Slashdot Mirror
Intel's Anti-Overclocking Technology Simplified
John Thorensen writes "Found a fantastic article on Intel's recent Anti-Overclocking patent at Fastsilicon.com. Worth the read, as it also explains some of the technical and ethical issues of overclocking. Good to see that some tech journalists can still write material understandable by an average person."
i keep getting 404 error when trying to read comments... what a great idea. force people to RTFA before making comments!
Beware - soon we will find people who sell overclocking devices going to jail for violating DMCA.
(yes, I forgot my password here.. again lol)
-Honestman
If this is not about control then they should give a way to turn it off. Otherwise, they are blowing smokes up our butts.
Seriously, wtf? Ethics should be something applied to lawyers, doctors, mechanics, etc.. not something that should be brought up when a kid (or adult i guess) is tweeking his hardware. The fact that overclocking voids any warranty should be enough of a precaution by manufacturers.
What ethical issues are there relating to overclocking? Overclocking a chip, and selling it at a higher speed is already called "fraud". There's nothing ethical or unethical about overclocking. Is redlining your car's engine unethical? Stupid maybe, but that's about it...
I want to delete my account but Slashdot doesn't allow it.
Writing material which is readable to the average tech-interested
layperson is easy. Doing that while avoiding insipidity and
simplification to the point of being misleading.... ummm....
priceless?
-I like my women like I like my tea: green-
This technique is nothing more than embedding an oscillator on-chip, and using that to monitor the main clock.
Since most CPUs internally multiply their clock (you don't feed a 3.0 GHz P4 a 3.0 GHz clock, you feed it a much slower clock and it multiplies it up), why then don't manufacturer's just use an embedded clock and do away with all this?
Simple - it is very hard to have an accurate clock embedded in the CPU. External clocks can use a quartz crystal to vibrate and make the clock - an embedded oscillator would have to use an on-chip delay line or RC network, which will drift over time, temperature, and voltage.
So all they can do with a system like this is catch you if you are overclocking by a fairly large amount - were they to try to trap you at a 10% overclock they would have false trips due to process variation.
To extend the analogy the article used: you will get a speeding ticket if you are going 20 over the speed limit. Keep it less than 10 over and you will be fine.
NOTE: this is not advice condoning overclocking or speeding! This is just an analysis of the technology involved.
www.eFax.com are spammers
Overclocking a CPU you purchased is no different than putting a turbo on your 4-cylinder and making it go faster than the v6.
"Comedy's a dead art form. Now tragedy, that's funny."
I think it's worth it to point out that the article stated that there are ethical issues only when a vendor sells a comptuer that is overclocked without alerting the consumer.
If this has been a problem, I agree with Intel that it's important to restrict overclocking to protect, not limit, the consumer.
If vendors are only rarely (or never) overclocking a CPU and selling it for for more then I think that while it's probably not a wise business decision by Intel to implement such a technology just to limit consumers, it is Intel's right as the manufacturer and there is nothing ethically wrong with it. There is still competition and the market will speak for itself.
No one is forcing you to buy Intel products after all.
That's the component that keeps breaking. Intel does not give one nuclear frog fuck about the life of their processors. They care about one and only one thing - the money they can extort out of you on the next upgrade.
All chips are baked to a manufacturing tolerance that allows them to run at any given speed. Each new batch is tested and if more than some number predictably run at a given speed then that is what they are rated. As their manufacturing process imporves with each turn of the Deeming crank then the rated speed goes up. But when you push more power through the chip to make it run faster it superceeds it's own manufacturing tolerances. It would be like putting a 767 in transonic dive. It might hold together but Boeing thinks that's pretty much your issue.
But Intel doesn't make airplanes they make CPUs and their revenue comes from locking you into THEIR upgrade path. Break that relationship and they will hose you.
Imagine that, YOU are paying the embedded costs for them to find a clever way to stop YOU from speeding up YOUR chip.
....is to buy an AMD. For the money you save you can go buy other goodies for your machine.
