IBM Germanian Tech in PPC?

arn

macrumors god
Original poster
Staff member
Apr 9, 2001
14,532
1,817
The Boston Globe gives some more information on the new 110Ghz communication chips from IBM announced last week.

While these chips are different from standard CPU's (including the PPC)... they do note that "Nintey-nine percent of the process steps are common.":

That means that IBM can build the chips on the same production line as its other silicon products, like the PowerPC chips used in Apple Computer Inc.'s Macintosh line. The only major difference is an extra step in which germanium vapor is deposited onto the silicon. Myerson said his firm's chips will not only be faster than those made with exotic materials, but also cheaper to make.
 

allday

macrumors newbie
Jul 17, 2000
9
0
Microsoft pullout

If this report is true, then I'd say it's time to start buying stock in Apple en masse. The work of porting OS X to PCs may be stepping up.
 

strider42

macrumors 65816
Feb 1, 2002
1,460
6
Re: Microsoft pullout

Originally posted by allday
If this report is true, then I'd say it's time to start buying stock in Apple en masse. The work of porting OS X to PCs may be stepping up.
you lost me, why would porting OS X to PC's be stepped up because IBM announces a technology that shares a lot in common with the PPC chips that are in macs already. Seems more like if this was something that really could be integrated into the PPC chip, it would make PC chips made by AMD and intel comlpetely obsolete.

For the record, its not going to happen anyway, not anytime soon.
 

allday

macrumors newbie
Jul 17, 2000
9
0
correction, retraction, and redirection

Sorry about that strider42, somehow my message got sent to this discussion instead of the MS pulling out thingy. Germanian technology is amazing though. There was a special on it on either Discovery or TLC this past week too. Good stuff!
 

Rocketman

macrumors 603
Re: IBM Germanian Tech in PPC?

Originally posted by arn
The Boston Globe gives some more information on the new 110Ghz communication chips from IBM announced last week.

While these chips are different from standard CPU's (including the PPC)... they do note that "Nintey-nine percent of the process steps are common.":

That means that IBM can build the chips on the same production line as its other silicon products, like the PowerPC chips used in Apple Computer Inc.'s Macintosh line. The only major difference is an extra step in which germanium vapor is deposited onto the silicon. Myerson said his firm's chips will not only be faster than those made with exotic materials, but also cheaper to make.
Apple has an annnoying habbit of manufacturing crippleware, presumably to make the product more manufacturable. If so and the IBM technology can produce 110 Ghz for less money than a silicon wafer, them perhaps Apple will simply release 10 Ghz versions?

Rocketman
 

unclepain

macrumors member
Jan 23, 2002
67
0
Va Beach VA
hmmm...

We talked about this last week, but here's my 2 cents. I'm curious to know if using this Germanian process in conjuction with the current PPC silicon would actually make a difference, or make it into a Mac anytime this century. Is there a reason that there is so much difference between the clock cycles of these chips and current PPC? I realize that they are talking about 2 completely different chips, but if it's the germanian that actually increases clock freqs on the silicon and does it at reduced power, and can be fabbed on the current PPC production lines, how hard is this going to be to get into a Mac quickly? Even if a fraction of this speed is possible by fabbing PPC with this technology, there would cease to be a MHz gap- In fact it would be the other way around. Who knows- by the time IBM actually gets this into Apple machines, Intel and AMD will be at 100 GHz themselves anyways. All I know is I still have to wait on my Mac to render my video compressions.
(falls asleep and dreams about instantanious video renders)
 

krossfyter

macrumors 601
Jan 13, 2002
4,297
0
secret city


My guess is, Moore's Law will remain at its current 'incremental' progress. Apple isn't scrapping the G5 quite yet, AMD is n't scrapping the Hammer, and Intel isn't scrapping the... oh.. what are they calling it now? if forget...

[/B]




how do you figure the incremental progress is going to stay the same? is there some history to this that backs that guess up? like for example..

back in the 1980s the chips were all running at a certan standard progessing so much within two years even though a certain company came out with higher speed chips.
 

