Sign in with
Sign up | Sign in
Your question

Intel moves Quad Cored to '06 in fear of AMD's 4x4

Last response: in CPUs
Share
July 20, 2006 3:46:10 PM

http://www.theinquirer.net/default.aspx?article=33151
http://www.hardocp.com


Looks like Intel is a lil worried about 4x4

although all the experts on this forum believe 4x4 will be crap

INTEL is not that sure..
July 20, 2006 3:54:59 PM

Antidisestablishmentarianism.
July 20, 2006 3:57:09 PM

Besides, even if 4x4 does suck(IF) who cares, we're getting Quad Cores!
Related resources
July 20, 2006 4:00:52 PM

Ooooh, HOCP has matured to the level of the inquirer... impressive...
July 20, 2006 4:05:14 PM

Quote:
http://www.theinquirer.net/default.aspx?article=33151
http://www.hardocp.com


Looks like Intel is a lil worried about 4x4

although all the experts on this forum believe 4x4 will be crap

INTEL is not that sure..


Um... IMO the Register has a better explanation of why... I really don't think its 4x4, its the earnigs they just reported (down 50% or so). At least this way they can show the street that they are on plan to regain profitibility.

http://www.reghardware.co.uk/2006/07/20/intel_quad-core...
July 20, 2006 4:17:03 PM

Please correct me if I am wrong (As if I really have to ask.)

The AMD 4x4 is planning on using 2 dual core processors....Right? This is a motherboard centric solution to the use of 4 cores. The motherboard will have two AM2 sockets which are managed by the motherbard.

How does the Intel solution seem to impliment this? Is it going to create a quad-core CPU chip - or is it 2 of the core 2 duo, and the solution is also motherboard centric.

Please illuminate the masses if you can.

Cheers.
July 20, 2006 4:21:40 PM

The quad core CPUs Intel is releasing is two Conroe dies on a single chip. Not two seperate processors in a dual socket motherboard.

The only problem with this first quad core CPU is that the two dies (2 cores and 2 cores) will have to talk to each other over the FSB instead of directly. That is in addition to each core on each die having to talk to each other via the FSB. AMD's first quad core, like their dual cores, each core will be able to talk directly to each other core via Hypertransport.
July 20, 2006 4:23:18 PM

The short answer...the Intel Quad Core is two Core2 duos in one package. And your right about the AMD 4x4 being a mobo solution...2 duals in a dual socket mobo.
July 20, 2006 4:23:30 PM

Quote:
Please correct me if I am wrong (As if I really have to ask.)

The AMD 4x4 is planning on using 2 dual core processors....Right? This is a motherboard centric solution to the use of 4 cores. The motherboard will have two AM2 sockets which are managed by the motherbard.

How does the Intel solution seem to impliment this? Is it going to create a quad-core CPU chip - or is it 2 of the core 2 duo, and the solution is also motherboard centric.

Please illuminate the masses if you can.

Cheers.


I think it is suppose to be like the D series except two Conroe's in the single chip.
July 20, 2006 4:27:47 PM

But I thought AMD was indicating that the performance of the X2 alone is more power than you need, and Conroe has gone power crazy where power is not needed, even tho Intel did it in a energy efficient fashion, but AMD is now talking out of both sides of their mouth trying to release 4x4, since it uses more power when admittidly they have produced enough power... :?
July 20, 2006 4:29:05 PM

Quote:
http://www.theinquirer.net/default.aspx?article=33151
http://www.hardocp.com


Looks like Intel is a lil worried about 4x4

although all the experts on this forum believe 4x4 will be crap

INTEL is not that sure..


I think Intel isn't taking any chances anymore and is doing whatever it can to cover EVERY segment of the market...very low budget, budget, mid range, high range, enthusiast, 2p, 4p, big iron, etc.
July 20, 2006 4:50:39 PM

Quote:
http://www.theinquirer.net/default.aspx?article=33151
http://www.hardocp.com


Looks like Intel is a lil worried about 4x4

although all the experts on this forum believe 4x4 will be crap

INTEL is not that sure..


