So Long CPU Wars. It Was Fun.

[Nik]

Originally posted by: Rumpltzer
Silicon is running out of physics. It's been a long time coming. Scaling can only take you so far when the mobility of the channel material isn't that great. The silicon scaling evolution is drawing to a close, speeds are not increasing, and the CPU makers are being forced into multiple cores.

Evolution isn't cutting it anymore. Throwing bodies at the problem isn't going to cut it. Something revolutionary needs to happen. They need to think hard about what their next step will be. The CMOS makers are working on it, but it's a tough nut to crack and it's going to require an investment of time and money that's much bigger than the huge amounts of time and money that have gone into scaling.

Give it six to ten years to reach the consumer market.

Do you have any links to show this "physical limit" of silicon?

 

[SludgeFactory]

PC makers and retailers will promote three flavors of AMD-based computers, called See, Share and Create models.

Yeah, that's real helpful and totally not confusing to novice and average buyers, a trio of arbitrary one-word labels. Let me guess, marketing is also planning to splash some nice Fisher-Price primary color all over each model.

And it may not be a big deal right away, but Intel without serious competition for a long period of time is not a good thing. There are a few people here old enough to remember what mid- to high-end, consumer-level x86 and Pentiums cost, before AMD stepped it up ~10 years ago. Hopefully they don't go away entirely.

 

[LikeLinus]

Originally posted by: dighn

Originally posted by: FetusCakeMix
I didn't know there ever was a war. Intel makes superior chips.

you must have short memory. there was a period when AMD kicked ass (around the P4 era)

What was the cause of that though? Intel created the P4 to compete against AMDs whole Mhz war. Intel won the Mhz war, but lost a bit of marketshare because the overall performance wasn't as good clock for clock. It was a lesson learned and Intel came back strong. So yes, AMD had a couple of good years and good chips (I had 2-3 of those chips), but Intel is now dominating again. I doubt they will let their grip slip again for awhile.

At this point AMD is a niche market for enthusiest and nothing more. They simply cannot compete against Intel in production volume and sales. It's good to have them around to keep Intel on their toes..but they really need something revolutionary to beat Intel at this point.

 

[techs]

Originally posted by: SludgeFactory

PC makers and retailers will promote three flavors of AMD-based computers, called See, Share and Create models.

Yeah, that's real helpful and totally not confusing to novice and average buyers, a trio of arbitrary one-word labels. Let me guess, marketing is also planning to splash some nice Fisher-Price primary color all over each model.

And it may not be a big deal right away, but Intel without serious competition for a long period of time is not a good thing. There are a few people here old enough to remember what mid- to high-end, consumer-level x86 and Pentiums cost, before AMD stepped it up ~10 years ago. Hopefully they don't go away entirely.

The Pentium 233mmx was top of the line for over 2 years and barely came down in price. Than AMD brought out the k-2's and despite the long lead time Intel had to rush an unready Pentium 2 on to the market.
Competition!

 

[Cogman]

Originally posted by: LikeLinus

Originally posted by: dighn

Originally posted by: FetusCakeMix
I didn't know there ever was a war. Intel makes superior chips.

you must have short memory. there was a period when AMD kicked ass (around the P4 era)

What was the cause of that though? Intel created the P4 to compete against AMDs whole Mhz war. Intel won the Mhz war, but lost a bit of marketshare because the overall performance wasn't as good clock for clock. It was a lesson learned and Intel came back strong. So yes, AMD had a couple of good years and good chips (I had 2-3 of those chips), but Intel is now dominating again. I doubt they will let their grip slip again for awhile.

At this point AMD is a niche market for enthusiest and nothing more. They simply cannot compete against Intel in production volume and sales. It's good to have them around to keep Intel on their toes..but they really need something revolutionary to beat Intel at this point.

Agreed, AMD's glory years where due solely to intels inept management. Intel learn it's lesson, pumped in a large amount of money into R&D and now AMD has absolutely no chance of beating them. The i7 has been around for a while now with really not much innovation from intel. My bets are that they are sitting on a fairly big advancement just in case.

 

[yhelothar]

Originally posted by: Rumpltzer
Silicon is running out of physics. It's been a long time coming. Scaling can only take you so far when the mobility of the channel material isn't that great. The silicon scaling evolution is drawing to a close, speeds are not increasing, and the CPU makers are being forced into multiple cores.

