Slashdot Mirror
Intel Fights Overclockers
Alejo sent us a link to a zdnet article about Intel
going after overclockers.
Not kids in the basement, but commercial vendors selling
systems with overclocked chips sold to unknowing customers.
But they've got software now to detect if you're overclocking.
Neato.
I know I once tried to overclock one of my chips, and found out it was a remark (it was already oc). So I like this approach...that way I can at least oc my chip on my own...and not be limited by someone making a buck or a few hundred as the case may be. When I oc my 300 celerons to 450 and pII 400 to 496 or 533, I want to get what I expect...
good show on intels part...now if we can just figure out how to burn out that darn psn thing for sure.
Intel isn't stopping people from overclocking. There's just allowing people to know the speed that their chip was intended to run at so they don't get ripped off by retailers who overclock (and overprice) their chips.
It's also an excuse to keep the processor ID enabled.
I know we used to when I worked at a pc shop on request only. We took 75Mhz P5, and overclocked them to 90+ all day long...Something about the overclocked chips made them run fast as hell. We continued that all the way up, amd and cyrix (though cyrix was a little iffy) and so on...and I guess they still do today.
The software only works on the new PIII "Big Brother Inside" processors...
> I don't know for sure, but I
> don't think an OEM could legally overclock a > chip, sell it at the price of the slower-speed > chip, and TELL you
> about the overclocking. Thoughts?
I think it would be irresponsible to sell an over-clocked chip to someone who doesn't understand what they're getting into. And if they DO understand what they're getting into, they can probably overclock it themselves, right? So what would be the point? As far as just a theoretical "can they do it?" Who knows? It wouldn't be false advertising, but...
Sure it does.
I remember hearing of a nefarious retailer that was selling P120 processors that were marked as Cyrix parts. When asked why the BIOS reports the processor as a P120, he told the "customer" that that was the way a BIOS recognised a Cyrix 150-MHz processor, it runs at 120-MHz, but performs like a 150-MHz processor.
Because of the confusion behind the P-rating stuff at that time, this worked on more than a few customers.
The scanning software puts information in the hands of the user. How is this a bad thing? It lets people know whether or not they're being scammed. The only reason I can think of as to why one would want to disable that software is if they were going to be selling overclocked computers and not telling the person they were selling it to.
I hope this isn't the reason why you're against this, and if it isn't, please tell us what is, because right now you sound like you sound rather guilty.
the article wasn't even about the PSN.
Intel marks the speed on the outside of some chips, the problem is that the ink can be removed by washing it off with a cocktail of soap and solvent. Then the remarker can print another higher speed on the chip, and sell it for a profit.
Then when you, the kid in the basement, buys a cheap system with the idea of overclocking it, its already overclocked. So you've just spent your hard earned money on a machine which is already maxxed out.
I still think they could mark the outside with a deep laser etching, but that takes time and money (lasers and electricity)
Ummm... ZDNet said a few days ago that Intel told them that the P2 had the ID chip too. The P3 is not the only "Big Brother Inside" to worry about but it's probably too late (or just a ploy to make people think it's too late anyway). TTFN & Shalom.
-PipTigger
But the processor ID could be used to prove you bought the cpu from a certain vendor.
Heh, just wait in the future you will not be buying the chip, but buying a liscence to run that chip, and only at its intended speed.
I gotta hand it to intel using this sweet propaganda to win ya all over.
Yeah, I'm with you on this one. There even is a website on overclocking chips! bxboards.com
You're talking about two different things... you are talking about the fabrication (which, if i understand correctly, has more to do with fab equipment manufacturers than intel themselves). He is (i assume) talking about architectural advances and making the processor do more work on each clock cycle.
Oh, you meant overclocking? Nevermind....
Go ahead Intel, do whatever you damn well please.
I'll wait till X-mas 2001 and then build me a CRAZY ass Intel free system...
I want a Kryotech case, a Dual k7 MB, Dual k7' so/c'd for all their worth, 256 MB RDRAM, 20(ish) Gigs of HD in the new standard bein develoes,
and I will be happy to write Intel a thank you note for converting me = )
Cha0s The0ry
Don't forget most intel cpus end up in corporate pc's, and the majority of the rest in home pc's owned by non-nerds. Overall only a tiny minority of intel cpus will be owned by people who've even heard of overclocking, yet alone want to do it or know how to.