I think Intel isn't taking any chances anymore and is doing whatever it can to cover EVERY segment of the market...very low budget, budget, mid range, high range, enthusiast, 2p, 4p, big iron, etc.Yes, they are hell-bent on total domination....and i bet they'll get it. 2006-2H 07 looks like all Intel.
July 20, 2006 4:54:02 PM

Quote:
http://www.theinquirer.net/default.aspx?article=33151
http://www.hardocp.com


Looks like Intel is a lil worried about 4x4

although all the experts on this forum believe 4x4 will be crap

INTEL is not that sure..


don't think they're worried. this more or less shows how agressively intel is going to compete with AMD going forward. 4x4 was supposed to be AMD's "answer" to conroe, and intel is already answering the answer before AMD has even released it. 4x4 probably won't compete well with conroe anyways because of price and high multithreaded performance which most users don't need. it'll probably just create another high-end segment which intel will now also have a part of.

imo, intel's quad core will more than likely give better performance at a lower cost. inexpensive motherboards, compared to the 4x4 boards which sound like will be expensive (can't say for sure yet though), and only 1 cpu compared to AMD's 2 high priced fx cpus.
July 20, 2006 4:55:57 PM

Don't be silly. Just like the race to 1 ghz, 2 ghz, dual core, etc., BOTH manufacturers want bragging rights to say they were the first to quad-core.
July 20, 2006 5:01:45 PM

Its a bet on multithreaded apps/OSes, that's all. And I think AMD is trying to give the option to go up to 8 cores on the two socket board (once the quads come out). That way they can have essentially an 8P machine with two sockets, while intel will have to either 1) do the same (everybody knows the FSB problems there), or 2) build an 8 core single socket processor. The first probably (speculative) not perform as well, and the second will take longer. I think that's what AMD is trying to leverage as advantage.

Would probably be very nice for a lot of workstation apps either way.
a b à CPUs
July 20, 2006 5:08:58 PM

Meh...who cares if Intel releases quad core...it's not like they are implementing a new design, they're just shoe-horning two Conroe/Kentsfield cores into the same package...Intel is just repeating the same mistake with quad core as they did with rushing dual core to market...

Intel should be worried about 4x4 as they've got nothing that can compare...4x4 will showcase AMD's scalibility and the advantages of hypertransport...
July 20, 2006 5:13:34 PM

Quote:
Meh...who cares if Intel releases quad core...it's not like they are implementing a new design, they're just shoe-horning two Conroe/Kentsfield cores into the same package...Intel is just repeating the same mistake with quad core as they did with rushing dual core to market...

Intel should be worried about 4x4 as they've got nothing that can compare...4x4 will showcase AMD's scalibility and the advantages of hypertransport...


but the difference with pentium D and kentsfield is that the core architecture is far superior to netburst, so packaging two dies in the same manner will give much better performance and is still worth it (imo) for 4 cores.

from the es samples floating around it already looks like kentsfield is a very good cpu.
July 20, 2006 5:18:49 PM

Quote:
Its a bet on multithreaded apps/OSes, that's all. And I think AMD is trying to give the option to go up to 8 cores on the two socket board (once the quads come out). That way they can have essentially an 8P machine with two sockets, while intel will have to either 1) do the same (everybody knows the FSB problems there), or 2) build an 8 core single socket processor. The first probably (speculative) not perform as well, and the second will take longer. I think that's what AMD is trying to leverage as advantage.

Would probably be very nice for a lot of workstation apps either way.



