Core 2 Duo thermal throttles
Remember the old days when Intel rated Netburst CPU at 3.73GHZ but in reality it often runs at half the speed due the thermal throttling? Same is seemingly happening with Core 2 Duo within normal operating parameters. Basically, you get a Core 2 Duo rated at 2.93GHZ, when it is not doing anything stressful and the room temperature is cold, it runs merry at that speed, most of the circuits are turned off anyway. However, once you add some load, the Intel chip slows down to save itself from meltdown.
BTW, EU may fine Intel $3.5 billion for breaking the law.
posted by Sharikou, Ph. D. @ 7/15/2007 06:50:00 PM
44 Comments:
Hmm, then how does Core win all the benchmarks if it's throttled down?
AMD is such a loser processor that even a throttled processor can hand AMD its ass........
My other system has the E6600 overclocked to 3Ghz from 2.4Ghz (set the FSB from 1066 to 1333) and it runs fine on the stock Intel fan under full load.
This comment has been removed by the author.
But of course IBM power6 is running at 4.7ghz around the clock apparently SOI technology is superior to intels sprained silicon.
IBM does indeed make good CPUs. Everyone has to admit that IBM makes good CPUs for the HPC PCs. These POWER CPUs aren't designed for regular lower end 2P/4P servers or desktop/mobile.
IBM's CPUs show that IBM's 65nm SOI process works quite well. It's AMD's broken CPUs and broken fabs that are proof that K10 can't scale past a pathetic 2Ghz.
Intel has quad core 3GHz CPUs shipping today. AMD is still stuck at dual core. Anything faster from AMD is just vaporwarre. AMD must use three or four good wafers just to get one K10 that works at 2Ghz.
Even K10 is for a pathetic low end only up to 8P. Tulsa now goes to 32P. Tigerton will also go to 32P. The Itanium 2 dual core does up to 512P. Nehalem will also do 512P:
http://www.vr-zone.com/?i=5113
AMD is for low end crap only.
People use real CPUs from Intel, Sun or IBM when they want real work done.
AMD BK Q2'08.
Penryn is coming and will totally destroy AMD:
http://www.vr-zone.com/?i=5116
AMD BK Q2'08.
ONeexpert said
"Luckily AMD runs SOI so there should not be any problem with running the k10 up to 10 ghz."
That sort of comment reminds me of Netburst claims from Intel .. the magic 10GHz makes a re-appearance! Lets see 2GHz before talking about 10GHz.
I agree with Scientia's comments, I think that Intel were being wisely cautious with the top speeds when considering the different environments around the world and possibly substandard PSU's, cases, etc etc. They probably could have got 3.33GHz out the door but did not want to blight their new star with possible bad press ( not that it stops some areas from giving it :) )
K10 should flush out Intel's hand though on speed so it will be interesting to watch. They are already showing benches on 3.33Ghz whereas they did not on Conroe so it seems to indicate their 45nm process is doing OK. Maybe the cooling is not the standard though?
My own X6800 conroe does 3.66Ghz with 1.45v and Tuniq tower and I hit 55-60C, never used the standard hsf but you have to assume I would be more comfortable nearer to 3GHz.
throttling @ stock... i dont think so.
Im running E6600 @3,2ghz with 1,4V (which is a way out of stock) and with Intels TAT, doing 100% cpu workload on both cores, it reaches 71°C.
Yes thats quite much, but also room temp. is 35-36°C right now. (i hate summer :p )
Still i dont get any thermal throttling under this conditions.
This comment has been removed by the author.
Some odd results there. I had a quick look at Tom's and found the following review (as I posted in Scientia's blog)
They test a x6800 against a fx-62 and it looks like they used stock cooling on both parts.
"Under heavy loads, the overclocked Core 2 Extreme reads 151° F (66° C), and power consumption registers 95 watts. Don't forget that these readings occur because the clock rate was boosted from its stock value of 2.93 GHz to an impressive 3.66 GHz. In this situation, the stock cooler fan ran at a rotational speed of 2700 RPM, easy to notice at higher clock rates."
Looking at the various benchmarks I cannot see any throttling when comparing the o/c x6800, x6800 and e6700.
While we don't know the room temp that Tom's run the tests at, wouldn't you expect a o/c x6800 running at 3.66 to throttle ?
