AMD Stops Overclockers Dream Motherboard

nerdusa writes: "The Overclocking Community got a decidedly unwelcome jolt with news today on THG that shipping versions of the TBird and Duron will be clock locked and that the Asus A7V is shipping without multiplier unlocking switches. United Overclockers, which had been eulogizing AMD for its recent poliicies, is "saddened" at having to eat its words. Overclockers are used to having their dreams dashed by reality, but this is a particularly cruel blow."

Comments

groan..

[The Rock1699]

But I WANT to blow up my processor...no fair.

Quite a shame

[Animol]

It's really sad to hear this indeed - where else will future stories about insane overclocking of poor, poor 386 chips to 700MHz come from? I mean, every quickie or two we find someone who feels a need to make their old 8088 into a dutch oven...

Re:Quite a shame

[Omega996]

imagine a beowulf cluster of those!

The question is

[Zachary Kessin]

How do you allow people to overclock their own systems, but prevent 2 bit computer stores from
overclocking systems and selling them as the faster system. I would not be happy to find out that that 800mhz system I just bought was an overclocked 500mhz system.

I don't think there is a way to do it.

The Cure of the ills of Democracy is more Democracy.

Re:The question is

[alleria]

CPUID and similar utilities. But you have to be saavy enough to know to use them!

Re:The question is

[DevTopics]

...how to find out if a system is overclocked or not? That's easy: Every CPU has a unique serial number (like the PIII). AMD publishs a database on their web site which shows which serial number runs with which clock speed. You can even create a little program that will read that serial, log on to the web site, and display the clock speed of your CPU vs. the clock speed of your system. Of course this rises a few privacy questions... but this could be overcome.

Re:The question is

[mcc]

Umm, i can think of a way to do it, but it isn't a technological solution.

You do the same thing to a 2bit computer store that overclocks machines and sells them as legit that you do to a 2bit clothing store that sells fake versace jackets as legit. You call in the police, and the police charge them with fraud and/or misleading business practices and/or deceptive advertising. I guess they could claim that since that 500 mhz AMD overclocked to 800 mhz really was running at 800 mhz, they were telling the truth even if AMD intended it to run at 500 mhz, but i seriously doubt that they could get away with it. Not that i know anything about law.

As for how you catch them.. well, i'm not so sure about that.

Still, this isn't a technological problem. It's a socioeconomic problem. Don't try to find technological solutions to law enforcement problems; technology can always be worked around by just adding more technology to the mix. Laws can't really be worked around, and if there are holes in the laws the laws can be changed. This is, of course, assuming that the republicans don't suddenly decide that they will filibuster and block all "bills requiring computer vendors to state openly whether they have overclocked or otherwise modified the material they are selling from the state it was intended by the manufacturer" unless the bill contains a passage banning second trimester abortions. -mcc (score, 0: gibberish)

Re:The question is

[tzanger]

How do you allow people to overclock their own systems, but prevent 2 bit computer stores from overclocking systems and selling them as the faster system. I would not be happy to find out that that 800mhz system I just bought was an overclocked 500mhz system.

Processor serial numbers.

Oops did I say something wrong?

There is no reason not to have a unique number assigned to every CPU manufactured. If companies lock their software to a CPU serial number that's their own perogative but I certainly wouldn't be buying software from them. A unique ID which could be fed into a CGI script at AMD.COM and return what it was desgined to run at would be the best by far.

Re:The question is

[alecto]

The problem is, if the serial number were in every CPU, the dominant vendor (i.e. Microsoft) would lock the software to a given CPU, and there wouldn't be a darned thing you could (legally) do about it if you need to use it. Seeking a competing vendor isn't a realistic option, because the competition will likely do the same thing.

Re:The question is

[tzanger]

the dominant vendor (i.e. Microsoft) would lock the software to a given CPU

I don't believe they would do such a thing. I am usually pretty cynical but I just don't see it happenning. Why hasn't Sun locked their software to a CPU? I believe the alpha processors have been serialized since day one.

You can already easily get the IDE/SCSI serial numbers and nobody is locking to it.

Re:The question is

[mong]

Go Here.

Fixed.

Mong.

* ...Student, Artist, Techie - Geek *

Re:The question is

[alecto]

Why hasn't Sun locked their software to a CPU? I believe the alpha processors have been serialized since day one.

I don't claim familiarity with the Sun platform, but the biggest thing that comes to mind is Solaris. Since to use Solaris, you need a Sun anyway, I could see why they wouldn't bother, Sun being the only source of Sun workstations. Also, Sun and Alpha processors are probably sold to a more technically literate population that is more likely to rebel against copy protection and seek alternatives. Microsoft captives don't seem, as a group, to have the technical ability (or motivation) to make those sorts of choices.