I doubt Intel can afford an 8core at 65nm. I think if AMD does pulll this off they will be heralded as the more innovative company. At least by me.
July 20, 2006 5:22:19 PM

Quote:
I think if AMD does pulll this off they will be heralded as the more innovative company. At least by me.
That's a given, whether they pull it off or not. :roll:
July 20, 2006 5:26:42 PM

The performance penalties for "shoe-horning" two conroe's onto one package, is FAR less than "shoe-horning" two single core's was. Not that it isn't noticiable, I'm sure it would be. Just like a true quad-core single die cpu will perform BETTER than 4x4 with two dual cores will. However, It does sure take a lot of space up to go to 8 cores, I don't see how that would be cost efficient for a while. Wait and see what happens, but Intel's move has nothing to do with Fear. The whole CORE uArch family (including core 1 and core 2), was Intel's response.
July 20, 2006 5:27:55 PM

Inquirer = not reliable source
July 20, 2006 5:29:02 PM

Quote:
Its a bet on multithreaded apps/OSes, that's all. And I think AMD is trying to give the option to go up to 8 cores on the two socket board (once the quads come out). That way they can have essentially an 8P machine with two sockets, while intel will have to either 1) do the same (everybody knows the FSB problems there), or 2) build an 8 core single socket processor. The first probably (speculative) not perform as well, and the second will take longer. I think that's what AMD is trying to leverage as advantage.

Would probably be very nice for a lot of workstation apps either way.



I doubt Intel can afford an 8core at 65nm. I think if AMD does pulll this off they will be heralded as the more innovative company. At least by me.

So, if Intel can't afford an 8core at 65nm, then AMD would go broke even atempting to on either 90nm or 65nm. Do the math. Use a calculator. Start with small numbers.
July 20, 2006 5:30:00 PM

Quote:
I think if AMD does pulll this off they will be heralded as the more innovative company. At least by me.
That's a given, whether they pull it off or not. :roll:

We know he's another diehard AMD fanboy like 9-inch. And I realized everytime he posts he gets ONE star
July 20, 2006 5:31:49 PM

Looking at the current power dissipation of Conore, Kentsfield will be in the 120 W range... for the lower end chips.

For the 3.0 Ghz ones it'll be somewhere in the 160w range?!?!

I guess its not that bad considering its the equivalent of 4 prescotts EXs in there.
July 20, 2006 5:32:46 PM

His point is that due to the better interconnect, the Amd chips will suffer less than the "shoe-horn" approach of previous Intel chips and presumably future ones did/have. Two Quad cores at 65nm gives eight cores in about 6 months on a ATX board. That's great for some apps. Intel needs to get their interconnect (CSI) finished, or they will continue to have scalability issues, but that's a pretty moot point in the PC market, seeing as hardly anyone actually has a need or use for dual cores, let alone eight cores.
July 20, 2006 5:33:57 PM

i think intel is trying to rush kentsfeild to fill a hole in its prices

think about it... you can get a conroe now for (very rough) 200-500 dollars

then there is a five hundred dollar gap between the 6700 and the 6800. A gap that will PROBABLY be filled by kentsfeild! performance wise,, this means that kentsfeild will outperform the best conroe we have right now *obviously*, and that intel isnt really that concerned about 4x4. it seems to me that AMD is doing 4x4 because they cant scale their processors to quad-core. so they make up or it and say annother way is better

when intel gets into eight core... AMD will be left in the dust, because they havent developed a quad core, and it will be difficult to have FOUR SOCKETS on one motherboard.

4x4 may be good in the short term, but in the future it will create big problems

anyways, those are my theories... and i appologize if its hard to understand, or if there is any misspelling.... I just woke up
July 20, 2006 5:34:39 PM

Quote:
Meh...who cares if Intel releases quad core...it's not like they are implementing a new design, they're just shoe-horning two Conroe/Kentsfield cores into the same package...Intel is just repeating the same mistake with quad core as they did with rushing dual core to market...

Intel should be worried about 4x4 as they've got nothing that can compare...4x4 will showcase AMD's scalibility and the advantages of hypertransport...


but the difference with pentium D and kentsfield is that the core architecture is far superior to netburst, so packaging two dies in the same manner will give much better performance and is still worth it (imo) for 4 cores.

from the es samples floating around it already looks like kentsfield is a very good cpu.