Evolution isn't cutting it anymore. Throwing bodies at the problem isn't going to cut it. Something revolutionary needs to happen. They need to think hard about what their next step will be. The CMOS makers are working on it, but it's a tough nut to crack and it's going to require an investment of time and money that's much bigger than the huge amounts of time and money that have gone into scaling.

Give it six to ten years to reach the consumer market.

Moore's law doesn't seem to have slowed down yet.
On the other hand, our needs for computing power have not quite kept up with our hardware.
My 4 year old portable low voltage laptop runs everyday usage 90% as well as my OC'ed quad core gaming powerhouse.

 

[fleabag]

Originally posted by: techs

Originally posted by: SludgeFactory

PC makers and retailers will promote three flavors of AMD-based computers, called See, Share and Create models.

Yeah, that's real helpful and totally not confusing to novice and average buyers, a trio of arbitrary one-word labels. Let me guess, marketing is also planning to splash some nice Fisher-Price primary color all over each model.

And it may not be a big deal right away, but Intel without serious competition for a long period of time is not a good thing. There are a few people here old enough to remember what mid- to high-end, consumer-level x86 and Pentiums cost, before AMD stepped it up ~10 years ago. Hopefully they don't go away entirely.

The Pentium 233mmx was top of the line for over 2 years and barely came down in price. Than AMD brought out the k-2's and despite the long lead time Intel had to rush an unready Pentium 2 on to the market.
Competition!

233mmx was top of the line for two years? what are you talking about?? The PII came out in '97 and the 233mhz MMX Pentium ALSO came out in 1997. So if those two chips came out in the same year, how could the Pentium I MMX be top of the line for two years? They already had 450mhz PIIs in 1999 and so a 233mhz MMX Pentium I could no way be top of the line. Are you talking about 16bit performance or something???

 

[13Gigatons]

AMD throwing in the towel or trying to trick people into buying their stuff. Truthfully if the price is the same I would go with Intel but when AMD is cheaper for 90% same power then I would buy AMD.

Sadly AMD can't produce low price CPU's and make money. They also have to many CPU's on the market.

 

[zinfamous]

Originally posted by: Nik

Originally posted by: Rumpltzer
Silicon is running out of physics. It's been a long time coming. Scaling can only take you so far when the mobility of the channel material isn't that great. The silicon scaling evolution is drawing to a close, speeds are not increasing, and the CPU makers are being forced into multiple cores.

Evolution isn't cutting it anymore. Throwing bodies at the problem isn't going to cut it. Something revolutionary needs to happen. They need to think hard about what their next step will be. The CMOS makers are working on it, but it's a tough nut to crack and it's going to require an investment of time and money that's much bigger than the huge amounts of time and money that have gone into scaling.

Give it six to ten years to reach the consumer market.

Do you have any links to show this "physical limit" of silicon?

interesting.

I started googling b/c I remember hearing this recently, but knew no details. found a Science article (14 September, 2001) that mentioned how previous limit estimates were way off and assuming scaling with manufacturing, could achieve 1 trillion transistors per chip.
(not sure what that means with today's tech, as I haven't stayed updated)

Limits on Silicon Nanoelectronics for Terascale Integration

James D. Meindl,* Qiang Chen, Jeffrey A. Davis

Throughout the past four decades, silicon semiconductor technology has advanced at exponential rates in both performance and productivity. Concerns have been raised, however, that the limits of silicon technology may soon be reached. Analysis of fundamental, material, device, circuit, and system limits reveals that silicon technology has an enormous remaining potential to achieve terascale integration (TSI) of more than 1 trillion transistors per chip. Such massive-scale integration is feasible assuming the development and economical mass production of double-gate metal-oxide-semiconductor field effect transistors with gate oxide thickness of about 1 nanometer, silicon channel thickness of about 3 nanometers, and channel length of about 10 nanometers. The development of interconnecting wires for these transistors presents a major challenge to the achievement of nanoelectronics for TSI.

