes..in fact they could probably give Apple enough 2.4 GHz G5s for testing, but these are chips from an inefficient fab. By midyear, Moto will be pumping out G5 chips efficiently and in quantity, so the price will be reasonable.
If the reports from Architosh are to be believed, then we'll be seeing Powermac G5s soon, as the field testing phase is coming to a close.
<strong>Jobs is currently obsessed with the consumer market and sees the G4 as ideal for it.
I think pro users are far far down apple's list - something that the likes of Adobe, Macromedia and Quark have now realised (Alias must be weeping - all that R&D).</strong><hr></blockquote>
Jobs currently has the means to do something about the consumer market, which frankly was in a worse state than the PowerMacs are.
He isn't the sort of person who is open about unannounced developments so it is impossible to say how he feels towards the pro market.
From what I have seen though Apple is very aware and interested in the pro market but simply not ready to deliver.
It would be a safe assumption to assume the execs at Alias knows something you don't.
<strong>intel is not struggling to push the P4 at all. The northwoods should easily clock beyond 3 GHz.</strong><hr></blockquote>
The 2 and 2.2GHz P4 is based on the northwood core, is it not? The fact that the best improvement in clock speed intel could get from shifting from .18 micron with Williamette to .13 micron with Northwood was a mere 10% should tell you everything about how scalable Northwood really is.
Even Intel internal roadmaps show Northwood only hitting 2.53GHz by the end of 2002 - and that's with Northwood B, rather than the current version.
And by the end of the year, G4's (yes, G4, not G5) will be at 1.5GHz (conservative estimate) and possible 2GHz (less likely, but possible). That gap is closing.
If the pro market is so far down on Apple's list than why did they just buy Nothing Real. This is expensive high end software. I'm sure parts of it can be incorporated into consumer level software but this has "Major Final Cut Pro Upadate" written all over it if not something in addition to FCP.
Another thing. Don't go looking too hard on Moto's web site for clues. They probably wouldn't publish incorrect information but there's no rule that says they have to tell you everything that they plan to realease.
The 2 and 2.2GHz P4 is based on the northwood core, is it not? The fact that the best improvement in clock speed intel could get from shifting from .18 micron with Williamette to .13 micron with Northwood was a mere 10% should tell you everything about how scalable Northwood really is.
</strong><hr></blockquote>
They are getting a 10% increase now, on a brand new 130nm process.
That doesn't mean a lot in regard to future scaling.
The Original Athlon began shupping at 500MHz IIRC, and scaled up to 1GHz during the following years without any design change. Still, they obviously couldn't get it to clock that high right from the start.
The same is most likely true for the P4 now.
And besides, why would Intel make huge clock speed advancements at once, when it's much more economical to increase it step-wise?
Also it doesn't matter how much Northwood CAN scale, only how much Intel will LET it scale. Intel will only bump up the processors to be ahead of the competition. If by the end of the year AMD is only at 1.8GHz, then why would Intel be at 3.0GHz? They would have to update almost 100MHz every month. Intel will milk every step for as long as they can and continue to be competitive. 2.5GHz by the end of the year. I think Intel is a schizophrenic company. On one hand they have a wildly high GHz but inefficient processor called the P4 and put emphesis on more MHz= power. But they want to replace it with a low MHz design (and McKinley will scale less than the modern G4. It has a huge die size even though its on 130nm process and only 7, yes seven, pipeline stages. Very un-intel-like.). All the power to them.
The 2 and 2.2GHz P4 is based on the northwood core, is it not? The fact that the best improvement in clock speed intel could get from shifting from .18 micron with Williamette to .13 micron with Northwood was a mere 10% should tell you everything about how scalable Northwood really is.
Even Intel internal roadmaps show Northwood only hitting 2.53GHz by the end of 2002 - and that's with Northwood B, rather than the current version.
And by the end of the year, G4's (yes, G4, not G5) will be at 1.5GHz (conservative estimate) and possible 2GHz (less likely, but possible). That gap is closing.</strong><hr></blockquote>
The PIV is plenty scalable (with standard cooling overclocking to 2.6GHz isn't too hard of the Northwood chips) and will be well past 2.5GHz come 2003. If not they stand to lose substantial marketshare to AMD who will break the 3GHz mark by early 2003.