It is their property first, though, and if they want to add a mechanism to make overclocking difficult, if not impossible, then it is their right. It is your problem if you still buy it. If they came to your house and installed it on a motherboard that you already owned then you would have a complaint. Otherwise you are just a knee-jerking troll.
I was actually thinking about building a P4 box for my next main machine and looking forward to the quiet whirr of a stock Intel heatsink. After 3 years of Delta Fans on Athys, I thought a P4 was a great idea for silence and overclocking. And what does Intel do? They bend all the OCers over and kick us squarely in the nuts.
I have an older technology that fits nicely alongside Intels anti-overclocking technology, it's proprietary and only works with geeks, OCers and effects all systems we build, its called anti-Intel-purchasing technology and I suggest we all use it religiously.
what happened to owning things? I mean if i buy a car and slap a 454 bigblock (or what ever it is the gearheads at teh auto shop are alwasy drolling over some this year swimsuit model) in there, the car manufacturer does come knocking down my door, although i supposed that would akin to replacing a the processor witha faster one, but still if i tweak the motor the car came with men in black suits still do not show up.
I understadn the problem of scum sucking bottom feeders buying cheaper chips and overclocking them and then selling them at clock rates higher than they actually are. But look there will always be people out to screw you as long as money is the most sought after object, then there will always be people trying to make a buck either honestly or dishonestly. I'm sure everyone has heard of the used car dealer selling cars as having all orignial parts when the car does not. Is it up tot he car manufacturer to make sure their parts cannot be tweaked, or modified? hell no! it's up to the comsumers to make sure what they think they are buying is what they are actually buying. Stop trying to pass the blame, and take responsibility for yourselves. Manufacturers should not take the blame for other's wrong doings, nor should they inhibit the actions of the hobbists that support them.
but then again, that's just my opinion, i could be wrong.
The money and expense it takes to over clock a cpu usually costs more than just buying a more expensive processor. There isn't much difference from a 2.5 Ghz and 3 Ghz processor technically, but the price reduction in return for the minimal performance is huge!
Spend money working on other bottlenecks, such as more ram, a better graphics card and faster hard drives.
The anti overclocking mechanisms are there to stop people from accidentaly setting the wrong settings in the bios and therefore voiding their warranty
For a while now, Asus has had the bad habit of tweaking their FSB slightly out-of-spec, for example a 135 MHz FSB instead of 133. Although only a slight overclock, this can easily lead to speed increase of 100 MHz on the (currently) high-spec processors. If Intel were to tweak their overclock-detection to such a point that it can even detect minor increases, I'll be curious what happens to Asus...
People replying to my sig annoy me. That's why I change it all the time.
... An over-clock deterrent mechanism of a chipset which comprises an over-clock detection circuit for detecting over-clocking of a system ... and an over-clock prevention (thwarting) circuit.
Okay, to implement this, they're including a reference clock on the chip, which means that processors of different (marketed) speeds will have to be made with a different process (which has maybe been true for a long time, but I was lead to believe that, eg in the pIII days, the wafers that failed 1Ghz just got sold as 833MHz, etc).
So instead of doing all these calculations to decide if you're "speeding," and then doing even *more* calculations to penalize you, why don't they just expose this reference clock speed in a special interrupt call? And maybe even the relation to the operating speed (eg, "you are overclocked by 10%")? Then, they could release an app that would tell you how fast your computer was SUPPOSED to be, and how fast it IS.
Then, OC'ers could have their cake, and no one else could be taken advantage of by unscrupulous OEMs who overclock to bump up their margins. I concede the point that "most average people will never check anyway," but just having the information *available* should protect Intel from liability, which seems to be the essential idea. Plus, the threat of having the practice exposed at any time should stop at least some of the overclock-resellers.
