Not really, unless you use a highend SLI/CF setup. The appealing part would be quadcores. Personally I wait for nehalem. Since K10 looks like a R600 hype aswell.Originally Posted by Darkatom
Not really, unless you use a highend SLI/CF setup. The appealing part would be quadcores. Personally I wait for nehalem. Since K10 looks like a R600 hype aswell.
Crunching for Comrades and the Common good of the People.
Well by the time wolfdale is released i'll most likely be running 8900gtx SLI. But in what way do you mean that wolfdale will make a difference for a high end SLI setup ?
Since games wont be utilizing quadcores much if any by the time. The SLi/CF setup would most likely leave the CPU as the bottleneck.
Crunching for Comrades and the Common good of the People.
dumbest reply you ever made....
Well if you disagree that intel labels their processors with a 65W TDP
prove me wrong
(not server and not extreme parts....)
Did you already forget the episode on page 2 where you made a fool of yourself on that subject?
Crunching for Comrades and the Common good of the People.
@GoThr3k
So, lets see. You don't believe that 4GHz 45nm C2D with 65W TDP is possible?
According to the 45nm High-K + metal gate parameters(Vcore, IdSat, etc) it provides 30% lower transistor power consumption at 20+% higher frequency.
Lets consider a 65W TDP E6700 as base in the calculation.
2.67GHz * 1.2 = 3.2GHz
65W * 0.7 = 45.5W
If for the 45nm @4GHz the Vcore remains same as for 3.2GHz, it will waste 56.875W.
(4GHz / 3.2GHz) * 45.5W = 1.25 * 45.5W = 56.875W
To 65W, there is +14.3% headroom.
65W / 56.875W = 1.143
Or you can increase Vcore to +6.9% and still be in 65W TDP.
Square root of 1.143 is 1.069.
All E6700 can OC at least 25% and be stable at default Vcore. So, its not a miracle for 65W TDP Woldfdale to have 25% higher clock at 7% higher Vcore.
i see
thx for that calculation,very interesting
are you sure its a 30% less power consumption AND a 20% increase in frequency?
Or is it more like a 30% less power consumption,and as a result of the smaller proces a 20% increase in frequency is possible?
@Shintai,are you a 12year old kid that doesnt get enough love or something??
I was just referring that intel cpu's were labeled at 65W TDP,i dont say they consume that..., is say that they are labeled at 65W
if it consumes 50W in reality or 70W i don't care, as long as intel's tdp calculation remains the same, so that we can compare wolfsdale and Conroe at the same TDP,if its 50/10/170 W, that doesnt matter,they will have the same TDP...
LoL,comparing theoretical High K 45nm on paper with an actual product yet to be seen...This is not the perfect world you know.Things in theory don't come up the same way in reality.
I'm not saying it will not happen,just that your "calculation" is worthless.
If anything, the calculation is conservative.
Never count on a penryn @3.6ghz this year
actually it is liberal considering. A little unknown fact for you but if the 286, 386, 486, and (original) Pentium were made on the same process, they would have all run at exactly the same clock speed.
Fast computers breed slow, lazy programmers
The price of reliability is the pursuit of the utmost simplicity. It is a price which the very rich find most hard to pay.
http://www.lighterra.com/papers/modernmicroprocessors/
Modern Ram, makes an old overclocker miss BH-5 and the fun it was
It depends of K10 performance and frequency. If Intel needs 3.6GHz to counter K10, I am pretty sure we'll see a 3.67GHz this year. But if they don't need, they will sell lower clocked Core2's as long as they can milk money.Never count on a penryn @3.6ghz this year
Again, you are talking nonsense. 286, 386, 486 and Pentium are so much different. They have different stages and number of stages, so their per stage propagated delay is different and their frequency can't be the same.
Is the 30% lower power consumption = total power consumption, or just dynamic power consumption?
Also, you're forgetting that P_dyn = CfV^2, where C is just some constant.
http://arstechnica.com/articles/paed...p4andg4e.ars/2
Intel's explanation of this diagram and the history it illustrates is enlightening, as it shows where their design priorities were.
Figure 2 shows that the 286, Intel386 ™, Intel486 ™ and Pentium ® (P5)processors had similar pipeline depths – they would run at similar clock rates if they were all implemented on the same silicon process technology. They all have a similar number of gates of logic per clock cycle. The P6 microarchitecture lengthened the processor pipelines, allowing fewer gates of logic per pipeline stage, which delivered significantly higher frequency and performance. The P6 microarchitecture approximately doubled the number of pipeline stages compared to the earlier processors and was able to achieve about a 1.5 times higher frequency on the same process technology.
Fast computers breed slow, lazy programmers
The price of reliability is the pursuit of the utmost simplicity. It is a price which the very rich find most hard to pay.
http://www.lighterra.com/papers/modernmicroprocessors/
Modern Ram, makes an old overclocker miss BH-5 and the fun it was
I really think they can''t. 45nm from intel don't use immersion, and 4ghz penryn have desapered from intel roadmap.
45nm frequency are lower as expected.
Intel don't want let their core duo higher than 3ghz, because core 2 duo is gonna 'die.' Core2 can go 3.6 i'm pretty sure of this.
Intel don't want c2d get better score than c2q. Tha'ts why intel will not release 3ghz+ c2d.