Take into account they also make more efficient chips and Intel would truly be a long way behind. That simply isn't going to happen.
They are getting a 10% increase now, on a brand new 130nm process.
</strong>
And everyone was laying into Motorola for getting a 15% instant increase moving to SOI on the 7455. 15% sure looks better than 10% in a single jump, does it not?
<strong>
That doesn't mean a lot in regard to future scaling.
</strong>
Agreed. So why are people assuming that Northwood will "easily" scale to 3GHz?
<strong>
The same is most likely true for the P4 now.
</strong>
There's no reason to suppose that P4 can/will follow the same path as Athlon. The two are very different chips, made by very different companies, with very different strengths and weaknesses.
<strong>
And besides, why would Intel make huge clock speed advancements at once, when it's much more economical to increase it step-wise?
</strong>
Because unlike in the old days, it now has a formidable competitor - Athlon - on its tail. Make no mistake, in order to keep ahead of AMD, Intel can and will scale the P4 as fast as it can.
The PIV is plenty scalable (with standard cooling overclocking to 2.6GHz isn't too hard of the Northwood chips) and will be well past 2.5GHz come 2003. If not they stand to lose substantial marketshare to AMD who will break the 3GHz mark by early 2003.
They are getting a 10% increase now, on a brand new 130nm process.
That doesn't mean a lot in regard to future scaling.
The Original Athlon began shupping at 500MHz IIRC, and scaled up to 1GHz during the following years without any design change. Still, they obviously couldn't get it to clock that high right from the start.
The same is most likely true for the P4 now.
And besides, why would Intel make huge clock speed advancements at once, when it's much more economical to increase it step-wise?
Bye,
RazzFazz</strong><hr></blockquote>
I recall that the K7 was 500-700 MHz and the K75 550-1000 MHz, although I'm not sure as to what the differences are.
<strong>Agreed. So why are people assuming that Northwood will "easily" scale to 3GHz?
</strong><hr></blockquote>
Because increases like this during the life span of a processor revision have been very common in the x86 world.
[quote]<strong>
There's no reason to suppose that P4 can/will follow the same path as Athlon. The two are very different chips, made by very different companies, with very different strengths and weaknesses.
</strong><hr></blockquote>
The same has also been true for the P3, the P2 before that, and the 250nm P1.
Basically, with the exception of the PPro, any recent x86 processor has almost doubled it's clock rate during its life cycle.
[quote]<strong>
Because unlike in the old days, it now has a formidable competitor - Athlon - on its tail. Make no mistake, in order to keep ahead of AMD, Intel can and will scale the P4 as fast as it can.
</strong><hr></blockquote>
No. They will scale it just enough to stay ahead of AMD. Anything more would be both unnecessary and economically unreasonable.
I recall that the K7 was 500-700 MHz and the K75 550-1000 MHz, although I'm not sure as to what the differences are.</strong><hr></blockquote>
Wasn't the K75 the so-called Thunderbird, i.e. the chip w/ on-die cache? I seem to remember there were, at least at one point, GHz versions w/ off-die cache (clocked at 1/3 core freq. IIRC). I might be wrong, though.
EDIT: Just checked it. The first (march 2000) chip to break the GHz barrier was an Athlon w/ off-die L2 cache, i.e. the original one. The Thunderbird core was only introduced in april 2000.
Well, if I'm to believe the article on MacCentral than I will be losing my G5 bet. However I think this is just a way to get people to buy G4's and stop waiting for the G5.
I've been holding off for over a year now, and will continue to do so (I don't NEED a faster computer, I just choose to wait for my next purchase to be a G5).
Apple is like the CIA, full of mis-information about products and releases, I think we will see the G5 before years end.
Well, in the MacCentral article it said " for the coming months ". What the hell does that mean? MWNY is 5 and a half months away. We are still very early in the year. Coming months to me means 3 to 4 months. No new hardware for spring. That doesn't say anything about the middle of summer. Plus I can imagine they want to have a set hardware plan, so they can finish up OS X without the unpredictability of having to support new options.