I get sick and tired of all these hardware nerdz acting like they're electrical engineers (which I happen to be). The out of a flop, thru a logic cloud and into the next flop is determined by 1) the output resistance of each logic gate, and 2) the capacitance (load) that is driven. The output resistance of a fet goes up when temperature goes up. This is why cooling your processor allows it run faster. However, there is a limit. No matter how much you cool your proc, it'll never go to 0 resistance. The amount of heat produced by your proc doesn't determine the speed bin it went into. The "fastness" of the batch of wafers that die was cut from determines the speed bin. Fet threshold voltages drift from wafer to wafer (and die to die within a wafer). Higher thresholds = slower fets. Heat is proportional to Capacitance*Voltage^2*Frequency. Note that fet resistance isn't in that equation.
Because then the unscrupulous would just hack that message out of the BIOS.
...phil
"For a list of the ways which technology has failed to improve our quality of life, press 3."
If they can make the clock internal to the CPU, why do they bother with the external clock at all?
Many may ask at this point, "If the processor does more work with higher clock speeds, why are there limitations on the clock speed - why can't one run a processor as fast as they want?". Although there are many factors that contribute to the answer to this question, the basic answer to this is heat. With every clock pulse, electricity flows through the processor. Because of resistance in the processor's pathways (think of it as a sort of electrical friction), some of this energy is converted to heat, similar to what happens when you rub your hands together very quickly. The higher the clock speed, the more often the clock pulses come, which means that more heat is generated at higher clock speeds. Because processors don't react well to the effects of this heat, testing is done to determine the maximum clock speed that they can run at safely.
That is not really accurate. While it is true that power and clock speed are approxmately linearly related (double the clock speed, double the heat output), the way the article explains the max speed is wrong. This implies that if you took a 2ghz P4 and clocked it at 2.4ghz, it would run hotter than a "real" 2.4ghz P4. This is not the case. All P4s will put out the same amount of heat at a given clock speed.
The actual reason that chips clock at different speeds has to do with precision of manufacture. I'm not really a car person, but I would imagine that better quality parts would let an engine go faster. If a spark plug has a problem, you might get misfires at higher RPMs (?). When a CPU is made, sometimes some of the wires are too thin, and because of the higher resistance it takes more time for enough charge to flow through the wire to get a 1 to change to a 0 (or vice versa). Now, you cannot clock it as fast or the CPU will produce erroneous results.
Another possible defect would be two wires ending up too close to each other. The faster a wire changes voltage, the more interference it creates in wires nearby. With the two wires closer than expected, they might start to experience "crosstalk", where the signal on one of the wires is affected by the other wire. At lower speeds, crosstalk is less of a problem.
There are many more things that cause variations in the max stable speeds of processors, but I won't go into them.
You might next ask, "What about the 'perfect' chips? Why can't they go faster?". The answer to that question is that even the best transistors can only switch so fast, and an electrical signal can only travel so far in a given period of time. When you're working with frequencies in the GHz, light can travel no more than a few feet, and the speed of electricity in wires is much lower.
The processors are then labeled with this clock speed, and they go out the door with a designation such as, "Pentium 4 - 2.4GHz". In this particular case, Intel has tested the processor and has determined that to run properly, it needs a clock that runs no faster than 2.4 billion times per second
The reason you can overclock is that Intel's tests are brutal. If they sell a processor as 2GHz and someone builds a computer with poor case ventilation and a cheap heatsink and low quality power supply in the sahara desert, the computer needs to be stable. Processors can run faster at lower temperatures (there are some equations describing the effects of temperature on various parts and generally higher temperature slows things down), so in a properly ventilated case with a good heatsink (and reliable power supply), the processor can operate reliably at higher-than-rated speeds.
It is important to note that just increasing the clock speed won't have as drastic of an effect on processor lifetime as many people say. What WILL have serious effects, though, is increasing the voltage. Why do overclockers like to raise the core voltage? More voltage means more current and stronger signals. In the thin wire scenario above, more voltage and more current means that even with the higher resistance,
My server
The suggested alternative was to place a speed identification in the processor . . . Intel balked at this . . . because it didn't address their real concern.
I agree so far.
Their real concern is limiting you to a set level of performance that you pay for.
