I don't think so -- I read just recently that Intel is adopting a technique that will let them get below 0.09 micron, whereas IBM is going to take SOI down to 0.09 micron before switching to this other new technique. I can't remember the details, but it sounds like there are different roads to travel and Intel and IBM aren't following eachother. IMO this is a good thing.</strong><hr></blockquote>
I believe it's called stretched or strained SOI, but still SOI.
"To date, IBM Corp. has been the main proponent of strained silicon, with a plan to add its version of strained silicon to IBM's 65-nm process node, which is expected to move to first manufacturing in 2005. IBM plans to combine its strained silicon expertise with its silicon-on-insulator capabilities. However, IBM researchers have indicated at least a 10 percent cost adder for strained silicon, and said much work remains to balance the mobility enhancement in the NMOS and PMOS portions of a CMOS device.
Dan Hutcheson, president of VLSI Research Corp. (San Jose, Calif.), said "it is pretty clear that Intel has made a major breakthrough here. It is amazing that they would use strained silicon at the 90-nm node, and if the cost adder is only 2 percent then the process additions would need to be pretty trivial.""
[ 12-10-2002: Message edited by: rickag ]
One last lament Oh great googly moogly, I could have sworn I read somewhere that Intel was going to combined strained silicon and SOI. me goes to corner to mend my ways.
Well, AMD haven't achieved 'critical mass' and have done a reasonable job of competing with Intel 'head on' for the last several years.
The 'Clawhammer' and 'Opteron' offer the chance to get some of those meaty bus' contracts and start making some money. Being the gamer's friend alone won't help them. It's AMD's chance to go after Intel on their home turf.
I can't accept Macluv's insinuation that IBM is too bloated to cope with Intel. A company, suffeting from bloat itself. And even with 'critical' mass has found it hard going with its Itanic cpu and has hardly blown AMD out the water with its Pentium 4. Too me, Intel look like a bloated monopoly who look vulnerable to any company with a reasonable business plan and a billion or so. Especially as 'we' move to a post x86 era. I think the whole market is up for grabs. I sez Intel and M$ will last a little less longer than the Romans did...
Intel and M$ coalition will put up a fight. They do have billions to protect their yard. But I think some of that "95%" is up for grabs.
AMD, a company who are still bleeding to death did a remarkable job of taking desktop cpu share away from Intel. Their share has fluctuated and petered downwards since the Pentium 4 dug in. Point? IBM has much more in the way of resources to bare against Intel and will bring it to bare if Intel aims its guns at IBM's turf.
The 970 is an IBM cpu where IBM's barely flexed its muscles! It's a derivative of another world class cpu. It'll scale better for Apple's desktop purposes.
IBM as a bloated company can't compete? If that's a central argument... It may have been true one day a decade ago. But these days? IBM is the kinda company you wouldn't want to take on. Not in terms of CPU tech'. IBM have been state of the art with their cpu tech R$D labs for years now. All that's lacked from a cpu market place point of view is a reason to go after the desktops. If Intel/Linux are even thinking about storming IBM's server and service gates then they'd better think again.
I think the IBM and Apple planets are in alignment for the next PPC chapter.
I don't think so -- I read just recently that Intel is adopting a technique that will let them get below 0.09 micron, whereas IBM is going to take SOI down to 0.09 micron before switching to this other new technique. I can't remember the details, but it sounds like there are different roads to travel and Intel and IBM aren't following eachother. IMO this is a good thing.</strong><hr></blockquote>
Intel have said they will adopt what they call "Fully depleted substrate", which is very similar to SOI in behaviour, although acheived slightly differently.
The other technology you may have heard of is strained silicon, which IBM have said they will use at .065micron, with SOI. This technology increases the mobility of holes, and hence the speed of p-channel transistors, giving a slight improvement to switcing speed.
Another technology that may appear is isotopically pure silicon28, another way of increasing mobility, it looks as if AMD may be the first with this one.
I expect they will all be in use by the .065micron node.
LBB-- really quick reply to 1st part of your post. I was wondering why you were saying "AMD" hasn't achieved critical mass, as companies don't achieve critical mass, technology does. Are you confused by this statement?
[quote]originally said by me<strong>
Intel, on the other hand, has almost *unlimited* resources to push the x86 further as >it< has already achieved critical mass.