Since it doesn't seem to have been mentioned in this thread yet, I'll do it: The Register has just posted an article about G4 enhancements that are coming soon. The 7460, 7470, and 7500. These variously include a faster MPX, higher clock rates, longer pipelines, and RapidIO. The chips described could be substantial improvements over the current G4, and for that last one it wouldn't bother me if they called it a G5. This article seems in line with the recent MacCentral article, and it could explain the substantially faster test boxes that have been rumoured.
Comments
<strong>G4's will be at 1.5GHz in September, and 2Ghz by the end of the 1st quarter of 2003. I think that will be the last iteration of G4 that Apple uses - so we're talking at the earliest Summer 2003 for G5.</strong><hr></blockquote>
I'm no engineer, but obviously you aren't either. If you talk to sources who understand the design and technologies currently being put into play by MOT, there is agreement that there is NO WAY the G4 design will reach 2 GHz. Perhaps 1.2 - 1.4 GHz, but not more without a significant redesign.
Of course if you can cite a logical reason why MOT would sink millions into redesinging an existing chip when they've already toiled on the G5, I'm all ears.
We may end up using G4's for another year yet, but that's NOT because the G4 will magically scale another 1000MHz in one year when it basically took three to go the first 1000.
I think pro users are far far down apple's list - something that the likes of Adobe, Macromedia and Quark have now realised (Alias must be weeping - all that R&D).
Bu the time we get a G5 the wintel world will be at 2GHz Itanium, DDR2 et al.
Every mac user I know (and I know hundreds - we've had our own poll over email) has resisited the urge to buy the new imac or the G4s as we are waiting for the G5. Even if it doesn't come out - the rumours are so big (fuelled in many ways by apple themselves) that they will never get the kind of sales they might have expected from the current crop of QS towers.
lets face it - its the G5, not the Mhz, that are the myth.
<hr></blockquote>
I think what this means is that the G5 won't be available in quantities for sale by Motorola. But Moto can still supply Apple with prototypes of the chips so that they can design Powermacs around them. Otherwise Moto would have new CPUs all ready for sale but Apple wouldn't buy them for 6-9 months because they have to design new mobos around them...not a very efficient design strategy.
Moto can supply Apple with plenty G5s for testing purposes..in fact they could probably give Apple enough 2.4 GHz G5s for testing, but these are chips from an inefficient fab. By midyear, Moto will be pumping out G5 chips efficiently and in quantity, so the price will be reasonable.
If the reports from Architosh are to be believed, then we'll be seeing Powermac G5s soon, as the field testing phase is coming to a close.
<a href="http://www.architosh.com/news/2002-02/2002b-0201-futurg5.phtml"; target="_blank">http://www.architosh.com/news/2002-02/2002b-0201-futurg5.phtml</a>;
<strong>Jobs is currently obsessed with the consumer market and sees the G4 as ideal for it.
I think pro users are far far down apple's list - something that the likes of Adobe, Macromedia and Quark have now realised (Alias must be weeping - all that R&D).</strong><hr></blockquote>
Jobs currently has the means to do something about the consumer market, which frankly was in a worse state than the PowerMacs are.
He isn't the sort of person who is open about unannounced developments so it is impossible to say how he feels towards the pro market.
From what I have seen though Apple is very aware and interested in the pro market but simply not ready to deliver.
It would be a safe assumption to assume the execs at Alias knows something you don't.
(apple's main source of revenue rates down the bottom?)
Now this is another entry on my list of moronic things uttered on rumor forums
(i can't remember all of them ad verbatim, but if you please
1) "There is no G5", often linked with "It's all up to Motorola" (or similar)
(sure, Apple's just sitting on their hands, doing nothing wrt a next generation processor.)
2) "There's no G5 soon because Apple doesn't listen."
(Apple isn't your spouse, in case you haven't noticed yet; if they can
release something to maximise revenue, believe me, they will. Corporations
are not impetuous humans [all the moto-killed-ppc-dev-out-of-spite theory
fans please take excruciatingly detailed note]; rather, they work to
please their owners, out of necessity [please, nothing about `Apple's
stock price tanking'; Apple's market cap is irrelevant).
3) The third of the existing doesn't take the form of a cutnpaste quote,
but rather the general sentiment that a desktop G5 isn't imminent because
Motorola has already announced their first Book E compliant processor will
be released in the second half of the year (usually followed by
something about `ruling out a G5 release until SF03').