You can already easily get the IDE/SCSI serial numbers and nobody is locking to it.

Replacing a disk drive isn't nearly as big a deal as upgrading a CPU (realistically, this almost always involves a motherboard swap and replacement of other components, as well). It is also quite likely to happen in the course of warranty repair of mass market PCs. Also, given Microsoft's recent penchant for "recovery" CD's that arelocked to at least a single computer model, being able to rely on a CPU serial number would give them the hammer to enforce an intellectual property "right" they already believe they have--to keep the user from transferring the license they legitimately purchased to another PC upon removal of the software from the old.

Re:The question is

[tzanger]

Replacing a disk drive isn't nearly as big a deal as upgrading a CPU (realistically, this almost always involves a motherboard swap and replacement of other components, as well). It is also quite likely to happen in the course of warranty repair of mass market PCs.

Doesn't matter, in my mind... A serial number is already available on a component of a PC and vendors aren't locking to it, why would they lock to a CPU serial number? Overclockers fry CPUs, people move from Celeron to P2 or P3s... Personally the fact that the components are serialized is meaningless to me. If I sell, upgrade or move to another computer I should be able to migrate my software. Why should my audio CD only play on my one deck?

Also, given Microsoft's recent penchant for "recovery" CD's that arelocked to at least a single computer model, being able to rely on a CPU serial number would give them the hammer to enforce an intellectual property "right" they already believe they have--to keep the user from transferring the license they legitimately purchased to another PC upon removal of the software from the old.

This kind of bonehead move would be defeated in court long before it became a problem, just as most shrinkwrap licensing is just usless words, so would their argument that it's for a particular PC. I'm sure of it.

not entirely true

[matticus]

the fact is, the new t-bird athlons will be locked to those who don't know the secrets. some hardware web sites have guides to changing the multiplier on the chip by connecting different contacts on the ceramic surface with a conducting ink pen. this just means the multiplier lock will be more difficult to overcome. before anyone flames AMD for this, remember that Intel has been locking processors since the p2-350. AMD provided a way to change the multiplier on the classic athlon (pretty ingenious if you ask me) with the goldfinger device, and they'll probably come through again. give them some credit, enough anti-amd sentiment.

Re:not entirely true

[pozzy77]

Why dont the chip companies just put a little write once ram with the chip that would spout out what speed it is. That way you wouldnt be buying bogus chips and you could overclock to your hearts content. that is all I have to say

Re:not entirely true

[orbital3]

remember that Intel has been locking processors since the p2-350

Actually, my old P200MMX was clock locked... all I could do was raise the bus speed to 75 giving me a whopping 225mhz... whee!

Re:not entirely true

[Tower]

Too bad - my PMMX 200 wasn't locked, and it would run at 100MHz bus speed. Stable at 250, 266, and almost at 300 (for a few hours anyway). Those were the days... the 300a @ 450 isn't too much of a slouch, either...

Re:not entirely true

[rcw-work]

That's entirely feasible, however chips can't just 'spout out' what speed they are supposed to be running at. If you want that information printed at bootup, the BIOS has to do it.

BIOS's are flashable, and it's fairly easy to change a text string (like "533") to another one (like "800") with a hex editor.

Once you have the thing in protected mode running a known OS (something you have loaded yourself, lest the evil bios flashers of doom also edit a couple DLL's or kernel /proc code), then the BIOS is, in theory, out of the picture and cannot interfere with a protected mode API for accessing this PROM.

I'd imagine you could put a protected mode CPU identifier program like this on a floppy disk (after all, memtest86 compiles down to 18kb or so).

But, you can hardly expect the target market that AMD or Intel is attempting to defend to do that. So that's why AMD and Intel don't do that. Oh, and people don't buy motherboards with non-flashable BIOS's anymore.

bus

[Anonymous Coward]

well I wonder if they are clock locked, then can you change the bus? The bus is already 200mhz, so what, pump it up to 266mhz??

Re:bus

[-brazil-]

The question is whether you can find memory that works at such speeds.

Re:bus

[KickKat]

Yes, you can alter the bus, but the best I have seen is where the bus has been bumped up to ~220mhz. Anything above that and the system ecounters pesky little problems, like boot failure. I suppose altering the core voltage could help with this, but on the boards I have seen, the core voltage changer thingy doesn't have a great enough range to compensate effectively.

Bad or Good

[Kinlan]

Is this a bad thing or a good thing that they are stopping people from Overclocking there AMD's.

I mean it will mean extra income for AMD people who want the extra horsepower in the PC will have to pay for it, instead of getting it "for free".