School of Electrical and Computer Engineering, Microelectronics Research Center, Georgia Institute of Technology, Atlanta, GA 30332-0269, USA.
* To whom correspondence should be addressed.E-mail:james.meindl@mirc.gatech.edu

Then I found this article:
http://www.electronicsweekly.c...l-limit-of-silicon.htm

This presents a processor speed which is 250 times faster than a 2GHz PC today, but before we get excited by this, it is worth noting that this ground-breaking speed for silicon was achieved by cryogenically "freezing" the chip to -451 deg F (4.5 Kelvins). This is just eight degrees above absolute zero- the coldest possible temperature in nature which occurs at -459.67 deg F.

so, assuming you're comfortable operating in the Kelvin range, you can then start to approach the actual physical limit

 

[ConstipatedVigilante]

Easy. Just ship liquid nitrogen instead of Arctic Silver 5 with all the chips.

 

[TehMac]

Originally posted by: bunnyfubbles

Originally posted by: TridenT
We are never going to get a CPU that plays Crysis at max settings now are we?

ah, the super original comment that comes along every time we have an article about something new CPU/GPU related

It's from TridenT. If you had your head stuck up your asshole the whole day I doubt you could think of anything clever either.

 

[zinfamous]

Originally posted by: ConstipatedVigilante
Easy. Just ship liquid nitrogen instead of Arctic Silver 5 with all the chips.

well, liquid nitrogen can get you to ~ -180 C, so -292 F.

not quite cold enough

 

[Fox5]

Originally posted by: LikeLinus

Originally posted by: dighn

Originally posted by: FetusCakeMix
I didn't know there ever was a war. Intel makes superior chips.

you must have short memory. there was a period when AMD kicked ass (around the P4 era)

What was the cause of that though? Intel created the P4 to compete against AMDs whole Mhz war. Intel won the Mhz war, but lost a bit of marketshare because the overall performance wasn't as good clock for clock. It was a lesson learned and Intel came back strong. So yes, AMD had a couple of good years and good chips (I had 2-3 of those chips), but Intel is now dominating again. I doubt they will let their grip slip again for awhile.

At this point AMD is a niche market for enthusiest and nothing more. They simply cannot compete against Intel in production volume and sales. It's good to have them around to keep Intel on their toes..but they really need something revolutionary to beat Intel at this point.

Based on the research time frames involved, I really doubt the P4 was a counter to anything AMD did. The Athlon was on the market for what, less than 2 years before the P4 came out, and only clearly took a Mhz lead toward the end of that.

No, the Athlon represented the best of microprocessor design at the time it came out. The P4 represented what most thought would be the next big thing (including AMD, IBM, and Intel), and only physical limitations held it back. And even despite that, IBM is still going down the path of high clock speeds, but can handle it since they're high priced chips, and Intel outdid AMD at making an Athlon like processor.

BTW, AMD isn't as far behind in processor design as everyone thinks. They're more behind in manufacturing, with a manufacturing process a generation behind Intel in both performance and size. Add a generational improvement to AMD's cpus (~25% increase in speed, ~50% size reduction) and suddenly they'd become a hell of a lot more competitive. I have a feeling AMD's 32nm cpus will be a return to true competition. They won't beat Intel, but Intel will no longer wipe the floor with everything AMD has.

 

[techs]

Originally posted by: Fox5

Originally posted by: LikeLinus

Originally posted by: dighn

Originally posted by: FetusCakeMix
I didn't know there ever was a war. Intel makes superior chips.

you must have short memory. there was a period when AMD kicked ass (around the P4 era)

What was the cause of that though? Intel created the P4 to compete against AMDs whole Mhz war. Intel won the Mhz war, but lost a bit of marketshare because the overall performance wasn't as good clock for clock. It was a lesson learned and Intel came back strong. So yes, AMD had a couple of good years and good chips (I had 2-3 of those chips), but Intel is now dominating again. I doubt they will let their grip slip again for awhile.

At this point AMD is a niche market for enthusiest and nothing more. They simply cannot compete against Intel in production volume and sales. It's good to have them around to keep Intel on their toes..but they really need something revolutionary to beat Intel at this point.

Based on the research time frames involved, I really doubt the P4 was a counter to anything AMD did. The Athlon was on the market for what, less than 2 years before the P4 came out, and only clearly took a Mhz lead toward the end of that.

