Apple's future speed bumps determined
It now seems somewhat predictable in how Apple upgrades it's product lines.
It's usually a 200 mhz speed bump or so at the most. None of this 600mhz jump as some have dreamed about.
Not that it's not technically possible, but from a business and marketing point of view, it's best to take the progressive route. They not only milk customers for more money along the way but don't piss off existing owners by introducing something better so soon.
So all those clamoring for a 2Ghz PowerMac, you've got a long way to go.
It's usually a 200 mhz speed bump or so at the most. None of this 600mhz jump as some have dreamed about.
Not that it's not technically possible, but from a business and marketing point of view, it's best to take the progressive route. They not only milk customers for more money along the way but don't piss off existing owners by introducing something better so soon.
So all those clamoring for a 2Ghz PowerMac, you've got a long way to go.
Comments
- The move from 4 stages to 7
So basically unless we're getting a G4++ I didn't even really think we'd get 1.6hz. I was thinking 1.4 VERY tops.SOI
Process Shrink?
Somehow I think that Apple will be using the G4 for sometime. That doesn't rule out a G5 within the next 12 months but the G4 still has enought life IMO
<strong>If you can show Apple a 2Ghz G4 I'm sure they'll be happy to build a super PowerMac.</strong><hr></blockquote>
I agree, I'm sure that Jobs would love to make another debute like when he announced the G4 towers, as long as he had the processors to back it up.
<strong>It now seems somewhat predictable in how Apple upgrades it's product lines.
It's usually a 200 mhz speed bump or so at the most. None of this 600mhz jump as some have dreamed about.</strong><hr></blockquote>That's only NOW becoming clear to you? Then you've been reading AI Future Hardware too much.
<img src="graemlins/lol.gif" border="0" alt="[Laughing]" />
<strong>A 2Ghz PM would require an architecture change. It's obvious that recent advances in G4 Megahertz have come from
- The move from 4 stages to 7
So basically unless we're getting a G4++ I didn't even really think we'd get 1.6hz. I was thinking 1.4 VERY tops.SOI
Process Shrink?
Somehow I think that Apple will be using the G4 for sometime. That doesn't rule out a G5 within the next 12 months but the G4 still has enought life IMO</strong><hr></blockquote>
And an interesting thing is that by sticking with the G4 design, but continuing to evolve fabrication techniques and process size alongside innovative engineering, you continue to leverage a very power-efficient chip and all of the amortized investment.
That's not to say I wouldn't want to see something a little more revolutionary, but if the rumours are true that Intel/MS are really having difficulty scaling the P4 (and presumably the Xeon MP/DP variants) to the vaunted 2.8GHz mark, it makes a lot of sense to keep the powder dry for the moment and ultimately release a more well-specced/developed solution.
1000 --> 1250 +25%
1250 --> 1600 +28% (MSWF 2003)
800 --> 1600 +100% in 18 months.
Moore's law wins again!
<strong>It now seems somewhat predictable in how Apple upgrades it's product lines.
It's usually a 200 mhz speed bump or so at the most. None of this 600mhz jump as some have dreamed about.
.</strong><hr></blockquote>
don't foreget that they also bump them down by a 50 mhz a month later
So powermac got speed bumped by two hundred mhz three times. that estabilishes some sort of pattern? we're talking about apple and motorola there's no telling what to predict
<strong>800 --> 1000 +25%
1000 --> 1250 +25%
1250 --> 1600 +28% (MSWF 2003)
800 --> 1600 +100% in 18 months.
Moore's law wins again!</strong><hr></blockquote>
Moore's Law says nothing directly about clockspeed. It only says that the numbers of transistors on a chip will double every 18 months.
<strong>That's not to say I wouldn't want to see something a little more revolutionary, but if the rumours are true that Intel/MS are really having difficulty scaling the P4 (and presumably the Xeon MP/DP variants) to the vaunted 2.8GHz mark, it makes a lot of sense to keep the powder dry for the moment and ultimately release a more well-specced/developed solution.</strong><hr></blockquote>
Not sure where that rumour came from but Intel is having no troubles increasing the PIV's speed yet. In fact they are ahead of their original schedules.
<strong>
Not sure where that rumour came from but Intel is having no troubles increasing the PIV's speed yet. In fact they are ahead of their original schedules.</strong><hr></blockquote>
Someone probably mixed up Intel and AMD here.