</strong><hr></blockquote>
"It" refers to x86, not Intel. Sorry for the misunderstanding. If you want to understand the importance of critical mass, look up Metcalf's Law and Law of Disruption. I was first introduced to the concept by reading the Bill Gates/Sculley memo about seven years ago. I didn't want to believe it back then.
If you would like me to elaborate I'll be happy to, right now I have to get some sleep.
Speaking as a non-participant in this argument, and therefore fairly objective....Macluv, you really need practice discussing/debating like an adult.
Fish
ps btw, before you use terms like "strawman" (in a post where you criticize someone's lack of knowledge of debate, no less), you should get a handle on what it means.
<strong>Speaking as a non-participant in this argument, and therefore fairly objective....Macluv, you really need practice discussing/debating like an adult.
Fish
ps btw, before you use terms like "strawman" (in a post where you criticize someone's lack of knowledge of debate, no less), you should get a handle on what it means.</strong><hr></blockquote>
Alrighty then. Thanks for stoppin' by. Give my best to Aunt Bess and the folks. I left an extra pack of condoms in your purse just in case you get lucky.
Alrighty then. Thanks for stoppin' by. Give my best to Aunt Bess and the folks. I left an extra pack of condoms in your purse just in case you get lucky.
Pa.</strong><hr></blockquote>
A childish reply to someone telling you you should learn to discuss like an adult. Great. Gee, you seem to _really_ have too much time on your hands - is your computer locker for pr0n or something?
Well, that was true to form, luv, and symptomatic of why your "arguments" are not given much merit here (based on this thread, at any rate).
Ignore the criticism if you want, but if you really care about getting your ideas across effectively, you should pay attention to the almost universal negative reaction to your posts. If you DO have worthwhile ideas you are positing here, they are lost in the endless (and often inane) stream of your smiley/sarcasm-laden posts.
If you don't care about getting your points across, then I suppose you are just enjoying trying to sound dismissive (although it comes off as childish rather than lofty). In that case - enjoy yourself!
Here I was checking up on this thread thinking there was actually some funded discussion on future hardware going on. <img src="graemlins/oyvey.gif" border="0" alt="[No]" />
I just heard that Bubba's CPUs was going to be using genuine melted Rubber Duckies for insulator and pairing them with Zilog Z80s from antique Kaypro II's to create some kind of whozit oven mit for Itanic iceberg lettuce processors!
AI could really use an IGNORE or FFW button. Why don't U guys meet at the Apple booth come MWSF and throw some punches. May the best loser win. I find myself scrolling till i get to a PROGRAMMER post and find something worth reading.
.... In the fall of 2000, IBM assembled its 970 development team at the request of Apple. The objective was to have the performance of the Power 4, at a much lower cost. IBM also saw potential in this chip for its linux solutions. IBM began delivering Apple engineering samples of the 970 in May of this year, about the time the 970 was taped out. Apple has numerous working prototypes, and will begin to send them to key developers later this month....
quote from the article [quote]"Asked whether IBM will develop a PowerPC that includes an AltiVec coprocessor, Elliott Newcombe, PowerPC product marketing manager at IBM's Research Triangle Park facility, said IBM is considering adding an AltiVec APU to one of IBM's designs.
"AltiVec could be an example of one of these coprocessors that could be plugged in [to a Book E-compliant PowerPC]," Newcombe said. "Nothing precludes IBM from doing that, but I cannot comment on whether a design is in progress. The market will decide whether we do that, and I can just say 'stay tuned.' ""<hr></blockquote>
If Apple had IBM assemble a team for the 970 in 2000, then, I am anxiously awaiting the introduction of the cpu Mr. Elliott Newcombe mentions in 1999. After all it has been in the planning stages a year earlier than the 970.
If Apple had IBM assemble a team for the 970 in 2000, then, I am anxiously awaiting the introduction of the cpu Mr. Elliott Newcombe mentions in 1999. After all it has been in the planning stages a year earlier than the 970. . .
</strong><hr></blockquote>
For every project carried to completion, there are likely ten or more that are started and then dropped at some stage. Many never make it past the analysis phase. The only thing we can pretty much bet on is the 970. Anything else is pure speculation, but we can dream -- what if? It sounds like you are hoping for a G4 replacement from IBM? That would be nice.