Does Intel really make 5.5ns L2 cache? That doesn't seem economical. It would make more sense to build only 4.5ns cache, making the tolerances low (this saves money), and then selling it as either a 400 or 450 MHz chip depending on how fast it runs when tested.
Well, that's what I'd do if I were Intel, anyway...
Posted by DarkYoda:
Intel should go after them...
as long as they only get the venders
and leave the us alone
Posted by RainForest:
I think the title is wrong. Intel is going after remarkers, and not users who knowingly over-clock their CPUs. Intel is releasing a software that will tell you what the CPU is rated at, to alert the user if he/she has bought a remark.
Posted by Pseudonet:
What would be really nice would be for Intel to use this technology to track illegal overclocking as an alternative to frequency and clock locks which prevent us true oc devotes forom getting as much as we can from these chips.
However, the way it looks at the moment (which we have seen in the revision A celerons) Intel will use this technology to further sercure the settings of the chips, so that they can mass produce one chip capable of all speeds, and then lock them to a specific one, hence cutting production costs.
Posted by EasySleeze:
:)
jeez!! "Big brother inside" had been around for a while. Connected to microsoft.com the other day cause I needed a patch foe Winblows. Got a cable modem, and I watched as the server tried to scan my machine
With this kind of thing, perhaps they could now remove the multiplier lock and allow hackers to play around with the multiplier to their hearts' content.
I don't think an OEM could legally overclock a chip,
Sure they could. After all, they gave the customer exactly what they said they would. If they don't want returns and unhappy customers, they will be SURE to explain the risks, and do a good long burn-in to make sure it's working reasonably well.
The reason remarking is illegal has nothing to do with selling a shoddy product. It's the fraud that gets them in legal trouble.
With MIPS, Sparc, Alpha, PPC and others out there, you can take a long walk off a short pier. Not to mention of course, AMD and Cyrix and um, (cough) Transmeta.
My last computer wasn't an Intel-based one and neither will my next one. :P
Welcome. :)
Anybody notice that the only links embedded in the article were stock tickers? If I were a buyer of high speed chips (I've only got a 233 MHz PII), I'd be more interested in downloading the clock testing program than in Intel's stock price.
But then again, I'm not a technology executive...
Suuure. Sorry intel, I heard that one already. Wasn't that the justification for the anti-OC'ing BX chipset? That said, can some enterprising individual come up with a way to block Intel's overclocker-scanning software?
--
As long as each individual is facing the TV tube alone, formal freedom poses no threat to privilege.
"Reactionaries must be deprived of the right to voice their opinions; only the people have that right." - Mao
Hey, man, I've got the BH6/450A combo and I'm 34 and live on the 4th floor! ;)
unless you put a clock in the processor... but taking the clock out of the processor is what the bus speed and multiplier were invented for in the first place, so that would make no sense at all.
)O(
the Gods have a sense of humor,
Never underestimate the power of stupidity
To err is human, to moo bovine
Offtopic, but yes you are very correct about PowerPC being very overclockable. As soon as I feel my new G3/300 has been "burned in", say in 3 months, I'll be attempting 350 then 400MHz with it.
With PPC, your success at overclocking is more likely to depend on the quality/tolerances of your cache RAM than what the processor might be capable of. Overclocked PPC's DO NOT overheat when clocked up 1 or 2 steps.
MacIntouch did a Review of an aftermarket G3 upgrade, using the new IBM 450 MHz chips. The user clocked it up to 560 MHz with no stability problems, and it ran at a relatively cool 35C/95F. The story is here:
http://www1.macintouch.com/g3zif466.html
To me, this is like tuning your car. No laws PREVENT a crime, but there are laws on the books against fraud. If a parts shop sells you a motor that is dangerously overbored and misrepresents it as a higher volume engine.. that is plain fraud, and auto consumers shouldn't all have to pay extra so they build in some kind of protection against overboring.
It's just life. People think they are saving money when buying cheep pee-cees from fly-by-night show vendors who won't be around half as long as their promised warranty.
You'll end up paying more anyways, as Intel had to draw resources from somewhere to pay for this.
As for privacy, the Net is like TV, and like it or not WILL replace TV. We have enough censorship here in the States without a big brother [the government OR Microsoft... take your pick] indexing what we watch, don't want, and studying what was on the screen the moment we changed programming.