As Tom's did not seem to experience the same problem as Digit-Lifes did someone get lucky on the quality of their proc or dodgy/badly applied HSF.
p.s. deleted final comment as it doesn't apply here
AMD, in more detail, admits they are heavily in debt and are running out of cash and may not be able to borrow more cash:
As of December 31, 2006 we had consolidated debt of approximately $3.8 billion. In addition, a significant portion of our consolidated
debt bears a variable interest rate, which increases our exposure to interest rate fluctuations. Our substantial indebtedness may:
• make it difficult for us to satisfy our financial obligations, including making scheduled principal and interest payments;
• limit our ability to borrow additional funds for working capital, capital expenditures, acquisitions and general corporate and other
purposes;
• limit our ability to use our cash flow or obtain additional financing for future working capital, capital expenditures, acquisitions or
other general corporate purposes;
• require us to use a substantial portion of our cash flow from operations to make debt service payments;
• limit our flexibility to plan for, or react to, changes in our business and industry;
• place us at a competitive disadvantage compared to our less leveraged competitors; and
We may not be able to generate sufficient cash to service our debt obligations.
• increase our vulnerability to the impact of adverse economic and industry conditions.
Our ability to make payments on and to refinance our debt, or our guarantees of other parties’ debts, will depend on our financial and
operating performance, which may fluctuate significantly from quarter to quarter,
AMD admits that they may very well run out of cash and cease to be competitive against Intel and Nvidia:
We cannot be certain that our substantial investments in research and development will lead to timely improvements in product
designs or technology used to manufacture our products or that we will have sufficient resources to invest in the level of research and
development that is required to remain competitive.
AMD also successfully predicted what would happen to their creditability if they delivered parts that were late or underperforming (R600, K10):
If we are delayed in developing or qualifying new products or technologies, such as what occurred with the multiple delays
in the launch of our R600 GPU for the high-end category of the PC market, we may lose credibility and our competitors may be able to take
advantage of these delays by launching higher performing products before we do, which could cause us to lose market share and force us to
discount the selling price of our products.
All this from AMD's Form 10-K.
You keep mentioning VIA CPUs onemoron, Intel had better keep an eye on VIA. VIA is much more of a serioues threat than AMD.
AMD is a joke. 65nm? Way behind Intel. DDR2 support? Way behind Intel. Dual core mobile CPU? Way behind Intel. Direct X 10 GPUs? Way behind Nvidia. Quad core CPUs? Well, there still missing. At this rate I wonder if AMD will ever get Barcelona above 2Ghz. They had the same MHz problem with K8. The only reason they ever got K8 out was because of IBM's assistance.
AMD's a pathetically small and weak company. Worth only a meagre $7.5bn vs. $150bn for Intel. Even Nvidia is worth more than twice as much as AMD
AMD BK Q2'08.
Intel announces 3Ghz Quad core CPU for desktops: http://www.tgdaily.com/content/view/32909/118/
As well as faster Core 2 Duo models. AMD 6000+ is left competing with the E6550 2.33Ghz. The above $150 market is 100% Intel. AMD has nothing to sell for over $150 except for the pathetic 4x4 that gets fragged all over by Intel's quad core CPUs.
AMD BK Q2'08.
New 3Ghz Quad Core and Dual core benchmarked: http://www.xbitlabs.com/articles/cpu/display/core2extreme-qx6850_2.html#sect0
This is blazing fast performance. A clean kill for Intel. The only benchmark AMD won in was Science Mark! Intel won all other tests.
Sharikou
Since you linked to my blog you might want to mention that the Digit Life article is from 2006. My article was to refute the notion that Intel was capable of delivering chips faster than 3.0Ghz in 2006 as they originally stated.
This does not prove that C2D is unable to clock above 3.0Ghz today. I can't say that without more recent test information which I don't have.
slack
As I mentioned on my blog, none of the benchmarks run on the THG review would thermally load the core. Also, running the processor just under a temperature that causes throttling would still damage the CPU; the lack of throttling does not mean that this is a safe temperature. Most OC'ers are aware and accept that OC'ing shortens the life of the chip.
I would like to point at xbitlabs.com and their preview on the QX6850.
They were so kind to offer power consumption numbers (something that all the AMD paid sites (so nearly every site there is) of course wouldnt do) and it clearly shows the 6000+ consuming more power than the QX6850.
So now its official: Intel processors consume less than half of AMD processors when put under high stress (i.e. high clock speeds). Clear proof of how far behind AMD technology is.
ycon
6000+ is still 90nm. If Intel's 65nm can't beat AMD's 90nm then that would certainly be a problem for Intel. The fact that AMD does not yet have its fastest speed in 65nm is of course a problem for AMD.