<strong>It sounds like you are hoping for a G4 replacement from IBM? That would be nice.</strong><hr></blockquote>
???
The ppc970 is claimed by IBM to be roughly twice as fast _at the same clock_ as the current G4s. Sure, it runs a LOT hotter when run at 1.8 GHz on a .13 micron process, but picture this:
At 0.6 GHz, it should be as fast as one of the fastest current G4s. It should produce a relatively miniscule amount of heat at that speed. I don't have the power/Hz curve, but that's obvious.
So either the ppc970 could work fine (at a lower clock), or at the least a mild variant would _certainly_ be fine. I mean, the ppc970 would work fine with one FPU disabled completely - which reduces heat.
And IBM is expected to convert to 0.9 by the end of 2003 - that's before I'd expect the laptop lines to switch over to a totally new processor _anyway_.
For every project carried to completion, there are likely ten or more that are started and then dropped at some stage...</strong><hr></blockquote>
No Doubt.
Sorry I didn't better clarify. I was feebly trying to point out a possible discrepency in the original macrumors report that started this topic. Hence, my reference to the 2 choices of clickable smilies.
If the 1999 reference was to a future 970 it blows a hole in the article starting this topic. On the other hand, it may be reference to another desktop cpu.
If this isn't the 970, IBM will be introducing another cpu(re: most probably 32 bit) for desktops in addition to the 970(re: 64 bit). If it is referencing the 970, IBM may have begun assemblying the team in 1999, a year before Apple requested this development team, according to the macrumors report.
A project started earlier than the 970 could have been converted into the 970 project later. Or, if it is a separate project that is still alive and well, it could be like you suggest, a 32 bit processor. It might be advantageous to have a G4 like processor from IBM, optimized for lower cost and power, but with the same processor bus type as the 970. It may allow using the same motherboard chip set for both pro and consumer Macs, and consumer Macs would benefit from better memory performance. However, this is still just wild speculation.
Like Nevyn seems to be saying above, once the 970 process is shrunk low enough, the price and power could be very low in the single core version. In that case, why not use it in everything? The top performance chips could have two cores, or more eventually. The top chip will someday be based on the Power 5 no doubt.
A project started earlier than the 970 could have been converted into the 970 project later. Or, if it is a separate project that is still alive and well, it could be like you suggest, a 32 bit processor. It might be advantageous to have a G4 like processor from IBM, optimized for lower cost and power, but with the same processor bus type as the 970. It may allow using the same motherboard chip set for both pro and consumer Macs, and consumer Macs would benefit from better memory performance. However, this is still just wild speculation.
Like Nevyn seems to be saying above, once the 970 process is shrunk low enough, the price and power could be very low in the single core version. In that case, why not use it in everything? The top performance chips could have two cores, or more eventually. The top chip will someday be based on the Power 5 no doubt.</strong><hr></blockquote>
I would like to point out that we still don't know that much about the 970. And, correct me if I'm wrong, the 970 is a processor that is extremely dependent on a GREAT compilier. In that case I think that Apple has chose well. At least that is where I would want the performance ball to be at, in the compilier and not relying on brute force number crunching, ie. lots of heat. I can't wait for Hannabls' next installment. My point about the compilier is that it is easier and quicker to fix than the CPU.
It may allow using the same motherboard chip set for both pro and consumer Macs, and consumer Macs would benefit from better memory performance. However, this is still just wild speculation.
...the price and power could be very low in the single core version. </strong><hr></blockquote>
OT fantasy about IBM pimping high performance, low cost ATX boards to the Linux masses. If Linux is an acceptable solution for corps then the age-old "we won't risk anything but x86" hardware really starts to fall apart. It would benefit Apple trememdously even if they keep their hardware separate.
Comments
<strong>
I don't think so -- I read just recently that Intel is adopting a technique that will let them get below 0.09 micron, whereas IBM is going to take SOI down to 0.09 micron before switching to this other new technique. I can't remember the details, but it sounds like there are different roads to travel and Intel and IBM aren't following eachother. IMO this is a good thing.</strong><hr></blockquote>
I believe it's called stretched or strained SOI, but still SOI.
<a href="http://news.com.com/2100-1001-268083.html"; target="_blank">web pageIBM's version</a>
<a href="http://zdnet.com.com/2100-1103-976120.html"; target="_blank">Another article about IBM's</a>
Opps, my mistake. SOI and strained silicon not related.