After penalizing the hobbyist overclocker for so long (with hardware-based locks), Intel has apparently finally purchased themselves a clue.
Now they can crack down on these scum who sell low-end systems overclocked and remarked as high-end. And the hobbyist can still clock his system into the stratosphere if he likes.
Who says gigantic, Big Brother-like corporations are COMPLETELY clueless?
Chas - The one, the only.
THANK GOD!!!
Chas - The one, the only.
THANK GOD!!!
Right now, I'm very happy to have my cool, speedy, modern PowerPC 604e. Another nice touch - it's smaller than a brick.
Is it me, or Intel taking lessons from Microsoft when it comes to mediocracy? First their attempt to pry into their customers' privacy, then the yawn inspiring performance of the Pentium III. Now, this.
And before someone says that I'm not understanding it correctly, let me explain myself.
1. Intel has taken great strides to stop overclocking, even taking away the ability to overclock your own machine (who knows, maybe your P-III will dial up Intel and tell on you?).
2. There are already laws against selling something different than what you're advertising. Just ask Ralph Nader. Intel is serving itself here, not customers.
Seems to me that the PowerPC world isn't such a comparitively 'closed' platform as many assume - it just has fewer active players. I've never heard about IBM/Motorola caring one bit about overclockers.
- Darchmare
- Axis Mutatis, http://www.axismutatis.net
- Jeff
no he's not. it's dark inside the box
---
I've spoken before with an Intel person about overclocking.
What he essentially said was, Intel have no problem at all with home users overclocking, so long as they know the risks they're taking. However, they had a problem with manufacturers actually altering the chip case so that they were a clock speed higher, and then selling them on at that price.
The risks, he said, were simply that of instability. Intel, quite rightly, doesn't want its reputation for reliability - and I don't think people would disagree that they have one - tarnished by people misleading the buyers.
It's not a "Nobody must overclock" thing, it's more a simple case of fraud.
Anyways, must get back to working out how to oil-cool my PC, using Peltier effect heat pumps.
Chips in the same family are already pretty much the same. CPU's with same specifications but for clock speed are already the same chip. The difference is that some chips pass more stringent QA procedures @ higher speeds and so they are 'marked' as running @ that speed given the assumption that some number of chips out of the oven passed "x" test. Chips that don't are tested less rigorously until they do pass @ some lower speed. You see it's not a technology or chip design problem - its a manufacturing process & QA problem associated with being able to bake-off "z" numbers of chips from a die of a given size and passing a given QA threshold. There's nothing inherently difficult with producing an Intel chip or any CPU that can run at any speed you wish if you can cool it off. The problem is being able to manufacture enough of them reliably & economically on the same die without a very small number of rejects.
So far Intel made a good deal by not marking their CPUs with their intended speed. With all the CPU faking going on, they probably managed to sell quite a few more chips - one middle speed CPU to the professional fakers and one high speed CPU to the end customers whose overclocked CPU didn't work properly.
Until now, if the chip did work most of the time (with all the people using Windows, a little instability goes unnoticed), there was no way of telling whether you really got what you paid for when you bought a P-II. Only details e.g. on the L2 cache implementation could be used to distinguish real from faked 300MHz CPUs.
That would be one of the simpler fakes. Real repackaging is also done, not only overwriting. These are so convincing that even official Intel distributors reportedly sold fakes without knowing.
"My intent here is not to advocate fraudulent sales from remarking"
:-)
Don't worry, the real remarkers are doing batches of thousands and sell them through many nations to be untraceable. This is not about some home remarkers
Intel couldn't go after the "kids in the basement," even if they wanted to, (which they don't.) Overclock it, underclock it... if someone found a way to make a 286 function at 2Thz, intel would say more power to 'em. The vendors selling overclocked chips as faster chips, though, should be penalized. I don't know for sure, but I don't think an OEM could legally overclock a chip, sell it at the price of the slower-speed chip, and TELL you about the overclocking. Thoughts?
-- Minds are like parachutes... they work best when open.
I expected someone to totally misread Intel's intent and fly off the handle with righteous prattle ... thanks for not disappointing me.
I've finally had it: until slashdot gets article moderation, I am not coming back.
...According to the article, at least.