That's quite a story, Enrique. I like stories. I like stories about pinatas. In fact, I like everything you have to say.
VR-zone have some interesting Intel information on 45nm processors today. Server is pencilled in at 3.16GHz rather than 3.33 with desktop problably 3.3GHz.
Not sure if this is a general announcement but if it is a leak then it does come in the same week as the Q2 earnings releases.
Intel is doing 0.5x multiplicators.
Death trap for Barcelona (the CPU of course, the city is nice *lol*)
As I mentioned on my blog, none of the benchmarks run on the THG review would thermally load the core.
The flip side to this, which you don't seem to acknowledge, is that no real workload is as heavy as TAT. TAT is not real code; it is code designed to make maximal use of the processor and make it as hot as possible. That isn't in any sense realistic, but it's useful for designing cooling systems (as it means you can easily generate a worst case thermal loading).
Of course, the reality is that Intel can hit high clockspeeds without exotic cooling, and AMD can't. And new steppings are reducing the power consumption and increasing the temperature envelope anyway.
If the X6800 is throttling at 80F !@%#$%#!, then that is clearly a problem of installation of the setup. Anyone that concludes otherwise is an idiot.
Sharikou said from previous post: Intel is still alive for one single reason: AMD's complex 65nm SOI process is not producing the clockspeeds as expected.
OH NOoOooOOo! does this mean Intel won't BK in Q2'08? Are you slowly taking back your famous prediction? Say it ain't so!!!
aeidyrun: a course of action when encountering something really scary.
i'm just amazed ... all along, scientia was playing around the pro doing the overclocking doesn't mean Intel can produce those chip at the said frequency ... and some were telling him that frequency is not the real issue, but the TDP that intel is willing to put. now he talked like he just discover a new land ...
3.5 billion in USD fine against Intel for anti-trust violations as ruled by an EU investigation.
This is a followup from the EU's raids of Intel's offices in '05.
link
This comment has been removed by the author.
oneexpert
However when you compare the energy saving qx6850 energy saving model to AMDs x2 6000 clearly you can see AMD is a power sucking loser.
According to Xbit Labs the energy saving qx6850 draws 209 watts idle and 337 watts loaded with its platform.
According to xbit the AMD x2 6000+ power sucker cpu and platform take 213 watts idle and 385 watts loaded.
Clearly the AMDs x2 6000 is another AMD space heater.
Intel cpu and platform qx6850=209 watts idle 337 watts loaded.
AMDs x2 6000 cpu and platform=213 watts idle and 385 watts loaded.
When Intel lables a cpu a energy saver it actually is, when AMD says they have built a energy saver its fud and spinola like all AMD's previous claims.
Stop wasting energy buy Intel and save our planet.
This comment has been removed by the author.
This comment has been removed by the author.
Oneexpert: Nobody except you EVER labeled Intel's highest-end Quad Core part an 'energy saver'. And when you figure in the fact that AMD's real competitor to the Q6850 is the gas-guzzling SUV of systems: The 4x4 it's pretty obvious who is more efficient.
You also seem to forget what efficiency actually means. Time for some math, Sharikou & Oneexpert, please overdose on the cyanide you use in the Kool-Aid while I put on some facts:
What the hell is efficiency? Well, when talking about efficiency, it basically means how much work can get done for a given expenditure of energy. Now, in the physical world work can be measured in simple ways like, how much energy did it take to pull a block with a mass of 100kg up a distance of 10m assuming the Earth's gravity was pulling against it? There is a theoretical minimum amount of energy to push that block up 10m, but (due to thermodynamics) no real-world system will ever quite meet that goal no matter how well designed it is. The efficiency in this case is measured by looking at how much useful work you got done, vs. the amount of actual energy you expended to do it.
In the computing world, the actual 'work' being done is more abstract, but we generally consider a set amount of processing to represent a unit of 'work' to measure. For example, rendering a scene in PovRay would be considered the unit of 'work' and the amount of measured electricity from the wall would be the 'energy' it took to perform that work.
So the next question is, how to watts and energy relate to each other? Well, watts are a unit of power. It is important to remember that power is NOT energy! Instead, power that is expended over a period of time makes up energy. Easy example: Look at any power (bad name) bill you get from the electric company. They don't charge you in 'watts' they charge you in Kilowatt-hours (Kw/h). Each Kw/h is just the energy that is expended by using 1 Kw of power for the time period of 1 hour. If I use 10 kilowatts for 6 minutes.... I have used exactly the same as 1 kilowatt for 1 whole hour! The power is different, but the energy used is the same.