The latter do seem to have some difficulties ramping up their clock speeds - at least moving to 130nm has only brought them a 66MHz increase (compared to 1733 MHz on 180nm!) in clock speed up to now.
Bye,
RazzFazz
<strong>800 --> 1000 +25%
1000 --> 1250 +25%
1250 --> 1600 +28% (MSWF 2003)
800 --> 1600 +100% in 18 months.
Moore's law wins again!</strong><hr></blockquote>
Give your math co-processor a whack man. 25+25+28=78% increase, not 100%
Also the Moore's Law thing was already pointedd out. Double the transistors not MHz speed, although that has usually been about on target too
Not that it's not technically possible, but from a business and marketing point of view, it's best to take the progressive route. They not only milk customers for more money along the way but don't piss off existing owners by introducing something better so soon.
So all those clamoring for a 2Ghz PowerMac, you've got a long way to go.<hr></blockquote>
I believe that Apple would try to introduce computers with the fastest chips they can stuff in them. This is from a business perspective.
How many people have said in the past, "Well, the new PM are great, but not compelling enough for me to upgrade"?
If Apple introduced 2GHz chips yesterday, the cybertumbleweeds rolling through these forums would be because everyone would be at the Apple store trying to place an order.
Give your math co-processor a whack man. 25+25+28=78% increase, not 100%
<hr></blockquote>
Funny, I could have sworn that two eights are sixteen....
<strong>
Give your math co-processor a whack man. 25+25+28=78% increase, not 100%
</strong><hr></blockquote>
Actually, he's completely correct. You can't just add up the percentages, as the base value is increased with each bump, like that:
1.0 GHz * 125% = 1.25 GHz
1.25 GHz * 125% = 1.5625 GHz
1.5625 GHz * 128% = 2.0 GHz
Bumps: 25%, 25%, 28%
Total increase: 100%
Bye,
RazzFazz
Well, well...Imagine in few years...we'll get to 5Ghz. Do you think then it will be a 200Mhz increase?
What about when we were at 100Mhz?
Your logic doesn't work.
MOBO simply can't produce good/fast enough.
Yes, logicaly, we should go the same path x86 went.
1Ghz, 1.2Ghz, 1.4Ghz, 1.6Ghz
200Mhz increment is a 20-30% speed increment, which is good. Talk in ration, not in number.
But still, we could go from 1.2 to 1.6, or could even go higher 1.2->1.8, 1.2->2.0.....IF apple change is processor manufacturer or moto got something big working on.
As always MHz don't mean jack.
SMP and bandwidth count for power users, Web surfers and spreadsheeters can get by with an 800 Mhz iMac.
<strong>
Give your math co-processor a whack man. 25+25+28=78% increase, not 100%
Also the Moore's Law thing was already pointedd out. Double the transistors not MHz speed, although that has usually been about on target too</strong><hr></blockquote>
1.25 X 1.25 X 1.28 = 2.00
Don't worry kid, you'll get straight on this when you hit percentages in sixth grade.
Your comment on Moore's Law is correct, of course, but in popular culture it has come to mean that CPUs double in power (= speed) every 18 months. This isn't entirely true of course - even within the G4 family the older chips with the shorter pipeline were faster MHz for MHz than the later ones, but it's been surprisingly close. Just another example of the Megahertz Myth.
<strong>
Funny, I could have sworn that two eights are sixteen....</strong><hr></blockquote>
bWHAHAHHAHAHA...i'm done
1.25GHz at 1.8v = 900MHz at 1.3v. So that's the top speed of the TiBook that Apple could manage with the current chip, assuming they use the same low-voltage 7455.
1.25GHz at .18u = 1.73GHz at .13u, assuming the same design and the same power consumption. So Mot can easily reach 1.6GHz with a process shrink - in fact, prototypes of this chip might be responsible for all the rumors about chips at that speed. That also means 1.2GHz low-voltage procs for the TiBook.
Note that I'm not taking bus speed multipliers into account here, so the speeds will have to be adjusted downward as appropriate.
Now, that's the 7455. If Mot releases a new chip, of course, things change. But if it's a 7 stage G4 with, say, another FPU and a RIO interface, these numbers should still be in the ballpark.
Mot's move to .13u is imminent (overdue, actually).
<strong>
Mot's move to .13u is imminent (overdue, actually).</strong><hr></blockquote>
Overdue is an understatement, it's criminal
remember the wanted posters
[ 08-14-2002: Message edited by: rickag ]</p>