<a href="http://www.eetimes.com/semi/news/OEG20020813S0012"; target="_blank">yet another</a> and this
"To date, IBM Corp. has been the main proponent of strained silicon, with a plan to add its version of strained silicon to IBM's 65-nm process node, which is expected to move to first manufacturing in 2005. IBM plans to combine its strained silicon expertise with its silicon-on-insulator capabilities. However, IBM researchers have indicated at least a 10 percent cost adder for strained silicon, and said much work remains to balance the mobility enhancement in the NMOS and PMOS portions of a CMOS device.
Dan Hutcheson, president of VLSI Research Corp. (San Jose, Calif.), said "it is pretty clear that Intel has made a major breakthrough here. It is amazing that they would use strained silicon at the 90-nm node, and if the cost adder is only 2 percent then the process additions would need to be pretty trivial.""
[ 12-10-2002: Message edited by: rickag ]
One last lament
[ 12-10-2002: Message edited by: rickag ]</p>
The 'Clawhammer' and 'Opteron' offer the chance to get some of those meaty bus' contracts and start making some money. Being the gamer's friend alone won't help them. It's AMD's chance to go after Intel on their home turf.
I can't accept Macluv's insinuation that IBM is too bloated to cope with Intel. A company, suffeting from bloat itself. And even with 'critical' mass has found it hard going with its Itanic cpu and has hardly blown AMD out the water with its Pentium 4. Too me, Intel look like a bloated monopoly who look vulnerable to any company with a reasonable business plan and a billion or so. Especially as 'we' move to a post x86 era. I think the whole market is up for grabs. I sez Intel and M$ will last a little less longer than the Romans did...
Intel and M$ coalition will put up a fight. They do have billions to protect their yard. But I think some of that "95%" is up for grabs.
AMD, a company who are still bleeding to death did a remarkable job of taking desktop cpu share away from Intel. Their share has fluctuated and petered downwards since the Pentium 4 dug in. Point? IBM has much more in the way of resources to bare against Intel and will bring it to bare if Intel aims its guns at IBM's turf.
The 970 is an IBM cpu where IBM's barely flexed its muscles! It's a derivative of another world class cpu. It'll scale better for Apple's desktop purposes.
IBM as a bloated company can't compete? If that's a central argument... It may have been true one day a decade ago. But these days? IBM is the kinda company you wouldn't want to take on. Not in terms of CPU tech'. IBM have been state of the art with their cpu tech R$D labs for years now. All that's lacked from a cpu market place point of view is a reason to go after the desktops. If Intel/Linux are even thinking about storming IBM's server and service gates then they'd better think again.
I think the IBM and Apple planets are in alignment for the next PPC chapter.
Lemon Bon Bon
<strong>
I don't think so -- I read just recently that Intel is adopting a technique that will let them get below 0.09 micron, whereas IBM is going to take SOI down to 0.09 micron before switching to this other new technique. I can't remember the details, but it sounds like there are different roads to travel and Intel and IBM aren't following eachother. IMO this is a good thing.</strong><hr></blockquote>
Intel have said they will adopt what they call "Fully depleted substrate", which is very similar to SOI in behaviour, although acheived slightly differently.
The other technology you may have heard of is strained silicon, which IBM have said they will use at .065micron, with SOI. This technology increases the mobility of holes, and hence the speed of p-channel transistors, giving a slight improvement to switcing speed.
Another technology that may appear is isotopically pure silicon28, another way of increasing mobility, it looks as if AMD may be the first with this one.
I expect they will all be in use by the .065micron node.
michael
[quote]originally said by me<strong>
Intel, on the other hand, has almost *unlimited* resources to push the x86 further as >it< has already achieved critical mass.
</strong><hr></blockquote>
"It" refers to x86, not Intel. Sorry for the misunderstanding. If you want to understand the importance of critical mass, look up Metcalf's Law and Law of Disruption. I was first introduced to the concept by reading the Bill Gates/Sculley memo about seven years ago. I didn't want to believe it back then.
If you would like me to elaborate I'll be happy to, right now I have to get some sleep.