No, the Athlon represented the best of microprocessor design at the time it came out. The P4 represented what most thought would be the next big thing (including AMD, IBM, and Intel), and only physical limitations held it back. And even despite that, IBM is still going down the path of high clock speeds, but can handle it since they're high priced chips, and Intel outdid AMD at making an Athlon like processor.

BTW, AMD isn't as far behind in processor design as everyone thinks. They're more behind in manufacturing, with a manufacturing process a generation behind Intel in both performance and size. Add a generational improvement to AMD's cpus (~25% increase in speed, ~50% size reduction) and suddenly they'd become a hell of a lot more competitive. I have a feeling AMD's 32nm cpus will be a return to true competition. They won't beat Intel, but Intel will no longer wipe the floor with everything AMD has.

I think you meant to say p3 there? The Athlon caught up to the Pentium 3 in mhz about the 600-800 megahertz range. AMD was the first to 1 ghz.
And from what I remember at the time, the P4 WAS a result of the Athlon catching up to the Pentium in actual clock speed and that Intel was determined not to let that happen again so the skewed the P4 towards greater clock speed at the expensive of real world performance. When the P4 came out it was a something like 1.6 ghz and then it could only beat the Athlon 1ghz in benchmarks that favored clock speed. Athlons still ruled the roost until they upclocked the P4's into the 2.4 to 2.6 range. Even then on a price performance basis the AMD's were ahead.

And I agree the only thing holding back the AMD chips is Intels manufacturing prowess that allows it to put much larger cache on the cpu.

In fact, I don't think either chip has much of a design advantage anymore, since executing x86 instructions has pretty much evolved to its limits. Its now a question of who can put the most cache on the chip and the most cores and who can do it most cheaply.
In that area intel is about, I guess 2 years ahead.

Oh yeah, to the guy who claimed the p233mmx and the p2 came out at the same time, the p2 came out at the very end of the year, was only available in limited quantities and only really became a mainstream chip with the Celeron 300a (the revised celery 300 which was slower than the p233).
Just selling a few thousand chips because thats all you can make doesn't actually make it a retail chip. Especially at 2000 bucks.
And the 233 was really out WAY before it was officially out. All the pentium mmx's were unlocked and all the mobos had manual jumpers to set the multiplier. Many more people were running p200 at 233 than had Pentium 2's for the first year and half they were out.

 

[Fox5]

Originally posted by: techs

Originally posted by: Fox5

Originally posted by: LikeLinus

Originally posted by: dighn

Originally posted by: FetusCakeMix
I didn't know there ever was a war. Intel makes superior chips.

you must have short memory. there was a period when AMD kicked ass (around the P4 era)

What was the cause of that though? Intel created the P4 to compete against AMDs whole Mhz war. Intel won the Mhz war, but lost a bit of marketshare because the overall performance wasn't as good clock for clock. It was a lesson learned and Intel came back strong. So yes, AMD had a couple of good years and good chips (I had 2-3 of those chips), but Intel is now dominating again. I doubt they will let their grip slip again for awhile.

At this point AMD is a niche market for enthusiest and nothing more. They simply cannot compete against Intel in production volume and sales. It's good to have them around to keep Intel on their toes..but they really need something revolutionary to beat Intel at this point.

Based on the research time frames involved, I really doubt the P4 was a counter to anything AMD did. The Athlon was on the market for what, less than 2 years before the P4 came out, and only clearly took a Mhz lead toward the end of that.

No, the Athlon represented the best of microprocessor design at the time it came out. The P4 represented what most thought would be the next big thing (including AMD, IBM, and Intel), and only physical limitations held it back. And even despite that, IBM is still going down the path of high clock speeds, but can handle it since they're high priced chips, and Intel outdid AMD at making an Athlon like processor.

BTW, AMD isn't as far behind in processor design as everyone thinks. They're more behind in manufacturing, with a manufacturing process a generation behind Intel in both performance and size. Add a generational improvement to AMD's cpus (~25% increase in speed, ~50% size reduction) and suddenly they'd become a hell of a lot more competitive. I have a feeling AMD's 32nm cpus will be a return to true competition. They won't beat Intel, but Intel will no longer wipe the floor with everything AMD has.