If you still don't believe me that energy is what you should care about, remember: the Electric company charges you for energy usage, not power!
So let's take the stuff above and apply it to some CPU examples. Here's a question: If my Q6850 uses 125Watts to get a job done in 30 seconds is that more or less efficient than a BE 2300 that takes 45 watts of power to do the same job in 90 seconds?
30 sec * 125 Watts = 3750 Watt/secs (a nonstandard but still accurate energy unit)
90 sec * 45 Watts = 4050 Watt/secs
So... while the giant Quad Core Intel part might use more power... it is actually more energy efficient since it needed less energy to complete the task.
The above example is somewhat contrived, although there's no reason to think that a high-end quad core Intel part couldn't be 3 times faster than a low-end AMD part... that's quite accurate.
Some of the better real benchmarks I've seen on power and energy are over at TechReport. This one right here is excellent in that it shows how many joules (an actual standard measurement of energy) it takes to complete the benchmark (bottom of the page).
The results are very interesting, since it turns out that the most energy efficient chips (at least for multithreaded apps) are actually Intel's Quadcores, beating out even the BE chips that oneexpert can't stop having convulsions over.
OK oneexpert, you can start posting the same tired old crap again and again now, the rest of us (hell even Sharikou is sick of you) will go back to ignoring you.
P.S. --> I neglected to talk about 'real' power, 'reactive' power, 'apparent' power, the power factor, inductive, and capacitative loads, etc. These are all factors, but power meters used in most reviews are based on kill-a-watts that ignore the reactive components in electric systems. These should not be ignored though, look for power supplies with high power factors (near 1) that can get the phases to match as closely as possible.
Oh one thing... in my previous post when I say '3750 watt/sec' I don't mean watts divided by seconds. I probably should have written it: (watt)(secs) to show that the units are multiplied.
Is it just me or does oneexpert look and sound more like a tool with every post?
It's not just you. Comparing a dual 2.1Ghz and a Quad 3Ghz in terms of power efficiency is just crazy. The QX6850 is still far more efficient than 6000+ or the 4x4.
In a oneexpert style posting: LV Dual core C2D mobile CPU is more energy efficient than the 3Ghz 4x4 setup. Who would have thought that?!
This comment has been removed by the author.
How does the qx6850 compare to 2 FX-74 which is AMD's equivalent part on power use?
You cannot say buy AMD save the plane when FX-74 is still around. Maybe when K10 is out.
For Penryn looks like there is a quad core at 2.66GHz and 50w for K10 to compete against.
oneexpert said...
"HOT HARDWARE
total power consumption
qx6850 idle=189 watts loaded=291 watts
AMD 6000 idle=158 watts loaded=246 watts
Well even in high power cpus and platforms looks like intel qx6850 sucks more power than AMD x2 6000."
Let's see if I'm reading this correctly. You are compared a quad core (4) cpu against a dual core (2) cpu. You seem to have some real issues with comparing similar systems.
You should be comparing the 6000+ against a x6800 (from the same review).
AMD 6000 idle=158W loaded=246W
Intel x6800 idle=135W loaded=180W
2.66Ghz 50W Harpertown server CPU is very impressive. That's TDP from 120W down to 50W, coupled with the IPC gain from the new Penryn core.
Two of them in a 2P server and you've one hell of a monster in performance with a TDP of just 50W each.
Before that, just in time to greet K10, is the new LV 2Ghz Clovertown with a 50W TDP.
Don't believe onemoron and his lies.
http://xbitlabs.com/images/cpu/core2extreme-qx6850/burn.png
http://techreport.com/reviews/2007q2/core2duo-e6750/cine-power-peak.gif
AMD is going to lose more than $500m:
http://www.fudzilla.com/index.php?option=com_content&task=view&id=1960&Itemid=1
AMD needs more cash badly:
http://www.fudzilla.com/index.php?option=com_content&task=view&id=1959&Itemid=1
AMD might get into serious trouble as the company is again going to post a larger than expected loss and that if it doesn't raise more cash, it might actually go out of business as soon as the end of the year.
Onemoron,
You never answered my question in the last thread.