Yeah, agreed. And since packaging is a completely separate process from manufacturing the cpu, it is a very logical interim step that has little to no affect on the manufacturing process.

Then once true quad comes out, Intel will likely have an interim 8core(2 quads in one package) following this same model.

Good times ahead!
July 20, 2006 5:36:01 PM

Quote:
Looking at the current power dissipation of Conore, Kentsfield will be in the 120 W range... for the lower end chips.

For the 3.0 Ghz ones it'll be somewhere in the 160w range?!?!


Heat caused dual core frequencies to drop, probably will do the same for quad cores, which reinforces the fact that these things are reaching a case of diminishing returns VERY fast. Great for the server's (which proper interconnects), but a WASTE on the desktop. We'll have to see what the next shrink actually brings.
July 20, 2006 5:38:23 PM

Quote:
His point is that due to the better interconnect, the Amd chips will suffer less than the "shoe-horn" approach of previous Intel chips and presumably future ones did/have. Two Quad cores at 65nm gives eight cores in about 6 months on a ATX board. That's great for some apps. Intel needs to get their interconnect (CSI) finished, or they will continue to have scalability issues, but that's a pretty moot point in the PC market, seeing as hardly anyone actually has a need or use for dual cores, let alone eight cores.


Agreed...that certainly is a drawback of the Intel model that will be realized around the 8core timeframe. And as you state, CSI is the answer. Let's hope they have it ready by then.
July 20, 2006 5:40:36 PM

I guess your screename really tells the truth. AMD has BOTH 4x4 and TRUE quad core in development. Furthermore, TRUE quad-cores have been on the roadmap for FAR longer than 4x4. 4x4 is a sudden move to try and counter Conroe, nothing more. More cores = more heat, higher frequencies = more heat, the both gotta to do a balancing act because of this. Just like the x2 4800 outperformed the FX57 in multi-threaded tasks, but the FX ruled in single-threaded ones, a Dual Core Conroe WILL outperform a Quad Core chip in many tasks (single and dual threaded tasks).
July 20, 2006 5:41:03 PM

Quote:
Looking at the current power dissipation of Conore, Kentsfield will be in the 120 W range... for the lower end chips.

For the 3.0 Ghz ones it'll be somewhere in the 160w range?!?!

I guess its not that bad considering its the equivalent of 4 prescotts EXs in there.
And just what do you think 2x125W FX-62's will be dissipating?
July 20, 2006 5:41:57 PM

Quote:
Looking at the current power dissipation of Conore, Kentsfield will be in the 120 W range... for the lower end chips.

For the 3.0 Ghz ones it'll be somewhere in the 160w range?!?!


Heat caused dual core frequencies to drop, probably will do the same for quad cores, which reinforces the fact that these things are reaching a case of diminishing returns VERY fast. Great for the server's (which proper interconnects), but a WASTE on the desktop. We'll have to see what the next shrink actually brings.

Diminshing returns hasn't been much of a factor in this industry. When one segment advances beyond the other(software vs. hardware and vice versa), the other typically catches up. Same with regard to heat/cooling, power consumption, storage, etc.
July 20, 2006 5:44:12 PM

i cant see why intel would be worried when its architecture for future processors is extremely efficient.
July 20, 2006 5:46:04 PM

Very easy. :idea:

:D  Intel just has to do the following steps:

1. EDIT -> COPY
2. Then EDIT -> PASTE
3. Repeat step 2 four times. :lol: 
July 20, 2006 5:46:21 PM

Two things here:
a) Having two separate sockets allows more surface area for proper cooling.
b) The first core *might* peg 125 watts once in a GREAT, GREAT, while, but the second core gonna spend the VAST majority of it's time at idle, not really pulling much at all.
July 20, 2006 5:48:03 PM

Not a factor?!?!? Bawawawawawawa!
Why exactly do you think we have two cores instead of one that's twice as fast??!?!?!?!?
July 20, 2006 5:53:44 PM

I think we are talking about two different kinds of diminishing returns...or maybe I misinterpreted your email on this regard. Can you provide more detail of what action you were referring to that denotes dimishing returns? Be back shortly.
July 20, 2006 6:05:51 PM

Quote:
Looking at the current power dissipation of Conore, Kentsfield will be in the 120 W range... for the lower end chips.