I think you meant to say p3 there? The Athlon caught up to the Pentium 3 in mhz about the 600-800 megahertz range. AMD was the first to 1 ghz.
And from what I remember at the time, the P4 WAS a result of the Athlon catching up to the Pentium in actual clock speed and that Intel was determined not to let that happen again so the skewed the P4 towards greater clock speed at the expensive of real world performance. When the P4 came out it was a something like 1.6 ghz and then it could only beat the Athlon 1ghz in benchmarks that favored clock speed. Athlons still ruled the roost until they upclocked the P4's into the 2.4 to 2.6 range. Even then on a price performance basis the AMD's were ahead.

And I agree the only thing holding back the AMD chips is Intels manufacturing prowess that allows it to put much larger cache on the cpu.

In fact, I don't think either chip has much of a design advantage anymore, since executing x86 instructions has pretty much evolved to its limits. Its now a question of who can put the most cache on the chip and the most cores and who can do it most cheaply.
In that area intel is about, I guess 2 years ahead.

Microprocessor design takes a long time, especially for a radically new design like the P4. The p4 was probably 5 years in the making when it came out. The effects of the Athlon's mhz advantage was only felt shortly before the P4 hit market, I really doubt the P4 was a reactionary design, especially because microprocessor design was skewing in the direction of the P4 anyway. Intel was the first with a P4 like design and the first to find it didn't work all that well, but everyone had one in the works.

 

[Nik]

Originally posted by: zinfamous

Originally posted by: Nik

Originally posted by: Rumpltzer
Silicon is running out of physics. It's been a long time coming. Scaling can only take you so far when the mobility of the channel material isn't that great. The silicon scaling evolution is drawing to a close, speeds are not increasing, and the CPU makers are being forced into multiple cores.

Evolution isn't cutting it anymore. Throwing bodies at the problem isn't going to cut it. Something revolutionary needs to happen. They need to think hard about what their next step will be. The CMOS makers are working on it, but it's a tough nut to crack and it's going to require an investment of time and money that's much bigger than the huge amounts of time and money that have gone into scaling.

Give it six to ten years to reach the consumer market.

Do you have any links to show this "physical limit" of silicon?

interesting.

I started googling b/c I remember hearing this recently, but knew no details. found a Science article (14 September, 2001) that mentioned how previous limit estimates were way off and assuming scaling with manufacturing, could achieve 1 trillion transistors per chip.
(not sure what that means with today's tech, as I haven't stayed updated)

Limits on Silicon Nanoelectronics for Terascale Integration

James D. Meindl,* Qiang Chen, Jeffrey A. Davis

Throughout the past four decades, silicon semiconductor technology has advanced at exponential rates in both performance and productivity. Concerns have been raised, however, that the limits of silicon technology may soon be reached. Analysis of fundamental, material, device, circuit, and system limits reveals that silicon technology has an enormous remaining potential to achieve terascale integration (TSI) of more than 1 trillion transistors per chip. Such massive-scale integration is feasible assuming the development and economical mass production of double-gate metal-oxide-semiconductor field effect transistors with gate oxide thickness of about 1 nanometer, silicon channel thickness of about 3 nanometers, and channel length of about 10 nanometers. The development of interconnecting wires for these transistors presents a major challenge to the achievement of nanoelectronics for TSI.

School of Electrical and Computer Engineering, Microelectronics Research Center, Georgia Institute of Technology, Atlanta, GA 30332-0269, USA.
* To whom correspondence should be addressed.E-mail:james.meindl@mirc.gatech.edu

Then I found this article:
http://www.electronicsweekly.c...l-limit-of-silicon.htm

This presents a processor speed which is 250 times faster than a 2GHz PC today, but before we get excited by this, it is worth noting that this ground-breaking speed for silicon was achieved by cryogenically "freezing" the chip to -451 deg F (4.5 Kelvins). This is just eight degrees above absolute zero- the coldest possible temperature in nature which occurs at -459.67 deg F.

so, assuming you're comfortable operating in the Kelvin range, you can then start to approach the actual physical limit

Bingo. The whole point of my asking that question was to set him up to prove him wrong

Nobody's "forcing" anyone to go multi-core instead of faster single core.