[ 12-10-2002: Message edited by: MacLuv ]</p>
Fish
ps btw, before you use terms like "strawman" (in a post where you criticize someone's lack of knowledge of debate, no less), you should get a handle on what it means.
<strong>....
If you would like me to elaborate I'll be happy to, right now I have to get some sleep.
[ 12-10-2002: Message edited by: MacLuv ]</strong><hr></blockquote>
No thank you
<strong>Speaking as a non-participant in this argument, and therefore fairly objective....Macluv, you really need practice discussing/debating like an adult.
Fish
ps btw, before you use terms like "strawman" (in a post where you criticize someone's lack of knowledge of debate, no less), you should get a handle on what it means.</strong><hr></blockquote>
Alrighty then. Thanks for stoppin' by. Give my best to Aunt Bess and the folks. I left an extra pack of condoms in your purse just in case you get lucky.
Pa.
<strong>
Alrighty then. Thanks for stoppin' by. Give my best to Aunt Bess and the folks. I left an extra pack of condoms in your purse just in case you get lucky.
Pa.</strong><hr></blockquote>
A childish reply to someone telling you you should learn to discuss like an adult. Great. Gee, you seem to _really_ have too much time on your hands - is your computer locker for pr0n or something?
Ignore the criticism if you want, but if you really care about getting your ideas across effectively, you should pay attention to the almost universal negative reaction to your posts. If you DO have worthwhile ideas you are positing here, they are lost in the endless (and often inane) stream of your smiley/sarcasm-laden posts.
If you don't care about getting your points across, then I suppose you are just enjoying trying to sound dismissive (although it comes off as childish rather than lofty). In that case - enjoy yourself!
Fish
Neat huh?
Now leave my straw man alone!
<strong>"I would not be just a nothin' my head all full of stuffin'
My heart all full of pain.
I would dance and be merry, life would be a ding-a-derry,
If I only had a brain."
♩♪♩♪♫♩</strong><hr></blockquote>
!!!! TONTON YOU ****IN' GOT IT! CHARLIE YOU WIN!
*SMOOTCH*
PS: <img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> <img src="graemlins/bugeye.gif" border="0" alt="[Skeptical]" /> <img src="graemlins/oyvey.gif" border="0" alt="[No]" />
[ 12-11-2002: Message edited by: MacLuv ]</p>
<strong>
Where does this leave Apple?
.... In the fall of 2000, IBM assembled its 970 development team at the request of Apple. The objective was to have the performance of the Power 4, at a much lower cost. IBM also saw potential in this chip for its linux solutions. IBM began delivering Apple engineering samples of the 970 in May of this year, about the time the 970 was taped out. Apple has numerous working prototypes, and will begin to send them to key developers later this month....
</strong><hr></blockquote>
<a href="http://www.eetimes.com/story/OEG19990507S0003"; target="_blank">This EETIMES</a> article was posted May 7, 1999.
quote from the article [quote]"Asked whether IBM will develop a PowerPC that includes an AltiVec coprocessor, Elliott Newcombe, PowerPC product marketing manager at IBM's Research Triangle Park facility, said IBM is considering adding an AltiVec APU to one of IBM's designs.
"AltiVec could be an example of one of these coprocessors that could be plugged in [to a Book E-compliant PowerPC]," Newcombe said. "Nothing precludes IBM from doing that, but I cannot comment on whether a design is in progress. The market will decide whether we do that, and I can just say 'stay tuned.' ""<hr></blockquote>
If Apple had IBM assemble a team for the 970 in 2000, then, I am anxiously awaiting the introduction of the cpu Mr. Elliott Newcombe mentions in 1999. After all it has been in the planning stages a year earlier than the 970.
I don't know if this
or this
is appropriate
<strong>
If Apple had IBM assemble a team for the 970 in 2000, then, I am anxiously awaiting the introduction of the cpu Mr. Elliott Newcombe mentions in 1999. After all it has been in the planning stages a year earlier than the 970. . .
</strong><hr></blockquote>
For every project carried to completion, there are likely ten or more that are started and then dropped at some stage. Many never make it past the analysis phase. The only thing we can pretty much bet on is the 970. Anything else is pure speculation, but we can dream -- what if? It sounds like you are hoping for a G4 replacement from IBM? That would be nice.
<strong>It sounds like you are hoping for a G4 replacement from IBM? That would be nice.</strong><hr></blockquote>
???