Au contraire: their real concern is crooked system builders scratching the "2 GHz" off the chip, scribbling in "2.4 GHz", slapping a big heat sink on, and pocketing the difference, leaving Joe Sixpack none the wiser. Joe Sixpack won't know to run a special utility they have to go to Intel to get just to check if their clock speed is legit.
It seems at first look that a better solution might be to have the BIOS display a warning and emit a great cacophonous PC speaker screech during startup if the input clock speed is greater than the CPU's specified speed, but big or clever crooked system builders will just patch the BIOS to remove that. A wonderful thing, flash memory...
This automatic checking thing has possibilities. I think they should also check the Vcore to make sure that's not too far out of line either, lest a misconfigured or broken motherboard fry the chip (*cough*K6-2*hack*Amptron*wheeze*)-- it would also further their anti-overclocking concerns since one way to make a balky CPU deal with overclocking is to raise the Vcore (CMOS switches faster when driven harder, but you quadratically raise the heat output... better have a truly mighty cooling system if you try it).
In any case, I've had enough systems get balky running at their specified speed that I really don't want to push the clock any more than necessary.
...when you're writing a game...tweak the difficulty of "Easy" to something [your mother] can cope with. -- onion2k
Preventing overclocking is just corportate bs
They have every right to limit how their product is used, just has you have every right to not buy it.
First off, I can't see why overclocking seems to work Intel's panties into such a froth. Overclocking a processor is no different then 'hot rodding' a car! Hobbyists take delight in getting those few extra horsepower out of something, whether it be a CPU or a Hemi. Bragging rights are also involved. Yet you don't see the car companies patenting devices that inhibit an engine's horsepower output. They see it as a tribute to their engineering designs that people can do this- and rightly so. Now comes along Intel -the spoiler- who pulls a hissy fit everytime someone even MENTIONS overclocking! Yes, I agree that remarking chips and selling them is wrong, but there are laws in place to deal with this. They're called fraud laws and they've worked quite well for decades! Frankly, I think that AMD has a MUCH more reasonable approach to overclocking...they make it possible -yet very obvious visually- when someone overclocks their CPU's. GROW UP INTEL...take the fact that you make great products that hobbyists love to: 'push the enevlope on' as a source of pride, instead of having a tantrum over it!
Okay, that's easy to fix. Just divide the CPU's clock by four and feed it to a clock driver for the rest of the board.
Of course. Once they sell it, however, it is no longer their product.
Do you really believe that a product you've purchased should be under the control of the guy who sold it to you? Maybe your car should limit you to the speed limit of the state you bought it in. Maybe women's underwear should have a license forbidding men from wearing them. Maybe when you buy fresh meat it can come with a contract forbidding you from freezing it "to preserve freshness." Maybe Microsoft Press programming books can come with a license prohibiting you from using the knowledge to create competing products.
Maybe everything should be licensed and nothing sold. Maybe every "manufacturer" should tell you everything you shouldn't do with their product and then warn you in the warranty that they're claiming "no fitness for a particular use or purpose."
Maybe when your car is leased, all your consumer products are licensed, your food is consumed on the spot at restaurants and your clothes are bought on credit you will really be free. You will be living in the very model of freedom for all the world to see. God bless America.
TW
We will sell you one 2 Ghz processor. But it will only be activated @ 500 MHz. You want it to run faster? mmmmmm well pal, you'll have to pay us $50. How it will be activated? send us the money, we'll send you a password. Oh naughty boy!, don't try to get this password at those ugly hacker sites. The password is wired to a serial number inside the processor. What, you want 2 Ghz? sorry, the $50 I was talking about was for 1 Ghz. For 2 GHz is $80. Oh, by the way, that is just the 1 yr fee. At the end of the period your processor will go back to 500 Mhz. But don't worry, you know where to reach us! Think about it this way: we love so much to increase the speed of your CPU that we will do it regularly!! In a way, we are liberating your CPU. (I love that word, you can use it for almost anything these days). Thanks for shopping Intel!
Step 1: Disable overclocking
Step 2: ???
Step 3: PROFIT!