 

[techs]

Originally posted by: Fox5

Originally posted by: techs

Originally posted by: Fox5

Originally posted by: LikeLinus

Originally posted by: dighn

Originally posted by: FetusCakeMix
I didn't know there ever was a war. Intel makes superior chips.

you must have short memory. there was a period when AMD kicked ass (around the P4 era)

What was the cause of that though? Intel created the P4 to compete against AMDs whole Mhz war. Intel won the Mhz war, but lost a bit of marketshare because the overall performance wasn't as good clock for clock. It was a lesson learned and Intel came back strong. So yes, AMD had a couple of good years and good chips (I had 2-3 of those chips), but Intel is now dominating again. I doubt they will let their grip slip again for awhile.

At this point AMD is a niche market for enthusiest and nothing more. They simply cannot compete against Intel in production volume and sales. It's good to have them around to keep Intel on their toes..but they really need something revolutionary to beat Intel at this point.

Based on the research time frames involved, I really doubt the P4 was a counter to anything AMD did. The Athlon was on the market for what, less than 2 years before the P4 came out, and only clearly took a Mhz lead toward the end of that.

No, the Athlon represented the best of microprocessor design at the time it came out. The P4 represented what most thought would be the next big thing (including AMD, IBM, and Intel), and only physical limitations held it back. And even despite that, IBM is still going down the path of high clock speeds, but can handle it since they're high priced chips, and Intel outdid AMD at making an Athlon like processor.

BTW, AMD isn't as far behind in processor design as everyone thinks. They're more behind in manufacturing, with a manufacturing process a generation behind Intel in both performance and size. Add a generational improvement to AMD's cpus (~25% increase in speed, ~50% size reduction) and suddenly they'd become a hell of a lot more competitive. I have a feeling AMD's 32nm cpus will be a return to true competition. They won't beat Intel, but Intel will no longer wipe the floor with everything AMD has.

I think you meant to say p3 there? The Athlon caught up to the Pentium 3 in mhz about the 600-800 megahertz range. AMD was the first to 1 ghz.
And from what I remember at the time, the P4 WAS a result of the Athlon catching up to the Pentium in actual clock speed and that Intel was determined not to let that happen again so the skewed the P4 towards greater clock speed at the expensive of real world performance. When the P4 came out it was a something like 1.6 ghz and then it could only beat the Athlon 1ghz in benchmarks that favored clock speed. Athlons still ruled the roost until they upclocked the P4's into the 2.4 to 2.6 range. Even then on a price performance basis the AMD's were ahead.

And I agree the only thing holding back the AMD chips is Intels manufacturing prowess that allows it to put much larger cache on the cpu.

In fact, I don't think either chip has much of a design advantage anymore, since executing x86 instructions has pretty much evolved to its limits. Its now a question of who can put the most cache on the chip and the most cores and who can do it most cheaply.
In that area intel is about, I guess 2 years ahead.

Microprocessor design takes a long time, especially for a radically new design like the P4. The p4 was probably 5 years in the making when it came out. The effects of the Athlon's mhz advantage was only felt shortly before the P4 hit market, I really doubt the P4 was a reactionary design, especially because microprocessor design was skewing in the direction of the P4 anyway. Intel was the first with a P4 like design and the first to find it didn't work all that well, but everyone had one in the works.

I wish I had the old Toms Hardware articles from those times. Intel really didn't make a decision on how to proceed after the p3 until pushed by the athlons (actually the p2's since that what the p3's really were). They were working on several designs. When the Athlons came out Intel, or if you believe what you read, the marketing department, pushed intel towards the design that would produce the greatest clock speed. Intel's designers revealed a few years later that they were going to produce an improved p3 on the P4 die process and make a great processor. When the p4 proved a dead end, especially in laptops, Intel went back and took that improved p3 design, improved it some more, and made it on the latest process and added a lot of cache. You and I know the design as the Core processors.

 

[zinfamous]

Originally posted by: Nik

Bingo. The whole point of my asking that question was to set him up to prove him wrong

Nobody's "forcing" anyone to go multi-core instead of faster single core.

to be fair, there is a difference between physical and practical limits. I think it's fair to say that the limits of silicon will never be reached on the practical real-world scale (I assume it will be put into use for massive industrial data processing centers or something similar)

however, the GT article did mention that while they achieved 500ghz only near absolute zero, they did squeak out 350ghz at room temperature. ....that's pretty damn impressive.