Unlikely, IMO. They still want anyone running at 450 instead of 400 to pay for a 450 (that is still a very expensive step, around here at least).
Why exactly they even did this much is hazy, but I suspect that it's so that unscrupulous manufacturers are forced to buy the higher speed chips from Intel instead of remarking. It might also be to try to gather public support for bus locking ("see? we caught all of these people gouging you for overclocked chips! let us stop them"). I know exactly how well that would fly here, but bear in mind that most PC buyers don't overclock and so might like this pitch.
My system at home is a P100. It works adequately. Ditto the PII systems at work.
Small denominator and relatively small numerator, sorry. Alternatively, "the solution to a binomial with small coefficients", but that's not terribly relevant.
If you put a purely transistor-based oscillator in the CPU, it won't be terribly accurate (i.e. useless for core clocking) but would catch people overclocking 66 MHz busses to 100 MHz. Or, you could put a crystal in the chip module (even the old socketed "chips" are modules) and verify the bus speed to within a few percent or better. This would be very cheap and would stop bus overclocking.
The reason why you'd still use the bus clock to generate your internal clock is synchronization. The core clock and the bus clock remain in lockstep under the present system, with clock edges precisely converging every (n) clocks. Using the bus-locking crystal to generate the core clock would cause the two clocks to drift with respect to each other, making data transfer troublesome. You could add synchronization circuitry to keep the clocks in step, but if you're doing this much you might as well just multiply the bus clock instead, as that would take less silicon.
Remarking a CPU, or decieving the customer is bad, wrong, and worse. But culling fast chips from the herd, and selling them as overclocked is not only beautiful, good for the environment, and wise.
You the customer may choose to take your place in time, or defy Intels laser inscribed number,
Go-forward, move-ahead, PLL-in-hand. (Devo!)
So just cause "the man" says the CPU is a 300 MHZ part (period) dosent EVER mean it wont go 301,415926 MHZ. The customer may find it zipps along at 449,000001 MHZ reliably, at all tempratures, under all system conditions. Or a vendor may sell this as value added by the retailer.
It's funny how NORMAL people accept XTAL speed detents as the homogenious comodities they are.
If you dont know how to play between the quantums, then leave the experts who do alone.
Joe Torre - X - HardwareEngineer @ Amiga Inc & ZapMedia Amiga, AmigaDE, BeOS, Linuxz, QNX, Rebol, Windoze, ZME: So
Part of Intel's motivation is to keep the end user from over clocking as well. The 400 and 450 are the same exact cpu. So why buy a 450 when you can get a 400 and run it at 450? Intel doesn't want to lose any of its profits.
-matt
I don't know about the false advertising. They could easily say "PentiumII 400 running at 450mhz". That's 100% true. If someone is ordering a large quantity they might want the systems to come pre-overclocked. I personally think that the stores should be able to do anything they want as long as they are 100% honest and upfront about it.
-matt
The serial# really only acomplishes one thing. It allows software to be locked to a specific CPU. Something that has been done in high end unix markets for years. It is 100% useless and quite dangerous for online sales. This means A) anyone w/ physical access to your computer becomes you. B) if you upgrade your CPU you no longer exist. C) Your identity can be easily spoofed. Now while spoofing the id on the system would be difficult, it could easily be intercepted by a firewall and then rewritten to something else. This is a very dangerous situation.
Also, I really would like to believe that Intel is not against home users over-clocking. I won't believe this until they remove the multiplier lock. Since intel has software that can detect overclocking they have proved that it is possible to check in software. Why not just add something to the bios that checks the speed the system is running at, and then displays the real mhz. It could even display a large warning if the cpu is not running at the rated mhz. It's really not a difficult situation to remedy.
It also makes economic sense for Intel to block overclocking. If I can buy a 400mhz cpu and run it at 450mhz why would I buy the 450mhz cpu?
I wish I could believe that Intel had nothing but honorable intentions but do to the facts as I see them I can't believe this.
-matt
I was sold a remarked PII/400 through the channel and the CPU was not stable. I think it was a 300. The guy who sold it to me took it back and acted surprised. The screening job on the chip was good but not perfect so it was possible he had been had. He replaced it with a boxed PII/400 for no additional charge. (I did email Intel with his name/number cuz I was pissed - wasted a day figuring out why the machine was crashing...)
