Latest from MacUser on chips...
Just because I'm posting this doesn't mean I believe it... more from the MacUser rumour mill:
*******
[summarised]
IBM set to announce 2GHz processor to replace the G4
IBM is set ot release new chip that *may* push performance beyond 2GHz. The as yet unnamed chip to debut at Microprocessor Forum in October that means it's two years from production, right!?). Will be the first 64-bit PPC for the desktop. Will include a "system" compatible with AltiVec.
Big cheese from IBM will give frist technical decscription of chip (that means it's five years away, right!?). IBM sources *claim* to be aiming for 2GHz.
Chip has "160 register" vector processing unit.
Because of the vector unit it must be for Apple - QED.
If, maybe, perhaps, pad, whatever, fudge... then smooth transition from Motorola AltiVec.
64-bit = heaven.
Based on Power4.
Despite this announcement we are stupid to print this article as if you are going to be able to buy this next week - there are huge gaps between such announcements and the introduction of real products.
(that last para graph is pretty much what the last paragraph of the article says).
*******
-- Clive
*******
[summarised]
IBM set to announce 2GHz processor to replace the G4
IBM is set ot release new chip that *may* push performance beyond 2GHz. The as yet unnamed chip to debut at Microprocessor Forum in October that means it's two years from production, right!?). Will be the first 64-bit PPC for the desktop. Will include a "system" compatible with AltiVec.
Big cheese from IBM will give frist technical decscription of chip (that means it's five years away, right!?). IBM sources *claim* to be aiming for 2GHz.
Chip has "160 register" vector processing unit.
Because of the vector unit it must be for Apple - QED.
If, maybe, perhaps, pad, whatever, fudge... then smooth transition from Motorola AltiVec.
64-bit = heaven.
Based on Power4.
Despite this announcement we are stupid to print this article as if you are going to be able to buy this next week - there are huge gaps between such announcements and the introduction of real products.
(that last para graph is pretty much what the last paragraph of the article says).
*******
-- Clive
Comments
Doesn't sound like there is much new
Can't find any info in the forums or the news. Maybe it's a subscription-only article.
Alex.
IBM announced the chip that was going to run inside the GameCube last October. Less than 3 months later, they started shipping the chips in quantity to Nintendo.
That being said, I *do* think we'll see these chips around this time next year in our Powermacs, but not before then. Think next September.
Will they be 2+ GHz... no idea. My humble opinion being no. I do expect them to be close 1.7 or so. Again, MHz/GHz doesn't matter. What does matter is how fast the chip is [pipeline, cache, superscalability, etc.]. This will be a good thing for all parties involved.
Stay Tuned!
The Visigothe
Lock please.
"its" ? possessive form of the pronoun "it".
It's shameful how grammar on the internet is losing its accuracy.<hr></blockquote>
Who gives a crap?
-onlooker (you can quote me on that)
I thought that Gamepuke chip was rumored here at AI long before they called it Gekko, and was codename dolphin, or something.
If there is a 2GHz powerMac it better be here much sooner than 2-3 years away. Like closer to Jan in SF.
I'm tired of Apples processor (moto's fault I know) updates. I have an 867/SP PowerMac, and I'm ready to upgrade now, and it's still pretty new really, but I'm not going to buy one those things they came out with last week. I like the design and all, but, it's the multitude of other things wrong with this picture of an upgrade that I wont pay for.
Do I need to list everything. I don't think so. Everybody loves the OS, but the damn computer is a lagging relic in a week.
nuff said
Sorry I strayed off topic. -o
<strong>No link, no article that I can find at macuser.co.uk, and there's already a robust thread on this topic. Way to go Clive!
Lock please.</strong><hr></blockquote>
Oh, a "junior" giving out the orders!?
The article isn't online, it's only in the printed magazine right now. They seem to have a weird policy about publishing to both. Either way it's in the 23 August issue.
Your "robust thread" just chatters on after the first page, this is comparitively "new" information, from a different source. Which may or may not add to the veracity of other sources/threads.
Although it could be that the MacUser writers just summarised from what they read here. :-)
If MacUser post this online it's likely to go up tomorrow or early next week. There's effectively no "facts" in it that I haven't summarised.
1)new case design
2)greatly changed design no the inside
to have people buy one computer just introduced and sell another that is better, isn't that just bad marketing?
<strong>I'll just say I'm getting pretty excited about the second half of 2003.</strong><hr></blockquote>
That's a long time off, is it.
Source was a IBM proceossor designer that I've done some research with. She's never told me anything secret before, so I don't know the realibility, except that she was fairly matter-of-fact about it, and didn't seem like she was hyping something.
the only bad thing about is altivec, which is even at mot, not believed to be continued in that form. future mot and ibm chips will for some reasons not com up with a 128bit altivec. test have shown that a 64bit vector unit is easyer to program and in most cases performs compareable. the situations where a 128bit unit performs better are the realy high edge of whats intellectually possible for a programmer (optimisations etc.).
the new ebook standard seems to be the bluebook for all new coming ibm and mot chips. this standard will be at the most tasks compatible to each other.
As for multiple cores, yes, that's been part of the PPC/Mac roadmap for a long time (if you hunt around on <a href="http://www.mackido.com" target="_blank">www.mackido.com</a> you'll find references), perhaps we'll get there at last.
This time we're talking about IBM, which is (as far as I can tell) basing this processor on existing ISAs. I'd say that Apple could have this in new Macs in well under a year.
I don't think a MWSF intro is at all out of the question.
In fact, with the architecture of the new towers, I think it'd be pretty easy to do drop the new chip in place of the G4s (making the thing a real DDR machine, finally), and get rid of the old El Capitan look with a newer metal case (which we've seen the start of already).
I'm guessing the new G4 with speedholes is Yikes! all over again - just a stepping stone to the real product.
[ 08-22-2002: Message edited by: O and A ]</p>
<strong>
Oh, a "junior" giving out the orders!?
The article isn't online, it's only in the printed magazine right now. They seem to have a weird policy about publishing to both. Either way it's in the 23 August issue.
Your "robust thread" just chatters on after the first page, this is comparitively "new" information, from a different source. Which may or may not add to the veracity of other sources/threads.
Although it could be that the MacUser writers just summarised from what they read here. :-)
If MacUser post this online it's likely to go up tomorrow or early next week. There's effectively no "facts" in it that I haven't summarised.</strong><hr></blockquote>
My apologies, Clive. Your first post made no mention of this article being print only, so I jumped to conclusions.
<strong>
My apologies, Clive. Your first post made no mention of this article being print only, so I jumped to conclusions.</strong><hr></blockquote>
No worries.
-- Clive