...but this is a per-transistor speed; so I'm not sure what this means in terms of heat if you want to cram 1 trillion of these transistors into one chip. Maybe we could power our computers based on the heat energy they output themselves--in theory recycling their own energy for their own use, everyone having a little fusion reactor, or micro-sun if you will for their desktop?

:Q

 

[0roo0roo]

Originally posted by: Marlin1975
Problem is most CPUs are faster then most software needs out there now. I upgraded to a Q6600 and my video editing software only goes up to 55-65% usage on all the cores.

That and Crysis is usually held back by video card, not cpu, from the reviews I see all the time.

when netbooks still run torturously slow when trying to multitask..cpus wars not over at all

i'll believe that cpus don't need to get much faster when converting a video to h264 for whatever happens in a blink of an eye. and yes it does matter, converting stuff to ipods and such is still a pain in the ass.

 

[TridenT]

Originally posted by: zinfamous

Originally posted by: Nik

Bingo. The whole point of my asking that question was to set him up to prove him wrong

Nobody's "forcing" anyone to go multi-core instead of faster single core.

to be fair, there is a difference between physical and practical limits. I think it's fair to say that the limits of silicon will never be reached on the practical real-world scale (I assume it will be put into use for massive industrial data processing centers or something similar)

however, the GT article did mention that while they achieved 500ghz only near absolute zero, they did squeak out 350ghz at room temperature. ....that's pretty damn impressive.

...but this is a per-transistor speed; so I'm not sure what this means in terms of heat if you want to cram 1 trillion of these transistors into one chip. Maybe we could power our computers based on the heat energy they output themselves--in theory recycling their own energy for their own use, everyone having a little fusion reactor, or micro-sun if you will for their desktop?

:Q

Yeah, sure... Let's all have fucking huge rooms dedicated to our "micro-sun." I already have enough heat coming out of my computer right now, I don't need more!!! :|

 

[sandorski]

This is just a Marketing Gimmick people. AMD isn't giving up or some other nonsense. Just 3-4 years ago AMD was "giving up" and supposedly leaving the Desktop Market, you may recall. Didn't happen and it's not happening now.

I suspect that they'll use a General CPU Design for all 3 categories, but it would be interesting if they came up with 3 unique Designs specifically designed for each. I suspect though that they will continue pretty much the way they always have, but just create a Standard set of Bullet Points that PC Makers need to meet to Brand their PCs with the 3 differing Classes.

 

[TridenT]

Originally posted by: sandorski
This is just a Marketing Gimmick people. AMD isn't giving up or some other nonsense. Just 3-4 years ago AMD was "giving up" and supposedly leaving the Desktop Market, you may recall. Didn't happen and it's not happening now.

I suspect that they'll use a General CPU Design for all 3 categories, but it would be interesting if they came up with 3 unique Designs specifically designed for each. I suspect though that they will continue pretty much the way they always have, but just create a Standard set of Bullet Points that PC Makers need to meet to Brand their PCs with the 3 differing Classes.

lol, desktop market is their only hope... I hear that their laptop CPUs are shittayyy!!! Oh, I suppose they could go server.

 

[sandorski]

Originally posted by: TridenT

Originally posted by: sandorski
This is just a Marketing Gimmick people. AMD isn't giving up or some other nonsense. Just 3-4 years ago AMD was "giving up" and supposedly leaving the Desktop Market, you may recall. Didn't happen and it's not happening now.

I suspect that they'll use a General CPU Design for all 3 categories, but it would be interesting if they came up with 3 unique Designs specifically designed for each. I suspect though that they will continue pretty much the way they always have, but just create a Standard set of Bullet Points that PC Makers need to meet to Brand their PCs with the 3 differing Classes.

lol, desktop market is their only hope... I hear that their laptop CPUs are shittayyy!!! Oh, I suppose they could go server.

Back when that rumour/hubbub was floating around, it was thought they'd go Server/Workstation only.

 

[Old Hippie]

Originally posted by: IGBT
it's a new business model. lets see what happens.

Yep. :thumbsup:

 

[alkemyst]

GPU WARZ!!!!!

RAWR!