MasterX (OSiX)

macrumors 6502
Sep 3, 2001
310
0
As I recall Motorola had recently discovered a simmilar super-speed (50Ghz+) system based not on Germanium, but light. As I understood it a computer based on this technology would use this new inexpensive crystal to produce all the chips in the system. Thus the buses are optical and processors are crystal. The entire computer would run at 50Ghz or whatever and was more heat effective. I'm guessing a little (was a while ago). Even so it's so awsome that the two leading chip manufactures are the ones producing the Apple's CPUs. As I see it:
G3: IBM- 1996-2002
G4: Motorola- 1999-2004
G5: IBM/Moto/Apple's new in-house- 2002-2006
G6: Apple/Moto (50Ghz crystal)- 2005-2008
G7: Apple/IBM (100Ghz+ Germanium) 2008+

So as you can see this technology could very well be WAY off. Unless one of the two super-speed architectures is available to form a G5 (not G6). But so far as I've heard the G5 is a traditional PPC, probably a smaller micron (0.15 to 0.1 or if IBM feels cool 0.05) and a bigger Floating/Fixed Point CPU, 64-bit Fixed, 128/256-bit Floating Point Altivec/2 engine.

This is mostly speculation, but then again what is this board for anyway!?
 

ThorPrime

macrumors member
Jun 6, 2001
41
0
Moore's Law


My guess is, Moore's Law will remain at its current 'incremental' progress. Apple isn't scrapping the G5 quite yet, AMD is n't scrapping the Hammer, and Intel isn't scrapping the... oh.. what are they calling it now? if forget...

by 2008 at the soonest we'll see something ~100Ghz in a desktop *that actually runs at something like that speed*
Moore's Law has held almost 100% true since the 60's. I doubtvery much it will be broken soon
 

elgruga

macrumors 6502
Dec 31, 2001
434
0
Canada
is there a Sonnett 10 Ghz upgrade available?

Yeah, I'll buy this German chip, I'm in.
The Germans always make good stuff, I mean look at their cars.

What? Oh. Germanium. ahhhhh!

Not German then? Pity.
 

madamimadam

macrumors 65816
Jan 3, 2002
1,281
0
graydecember, you bet me to it by MILES.... how can we even think about Macs with 110GHz processors anytime soon when the current Bus is holding back 1GHz processors.
 

PCUser

macrumors regular
Mar 1, 2002
123
0
"Gordon Moore made his famous observation in 1965, just four years after the first planar integrated circuit was discovered. The press called it "Moore's Law" and the name has stuck. In his original paper, Moore predicted that the number of transistors per integrated circuit would double every 18 months."

http://www.intel.com/research/silicon/mooreslaw.htm

Moore's Law has NOTHING to do with the *speed* of the processors. It has to do with doubling the transistor count every 18 months. That could be a 200% increase in speed, and it wouldn't matter. Just as long as the transistor count doubled in 18 months.
 

Xapplimatic

macrumors 6502
Oct 23, 2001
417
0
California
Originally posted by madamimadamtimallen
graydecember, you bet me to it by MILES.... how can we even think about Macs with 110GHz processors anytime soon when the current Bus is holding back 1GHz processors.
Sort of a repeat of another post.. This is backwards... The processor is holding back bus development, not the other way around...

Current G4 spec sheet..

Note that the G4's maximum bus is 133 MHz.. hence Apple can't release a faster bus.. it can't go faster than the processor..
 

Catfish_Man

macrumors 68030
Sep 13, 2001
2,579
1
Portland, OR
The bus/ram speed on current G4s...

...is crippling them. The DP 1GHz has a 7.5x bus multiplier for each processor. That means that the total clock speed of the processors is 13 times higher than the bus clock speed. In comparison, a 2 GHz Pentium 4 has a 5x bus multiplier. I think that if the G5 has the rumored 500MHz RapidIO/HyperTransport bus then all the other improvements in it(while very nice), will be pretty much irrelevant. Altivec needs a lot of data and the current bus just can't get it there fast enough.