For the 3.0 Ghz ones it'll be somewhere in the 160w range?!?!

I guess its not that bad considering its the equivalent of 4 prescotts EXs in there.
And just what do you think 2x125W FX-62's will be dissipating?

I'm not an AMD fanboy :roll:

Just stating that it's gonna heat up the room like a preshott.
Amd's is obviously going to be even worse.

250 W? I can miltiply :lol: 
July 20, 2006 6:07:19 PM

Quote:
Very easy. :idea:

:D  Intel just has to do the following steps:

1. EDIT -> COPY
2. Then EDIT -> PASTE
3. Repeat step 2 four times. :lol: 



looks like BM has a new apprenties...


hahahahaha
July 20, 2006 6:25:28 PM

Quote:

Looks like Intel is a lil worried about 4x4

although all the experts on this forum believe 4x4 will be crap

INTEL is not that sure..


No, it's the other way around. AMD announced 4x4 because they found out Intel was releasing Kentsfield early and they couldn't match them until mid 07. 4x4 was not designed to counter Conroe, but Kentsfield.
July 20, 2006 6:49:54 PM

Quote:
Two things here:
a) Having two separate sockets allows more surface area for proper cooling.
b) The first core *might* peg 125 watts once in a GREAT, GREAT, while, but the second core gonna spend the VAST majority of it's time at idle, not really pulling much at all.


I agree. Fiirst off, how long has dual core processors been out? Half a year? Ok... and how many apps out there make use of two processors, not to mention, that XP is not really designed to spread processes very well across two CPUs (just so everyone knows, I only use XP Pro). I don't know any. The only thing I can say that is useful, is it makes it easy to multitask. But who the hell does virus scanning, movie encoding, and gaming at the same time??? And in that case, if you did, your real bottleneck would be the motherboard, because it'd have hard enough time transferring data from RAM to HD to CPU to GPU and vice versa. For example, my dual core laptop is compiling data with one core and the other, as shabodah stated, is idle. So all in all, dual core is nice. Quad core??? WTF?! There goes the electric bill. What's real funny is that one day, computers will be the cause of global warming and not cars and factories! 8O

Until Windows figures out how to utilize dual cores, what's the point of going Quad Core? Not to mention, Vista needs to work on it's RAM usuage!!! Well, I guess it's about time to pull out the Linux book and start learning! (that is if I get a quad core! -.- ) You think by now, Windows would learn something from linux
July 20, 2006 6:57:17 PM

Not sure if you got the intelligence to understand it (well... I am sure that you DONT have the intelligence) but 4x4 is a child of AMDs fear.
Think about what came first and who reacted.
July 20, 2006 6:59:26 PM

The real a**kicker for these quad cores is the Xeon line. You will now be able to get an 8 core server for the price of a 2 or 4 core server a year ago.

The Xeon FSB is already 1333 MHz too...

If 65nm is cranking out this good already, expect 45nm to come earlier than expected as well.
July 20, 2006 7:12:46 PM

Quote:
Two things here:
a) Having two separate sockets allows more surface area for proper cooling.
b) The first core *might* peg 125 watts once in a GREAT, GREAT, while, but the second core gonna spend the VAST majority of it's time at idle, not really pulling much at all.