Adults are obsolete children. - Dr. Seuss
My intent here is not to advocate fraudulent sales from remarking, but couldn't the marking come off with acetone?
I say that because I used this acrylic "glop" (best term for it) as a liquid gasket to repair a leaky fuel pump on a small engine. I got some of the glop on my hands. Water, kerosene or alcohol wouldn't dissolve it. Acetone did. Stung my hands pretty good too.
Gleepy the Hen. More intelligent than the average hen.
I run a computer store. (yes linux, boxen go out my door)a year and a half ago, the normal channel suppliers didn't have a pentium mmx at the speed I needed for a machine. So, I went to a gray market supplier. (not an authorized intel supplier) The fools sent me a remarked chip, which luckily, the motherboard identified correctly. (not mmx) At the time intel had on their web site a program called cupid, cpu id program. I ran it and sure enough this wasn't what they had sold me, nor was the stepping number consistant with the speed that was marked on the bottom chip. These turkeys didn't want at first to replace the chip with the correct one, (they said that intel had made a mistake in marking the chip)until I told them that Intel wanted their company name. Needless to say I don't use them as a supplier any more. I really can't afford the grief that this kind of a supplier can cause my shop. Overclocking as something to ship out of the shop. Well, I shown users who seem to have a pretty good handle on technology, some of the web sites devoted to overclocking, but shipping one, nope, voids warranty on the machine. Overclocking isn't necessarly a bad idea, but if the user isn't aware what the risks and benifits are, shouldn't be done. My kid however has always had an overclocked system, and thinks its the only way to go. His system btw is completely stable, fast and can be left on always if you want, though he doesn't like the leds glowing at night (its in his bedroom) so he generally shuts it off at night.
Then it shouldn't matter, because you're not supposed to be able to overclock a PIII...from the mouths of Intel employees who came to talk to us here at KU...
It's the fact that they make them so well that means you can overclock them.
john
Soon, someone will post 'Intel reveals 2Ghz Pentium-V chip with built in quantum-processing & biological intelligence matrix' and the general slashdot response will be:
- Intel is just trying to force the software industry to make bulkier software so that we have to buy their stuff!
- Intel is imprisoning an artificial life form in their chip and should be shot!
--
The P-III serial# is an *EXCELLENT* idea, they just marketed it wrong (as others have said, this is absolutely not something new.. just new to Intel's x86 stuff).. besides, if you are going to use an online service, they have the right to ask you who you are. It's up to you whether or not you want to tell them or not... The chip-ID could be protected as vigorously as your address book. (never mind that this isn't how it'll be used anyway)
What intel just said makes *GOOD* sense. They don't want vendors fraudulently selling high-end machines that aren't, that may fail. They don't want the bad publicity of people saying 'your fast chips aren't reliable'.
They said they are *NOT* trying to stop people from overclocking chips, just to stop vendors from selling overclocked chips without notifying the customer.
"Although remarking is a crime, overclocking isn't," Johnson said.
Wow. They actually admit this. Maybe they do have a clue of whats going on..
"..since almost every remarked processor is overclocked."
Personally I have never heard of a remarked processor that wasn't overclocked. What else is there, remarking a processor to its same clock? Remarking it so that the processor is underclocked? Or maybe remarking it so that its a n AMD chip or a Cyrix chip? None of these other remarkings make sense.
-?-
-?-
intel sucks, but companies shouldnt sell overclocked computers. unless they charge the customers for a slower machine. i guess in this case intel is actually doing something smart.
-- toolie
Cryo-something, I can't remember the name. They make computers with really, really good cooling systems and overclock the processor. Their customers know exactly what they are getting.
There seems to be a misconception among a lot of people posting in response to this one that there
is no difference between the PII 400 and the PII 450. The processor core of these things IS the same. The differences between them are the level 2 cache and the casing. The specification for the 400 states that it comes with 5.5ns cache and the 450 comes with 4.5ns cache. The casing on the 450 also has two aluminium contact plates that are cast into the chasis behind the cache that keep it cooler.
That is the specification anyway, but, as anyone seriously into overclocking knows sometimes intel stick faster cache in slower CPUs because that's all they happen to have lying around at the time. So you'll find runs of 400s and even 300s with the 4.5ns chips in them. So with either a bit of luck or lot of research into manufacture dates and serial numbers you can end up with what amounts to a radically underated CPU.