Intel can't run 45nm penryn Quad higher 3.33ghz, if it was true, intel should have showed us.
QX6800 is already beyond the limit of a normal CPU. Too much power wasted.
If you could go 3.6ghz with a kenfield easy with stock volt, i would agree that, but it's impossible.
lol gojdo got owned by nn_step once again!
Its just the dynamic power consumption, which is more than 90% of total power consumption.
The transistors of the 45nm Intel process are needing 30% less energy (30% lower power consumption) for transistor switching, compared to their 65nm process.
The static power(the current leakage) on the 45nm is reduced more than 5 times in the drain-source and more than 10 times in the gate oxide, compared to Intel's 65nm at same frequency. At 20% higher frequency the 45nm CPU will have roughly same(a little beat lower) leakage than the 65nm. Intel can either reduce the leakage or increase the frequency for 20%.
C is the cumulative capacitance which is not a constant, but a variable. Its value depends of the number of transistors involved in a certain situation. The more the operations(more stages working) are being performed at once, the more transistors are involved and are forced to switch, thus to consume more energy.Also, you're forgetting that P_dyn = CfV^2, where C is just some constant.
If we have two same clocked CPUs with same architecture, working at same voltage, but the one built on Intel's 45nm and the other on Intel's 65nm, for the same task the dynamic component of the total power consumption 45nm CPU will consume 30% less energy to accomplish the task.
The 45nm Intel CPUs have lower Vcore than the 65nm, which will significantly reduce the power consumption further.
@nn_step
Again, you are unable to understand the quotes you are posting and you are unable to support your claims. The 286 is similar to 386, the 386 is similar to 486 and the 486 is similar to Pentium, but the 286 is very different compared to the Pentium. I can't figure out what the author of the article was thinking, but the number of stages is not the only factor on which depends the frequency. The 286 has 134000, the 386 has 275000, the 486 has 1200000 and the Pentium(1st gen) has 3100000 transistors. If what the author of the article said was true, then why the Intel 80286 and the Intel 80386, which were produced on exactly the same 1500nm process are clocked differently? The 80286 frequencies are 6-25Mhz, while the 80386 are 16-40MHz. The same story is about the 486 and the 1st gen Pentium, both are made on the same 800nm process. The 80486 has frequencies of 25-100Mhz, while the 1st gen Pentium has 60-66MHz.
The author I was linking to was directly quoting intel http://www.intel.com/technology/itj/.../pdf/art_2.pdf Page 3.
Some how I would believe Intel if they say that their processors would run at the same speed if they were on the same process but then again I don't expect you to believe Intel about Intel processors. Since heaven knows, how could Intel know anything about Intel processors
Fast computers breed slow, lazy programmers
The price of reliability is the pursuit of the utmost simplicity. It is a price which the very rich find most hard to pay.
http://www.lighterra.com/papers/modernmicroprocessors/
Modern Ram, makes an old overclocker miss BH-5 and the fun it was
@nn_step
OK, I think we are missing the point from the article. The article talks about the relative frequency, which is not an exact number. The graph is the relative theoretical possible frequency by Intel various uArch designs pipeline depth, when there are no pipeline stalls.
Says the one that claimed K10 NDA breaks today. lolNever count on a penryn @3.6ghz this year
http://processorfinder.intel.com/det...px?sSpec=SL9UK
http://techreport.com/reviews/2007q2...power-peak.gif
Last edited by red; 05-14-2007 at 02:11 PM.
http://dailytech.com/DailyTech+Diges...rticle7277.htm
Anyone read this yet?
Currently, static power is much higher than that.
On the limits of leakage power reduction in caches, Yan Meng; Sherwood, T.; Kastner, R., in HPCA 2005.
You quoted the extremes for the leakage, assuming either constant switching speed or constant leakage. You don't get both. It's either one or the other, or a mix.The transistors of the 45nm Intel process are needing 30% less energy (30% lower power consumption) for transistor switching, compared to their 65nm process.
The static power(the current leakage) on the 45nm is reduced more than 5 times in the drain-source and more than 10 times in the gate oxide, compared to Intel's 65nm at same frequency. At 20% higher frequency the 45nm CPU will have roughly same(a little beat lower) leakage than the 65nm. Intel can either reduce the leakage or increase the frequency for 20%.
Proof:
I meant that C is constant for a given design on a given process. You were trying to talk about at 3.6GHz Penryn scaling to 4GHz at 65W, so you need to take this into account.C is the cumulative capacitance which is not a constant, but a variable.
----
Also, I hear from certain people that the current Penryns floating around even have trouble doing Conroe speeds. Hope this gets fixed soon, and hope that it was just a bad CPU.
Last edited by Shadowmage; 05-14-2007 at 06:10 PM.
http://download.intel.com/pressroom/...0107_FINAL.pdf
*~30% reduction in transistor switching power
*>20%improvement in transistor switching speed or >5x reduction in source-drain leakage power
That implies 30% reduction in transistor switching power throughtout.. Now what is the difference between "transistor switching power" and "source-drain leakage power" as it relates to the real world?
The 30% switching power is dynamic power and the leakage power is static power.
Looks like old info. The busses have gone up since.
Also the product launch is in 2007.
Crunching for Comrades and the Common good of the People.
Posting Permissions
Reply With Quote
Bookmarks