Now there's a number of ways around this: pure BS (out of contractual
secrecy with Apple), semantic escapology such as the G5 Apple uses not being
from `Motorola' but instead from `Apple'; an `Apple G5', for example.
[the reasons for a desktop G5 development being conducted by Apple rather
than Motorola is obvious and has been discussed in depth already.]
>Bu the time we get a G5 the wintel world will be at 2GHz Itanium, DDR2 et al.
The Itanium isn't going anywhere near 2 Ghz. The maximum clockspeed of
McKinley is 1 GHz (taken from Kins Collins' CPUs.html page -- can't find
it right now), and it's possible that the even the third in line won't be
near 2 Ghz max, I suppose.
>lets face it - its the G5, not the Mhz, that are the myth.[/QB][/QUOTE]
Does this qualify as 1)? I believe it does, as (given that `G5' is the
name given to Apple's next-generation processor) it boils down to "the
G5... [is] the myth".
(Yes, you're correct, Apple's not doing anything at all with regard to
future processor development. They'd rather forsake the largest chunk of
their revenue and a large part of their installed base. Indeed, Apple
would forgo such revenues... *sarcasm off*)
All flames welcome and will be replied to in turn. Constructive opinion
also welcome.
/me steps into flame-proof suit
<strong>intel is not struggling to push the P4 at all. The northwoods should easily clock beyond 3 GHz.</strong><hr></blockquote>
The 2 and 2.2GHz P4 is based on the northwood core, is it not? The fact that the best improvement in clock speed intel could get from shifting from .18 micron with Williamette to .13 micron with Northwood was a mere 10% should tell you everything about how scalable Northwood really is.
Even Intel internal roadmaps show Northwood only hitting 2.53GHz by the end of 2002 - and that's with Northwood B, rather than the current version.
And by the end of the year, G4's (yes, G4, not G5) will be at 1.5GHz (conservative estimate) and possible 2GHz (less likely, but possible). That gap is closing.
[ 02-07-2002: Message edited by: lineker ]</p>
Another thing. Don't go looking too hard on Moto's web site for clues. They probably wouldn't publish incorrect information but there's no rule that says they have to tell you everything that they plan to realease.
<strong>
The 2 and 2.2GHz P4 is based on the northwood core, is it not? The fact that the best improvement in clock speed intel could get from shifting from .18 micron with Williamette to .13 micron with Northwood was a mere 10% should tell you everything about how scalable Northwood really is.
</strong><hr></blockquote>
They are getting a 10% increase now, on a brand new 130nm process.
That doesn't mean a lot in regard to future scaling.
The Original Athlon began shupping at 500MHz IIRC, and scaled up to 1GHz during the following years without any design change. Still, they obviously couldn't get it to clock that high right from the start.
The same is most likely true for the P4 now.
And besides, why would Intel make huge clock speed advancements at once, when it's much more economical to increase it step-wise?
Bye,
RazzFazz
<strong>
The 2 and 2.2GHz P4 is based on the northwood core, is it not? The fact that the best improvement in clock speed intel could get from shifting from .18 micron with Williamette to .13 micron with Northwood was a mere 10% should tell you everything about how scalable Northwood really is.
Even Intel internal roadmaps show Northwood only hitting 2.53GHz by the end of 2002 - and that's with Northwood B, rather than the current version.
And by the end of the year, G4's (yes, G4, not G5) will be at 1.5GHz (conservative estimate) and possible 2GHz (less likely, but possible). That gap is closing.</strong><hr></blockquote>
The PIV is plenty scalable (with standard cooling overclocking to 2.6GHz isn't too hard of the Northwood chips) and will be well past 2.5GHz come 2003. If not they stand to lose substantial marketshare to AMD who will break the 3GHz mark by early 2003.
Take into account they also make more efficient chips and Intel would truly be a long way behind. That simply isn't going to happen.
Edit: I need sleep. Lots of errors to correct.
[ 02-07-2002: Message edited by: Telomar ]</p>
<strong>
They are getting a 10% increase now, on a brand new 130nm process.