How much money will it's superior products make AMD this quater, or this year?
oneexpert
Anandtech's Low Power Server Shootout is a pretty good comparison for servers. It puts an HE 2218 (2.6Ghz) up against a Xeon 5140 (2.33 Ghz). Both are dual socket and use 8 1GB DIMMS. At least in terms of ASAP on these dual socket servers, AMD ends up with a small performance/watt lead. Intel's CPU draws less power but then the Northbridge and FBDIMM's draw more.
Now, ASAP is a good benchmark for servers but it doesn't really apply to desktop systems. Intel comes out ahead on a desktop performance/watt comparisons because it does better on typical desktop benchmarks.
Things should improve for AMD once K10 comes out since K10 has better power saving features as well as better performance than K8. However, Intel will soon follow with 45nm chips. So, we'll be revisiting this in couple of months and then again in Q4.
Given Intel's problems with FBDIMM I would guess that they will try to avoid dual socket desktop systems while presumably AMD would encourage them. Basically, in Q4 Intel will want to frame the comparison for dual socket systems (like V8) in terms of total power (which it should win) while AMD will want to frame the comparison in terms of performance/watt (which it could win). However, Intel should be happy to do performance/watt comparisons on single socket systems.
Intel's CPU draws less power
Intel's CPU draws less power
Intel's CPU draws less power
Intel's CPU draws less power
Intel's CPU draws less power
Intel's CPU draws less power
Clear enough oneexpert you simplton fuckwit?
As the big stars fashionable vane is always dress tide in the spring of 2011, louis vuitton handbags imitation, they will have the bright beautiful fashionable dress collocation how single product? Small make up take you to go to see star street snap, lists the most fashionable handbags this season's new list. Lv new handbag series and accessories designer Katie hillier cooperation has been completed and product six money at the end of this month, will land shelves, but the overall product has a good omen. Earlier on 2.net - a - ahead of porter UK auction, bid on your site for cowhide and grind a nap cortical ku shoulder bag for chain, another kind of bag, two paragraphs experienced alligator handbags in one hour will be fully subscribed. LV marcie "China red" edition of the red hot flashes, a LiuWen, actor GaoYuanYuan by supermodel, and the street pats burn all floridians Chen personally demonstration. Louis vuitton spring 2011 hand bag bag, surely new series a year let you red. Inspiration: the 1970s hippie style, simple lines, harmonious style, withstand time trial. Material: bags adopts advanced small cowhide leather, harmony by weaving effect of arc adornment; Details: leather woven winding handle, convenient portable and shoulder ku. The city of "lv" celebrity hot holds degree continues unabated. Madonna, dita von teese, Greene, Eva herzigova poses isla fisher, capotondi cristiana, alessandra ambrosio, kat dennings,,, mariacarla boscono waiting for movie star every time out, Louis vuitton saves lv bag as boon, wheezed. The 2011 summer series with white as the theme, the lv series series. Represent purity and frank white, become the new inspiration LV classic handbags. Material: soft leather, leopard pony fur, white fabrics and weave cane knits bring new idea for series. Details: Sicily lace and dazzling aureate full of Mediterranean baroque amorous feelings, also is small in Sicily to the characteristics.Louis Vuitton Shoulder Bags Modelling like small handbag equipped with long leather lining gold aglet.
Bag in use: if hand to sweat,coach bags outlet 2012 the best with the arm as far as possible or shoulder bag bag, and less with hand bag, or in the bag to sweat stains left. In the bag for sharp objects inside, must will sharp objects, complete packaged put again. While those with slit, damaged bag too items had better not put in baby bag. Don't forcedly or put overweight items, or you will make the deformation or damaged. Bag The lock of the bag in the switch it, 10 million don't in the lock,Coach Bags outlet with foreign body that this might damage the hardware that bag. Don't bags in sticking labels or adhesive tape, so it's easy to cause the skin when off spalling. Try to let vanity avoid and jeans clothes frequent friction, because will be very easy to happen dyeing phenomenon. The rain and snow, try to avoid using turn fur, USES, suede leather, the cortex, because once large bags with water can be hard to do, and cortex will harden. Long time point-blank light and heat,Coach Bags can make the bag bag decoloring, color and shape. Should try to avoid with sunshine, heating bag long time contact. When receiving: an outlet to collect bags, best receive before cleaned and nursing, must be in bags into a clean filler (for example: shredding group or cotton unlined upper garment), to keep the shape of the bag, and then will leather bag dust bag to put collection. Receive the cabinet of the bag must maintain good ventilation, it is better to have the shutters cloakroom or chest, Coach Pursesbest in the cabinet don't put too much items.
Post a Comment
<< Home