I agree. Fiirst off, how long has dual core processors been out? Half a year? Ok... and how many apps out there make use of two processors, not to mention, that XP is not really designed to spread processes very well across two CPUs (just so everyone knows, I only use XP Pro). I don't know any. The only thing I can say that is useful, is it makes it easy to multitask. But who the hell does virus scanning, movie encoding, and gaming at the same time??? And in that case, if you did, your real bottleneck would be the motherboard, because it'd have hard enough time transferring data from RAM to HD to CPU to GPU and vice versa. For example, my dual core laptop is compiling data with one core and the other, as shabodah stated, is idle. So all in all, dual core is nice. Quad core??? WTF?! There goes the electric bill. What's real funny is that one day, computers will be the cause of global warming and not cars and factories! 8O

Until Windows figures out how to utilize dual cores, what's the point of going Quad Core? Not to mention, Vista needs to work on it's RAM usuage!!! Well, I guess it's about time to pull out the Linux book and start learning! (that is if I get a quad core! -.- ) You think by now, Windows would learn something from linux

Again, there will always be a time when one segment of the industry advances beyond the other(software vs. hardware and vice versa), and the other typically catches up. Same with regard to heat/cooling, power consumption, storage, OS vs. # of cores, etc.

For example, I remember when I bought my first 40mb hard drive back in the early-mid 80's. At the time, I was told that spending $400 for so much storage space was crazy because I would never fill it up. Well spending that much was pretty crazy, but as you can imagine, it got filled because advances in software and other devices created a scenario where I needed more storage. The same applies in the case of adding cores. Sure, there are few apps out there today that can currently take advantage of the full power of multi cores, but once that mechanism is in place, let the fun begin.

A similar arguement could be made for multi-tasking with dual cores today. More and more applications are being designed to run in the background(virus scanning, system health monitoring, etc.) and the less they affect our ripping, gaming, etc. the better.
July 20, 2006 7:13:24 PM

Quote:
Looks like Intel is a lil worried about 4x4


Either that or their wanting to take it head on and demolish it. That seems to be Intel's more current strategy.
July 20, 2006 7:16:01 PM

Quote:


Intel should be worried about 4x4 as they've got nothing that can compare...4x4 will showcase AMD's scalibility and the advantages of hypertransport...


You idiot ... Intel has a product superior to 4x4 and its available RIGHT NOW: Two socket Woodcrest systems will cost the same as 4x4 be much faster and use half the power.

Its amazing to watch how using a meaningless marketing term like "4x4" will make people blind to the fact that its just a 2 socket motherboard, which exists as dual socket opteron and woodcrest systems right now :roll:
July 20, 2006 7:26:19 PM

I gotta eat lunch, too! lol
Diminishing returns: When the money/effort spent is not balanced by the improvement made by it. Many times you have to increase time and costs exponentially to get linearly better results. Apparently it is cheaper and easier to produce a dual core 2.0ghz processor, than a 4.0ghz one with the same uArch (gotta compair apples to apples, not P4 to Core/k8/Core 2 comparisons would be fair), and it also uses less energy. However, your not gaining twice the performance. CPU performance scales realitively linearly with frequency, but it is definately not a linear scale for multiple cores. Furthermore, increasing voltage and thus, power and heat, does not scale well at all. The FX-62 with a TDP of 125 watts, is only 200mhz faster than the 85 watt X2's. On other tech, GPU's rarely scale as well either, you have to use the pipelines available to get performance increases, regardless if CPU or GPU.
July 20, 2006 7:26:41 PM

Quote:
I think if AMD does pulll this off they will be heralded as the more innovative company. At least by me.
That's a given, whether they pull it off or not. :roll:


I'm a strict person when it comes ot keeping your word. Believe me I will have prolems. I won't buy ATi to this day because they made me wait a year for video card and it was CRAP.
July 20, 2006 7:28:10 PM

Quote:
The real a**kicker for these quad cores is the Xeon line. You will now be able to get an 8 core server for the price of a 2 or 4 core server a year ago.

The Xeon FSB is already 1333 MHz too...

If 65nm is cranking out this good already, expect 45nm to come earlier than expected as well.



agree....

my few additional,

Xeon 5100 series do support Dual FSB.. thats a 1333mhz x2
!