</strong>
And everyone was laying into Motorola for getting a 15% instant increase moving to SOI on the 7455. 15% sure looks better than 10% in a single jump, does it not?
<strong>
That doesn't mean a lot in regard to future scaling.
</strong>
Agreed. So why are people assuming that Northwood will "easily" scale to 3GHz?
<strong>
The same is most likely true for the P4 now.
</strong>
There's no reason to suppose that P4 can/will follow the same path as Athlon. The two are very different chips, made by very different companies, with very different strengths and weaknesses.
<strong>
And besides, why would Intel make huge clock speed advancements at once, when it's much more economical to increase it step-wise?
</strong>
Because unlike in the old days, it now has a formidable competitor - Athlon - on its tail. Make no mistake, in order to keep ahead of AMD, Intel can and will scale the P4 as fast as it can.
<hr></blockquote>
<strong>
The PIV is plenty scalable (with standard cooling overclocking to 2.6GHz isn't too hard of the Northwood chips) and will be well past 2.5GHz come 2003. If not they stand to lose substantial marketshare to AMD who will break the 3GHz mark by early 2003.
Edit: I need sleep. Lots of errors to correct.
[ 02-07-2002: Message edited by: Telomar ]</strong><hr></blockquote>
Can you give any sources for your claim that either Intel or AMD will be beyond 3GHz by 2003?
And have a good sleep!
FORGET IT
First of all 2Ghz would already require a major overhaul, most certainly 0.13 micron and deeper pipelines.
Then: Have you ever seen a PPC chip, or any other chip for that matter, scale 1000MHz in less than a year?
FORGET IT.
G-News
<strong>
They are getting a 10% increase now, on a brand new 130nm process.
That doesn't mean a lot in regard to future scaling.
The Original Athlon began shupping at 500MHz IIRC, and scaled up to 1GHz during the following years without any design change. Still, they obviously couldn't get it to clock that high right from the start.
The same is most likely true for the P4 now.
And besides, why would Intel make huge clock speed advancements at once, when it's much more economical to increase it step-wise?
Bye,
RazzFazz</strong><hr></blockquote>
I recall that the K7 was 500-700 MHz and the K75 550-1000 MHz, although I'm not sure as to what the differences are.
<strong>Agreed. So why are people assuming that Northwood will "easily" scale to 3GHz?
</strong><hr></blockquote>
Because increases like this during the life span of a processor revision have been very common in the x86 world.
[quote]<strong>
There's no reason to suppose that P4 can/will follow the same path as Athlon. The two are very different chips, made by very different companies, with very different strengths and weaknesses.
</strong><hr></blockquote>
The same has also been true for the P3, the P2 before that, and the 250nm P1.
Basically, with the exception of the PPro, any recent x86 processor has almost doubled it's clock rate during its life cycle.
[quote]<strong>
Because unlike in the old days, it now has a formidable competitor - Athlon - on its tail. Make no mistake, in order to keep ahead of AMD, Intel can and will scale the P4 as fast as it can.
</strong><hr></blockquote>
No. They will scale it just enough to stay ahead of AMD. Anything more would be both unnecessary and economically unreasonable.
Bye,
RazzFazz
<strong>
I recall that the K7 was 500-700 MHz and the K75 550-1000 MHz, although I'm not sure as to what the differences are.</strong><hr></blockquote>
Wasn't the K75 the so-called Thunderbird, i.e. the chip w/ on-die cache? I seem to remember there were, at least at one point, GHz versions w/ off-die cache (clocked at 1/3 core freq. IIRC). I might be wrong, though.
EDIT: Just checked it. The first (march 2000) chip to break the GHz barrier was an Athlon w/ off-die L2 cache, i.e. the original one. The Thunderbird core was only introduced in april 2000.
Bye,
RazzFazz
[ 02-10-2002: Message edited by: RazzFazz ]</p>
I've been holding off for over a year now, and will continue to do so (I don't NEED a faster computer, I just choose to wait for my next purchase to be a G5).
Apple is like the CIA, full of mis-information about products and releases, I think we will see the G5 before years end.
I'd buy one.
<a href="http://www.theregister.co.uk/content/3/24018.html"; target="_blank">http://www.theregister.co.uk/content/3/24018.html</a>;