The ppc970 is claimed by IBM to be roughly twice as fast _at the same clock_ as the current G4s. Sure, it runs a LOT hotter when run at 1.8 GHz on a .13 micron process, but picture this:
At 0.6 GHz, it should be as fast as one of the fastest current G4s. It should produce a relatively miniscule amount of heat at that speed. I don't have the power/Hz curve, but that's obvious.
So either the ppc970 could work fine (at a lower clock), or at the least a mild variant would _certainly_ be fine. I mean, the ppc970 would work fine with one FPU disabled completely - which reduces heat.
And IBM is expected to convert to 0.9 by the end of 2003 - that's before I'd expect the laptop lines to switch over to a totally new processor _anyway_.
<strong>
For every project carried to completion, there are likely ten or more that are started and then dropped at some stage...</strong><hr></blockquote>
No Doubt.
Sorry I didn't better clarify. I was feebly trying to point out a possible discrepency in the original macrumors report that started this topic. Hence, my reference to the 2 choices of clickable smilies.
If the 1999 reference was to a future 970 it blows a hole in the article starting this topic. On the other hand, it may be reference to another desktop cpu.
<a href="http://www-3.ibm.com/chips/products/powerpc/rdmap/roadmap_small.jpg"; target="_blank">IBM's Roadmap</a> shows the next generation PowerPC to have the following
1+ GHz
Multicore Superscaler
SMP Capable
Integrated SIMD engine(vroommm vrooommm)
Rapid I/O
n-way Crossbar Coreconnect
If this isn't the 970, IBM will be introducing another cpu(re: most probably 32 bit) for desktops in addition to the 970(re: 64 bit). If it is referencing the 970, IBM may have begun assemblying the team in 1999, a year before Apple requested this development team, according to the macrumors report.
[ 12-12-2002: Message edited by: rickag ]
and i was trying to get back on topic
[ 12-12-2002: Message edited by: rickag ]</p>
<strong>
and i was trying to get back on topic
</strong><hr></blockquote>
What a novel idea!
A project started earlier than the 970 could have been converted into the 970 project later. Or, if it is a separate project that is still alive and well, it could be like you suggest, a 32 bit processor. It might be advantageous to have a G4 like processor from IBM, optimized for lower cost and power, but with the same processor bus type as the 970. It may allow using the same motherboard chip set for both pro and consumer Macs, and consumer Macs would benefit from better memory performance. However, this is still just wild speculation.
Like Nevyn seems to be saying above, once the 970 process is shrunk low enough, the price and power could be very low in the single core version. In that case, why not use it in everything? The top performance chips could have two cores, or more eventually. The top chip will someday be based on the Power 5 no doubt.
<strong>
What a novel idea!
A project started earlier than the 970 could have been converted into the 970 project later. Or, if it is a separate project that is still alive and well, it could be like you suggest, a 32 bit processor. It might be advantageous to have a G4 like processor from IBM, optimized for lower cost and power, but with the same processor bus type as the 970. It may allow using the same motherboard chip set for both pro and consumer Macs, and consumer Macs would benefit from better memory performance. However, this is still just wild speculation.
Like Nevyn seems to be saying above, once the 970 process is shrunk low enough, the price and power could be very low in the single core version. In that case, why not use it in everything? The top performance chips could have two cores, or more eventually. The top chip will someday be based on the Power 5 no doubt.</strong><hr></blockquote>
I would like to point out that we still don't know that much about the 970. And, correct me if I'm wrong, the 970 is a processor that is extremely dependent on a GREAT compilier. In that case I think that Apple has chose well. At least that is where I would want the performance ball to be at, in the compilier and not relying on brute force number crunching, ie. lots of heat. I can't wait for Hannabls' next installment. My point about the compilier is that it is easier and quicker to fix than the CPU.
<strong>
It may allow using the same motherboard chip set for both pro and consumer Macs, and consumer Macs would benefit from better memory performance. However, this is still just wild speculation.
...the price and power could be very low in the single core version. </strong><hr></blockquote>
OT fantasy about IBM pimping high performance, low cost ATX boards to the Linux masses. If Linux is an acceptable solution for corps then the age-old "we won't risk anything but x86" hardware really starts to fall apart. It would benefit Apple trememdously even if they keep their hardware separate.