But does the consumer want to be restricted to what they can and can't do with there product.

I mean we see this in lots of other products, like cars, (some) OS's. There philosphy is: "You wnat it better, then pay for it". so is it really that bad that AMD are doing this with there chips now?

feel free to give you remarks

Paul Kinlan

-

Re:Bad or Good

[Kinlan]

why do you say that?
I was just raising some points, that I thought might have been valid.
Sorry to hear you didn't think the same
-

Re:Bad or Good

[RobertAG]

I would imagine that 90% percent of people out there have no idea what overclocking means. They flip a switch, the box lights up and they then proceed to happily pound away at the keyboard.

As a person with a fair amount of hardware experience, I personally have no interest in overclocking my equipment. Since I also buy equipment for the entity I work for, I certainly would like to be certain that the equipment I receive is what I paid for. In the overall picture of AMD vs Intel, this is probably just an issue of quality control. The last thing AMD needs is bad publicity. Their target market isn't overclockers.

This isn't to say that I don't respect the overclocking community. I DO feel that once you buy something, its YOURS to do with as you wish. In AMD's case, they feel it is better to repond to the needs of the many over the needs of a few.

Re:Bad or Good

[mrfiddlehead]

Sorry, but this kind of thing is just plain mean spirited. I could buy a car and soup it up so that it would run much faster than a stock engine and the automakers don't really care. It's not like they're putting governors on their cars to limit their speed.

And overclocking is not without risks, and thus cost. I'll bet there are lots of overclockers out there with fried chip stories. Crisps?

Re:Bad or Good

[lucid]

no, it isn't. as others have pointed out, overclockers are a small minority. if a large number of less clued-in customers repeatedly buy remarked CPUs without knowing it, and some of them fail, it soil's AMD's image. around this area (NW Ohio) there used to be a big problem with k6-2's being dead on arrival, or unstable at their rated clock speeds (25-50%?), leading many to run to Intel with their wallets. i think AMD wants to get away from having a reputation like that, being the 'flaky' chip maker. if a few hundred overclockers have to run their boxes at the rated speed, or alternately buy Intel, for AMD to have a good reputation for making dependable products, so be it.

i don't think the comparision to the auto industry is valid, either. by overclocking the CPU, you're using a 'lesser' CPU to do the job of a better one, whereas with a car, you're actually changing lesser parts out for better ones, i.e. changing the stock cam and putting one in with bigger lobes, etc. and the new cam would be designed to perform better than the previous one needed to. besides, even in australia i don't think there's been a problem with people finding 12 horse Briggs & Strattons under the hood where a big V8 was supposed to be.

Re:Bad or Good

[Frank T. Lofaro Jr.]

It's not like they're putting governors on their cars to limit their speed.

Umm, yes they do. The engine will stop producing power (it stops getting gas) once it hits the max speed (often the highest number on the speedometer). You'll slow down until it cuts back in again (you'll still be going at an ungodly speed).

You can't blame AMD (not completely anyway)

[Trans]

The reason they're forced to do this is because of those "Resellers" who overclock the cpu and neglect to tell the customer. When an OCed processor bonks, it's bad for the customer and AMD. It's not bad for the asshole seller though, he's probally running under a new name by now.

I think overclocking is a damn cool thing, but I wish AMD had another way of dealing with remarked chips.

Re:You can't blame AMD (not completely anyway)

[Anonymous Coward]

You answered it there.

You're asking AMD to increase production costs for everyone, so that 2% of their market has more freedom to fiddle around and use their parts out of spec.

I'm afraid the world doesn't work that way.

Ten years from now, 'Overclockers' will be remembered for what they are: the late 1990's equivalent of the maroons who put linear amps on the output of their CB radios in the past.

You've really gotta come up with some better ways of spending time.

Re:You can't blame AMD (not completely anyway)

[joshsisk]

Couldn't resellers just not use a BIOS that made the call? Or alter the BIOS to list it incorrectly?

Re:You can't blame AMD (not completely anyway)

[scotch]

You've really gotta come up with some better ways of spending time.

... like chastising people for their hobbies on slashdot

Overclocking Denied

[Calimus]

Personaly I find this to be a very odd play for AMD. I guess I would have to ask if maybe the reason for it is that the chips arn't as stable as advertized? Or now that AMD has stolen the spot light a bit from Intel, are they picking up Intels bad habbits?

Either way I see it becomming a sore spot fast. I've been a loyal AMD consumer and overclocker but I guess I knew this would happen at some point. It would be too esay to make your own 1ghz chip with those chips and the new Asus board. Not to mention alot cheaper.

Re:Overclocking Denied

[Schnedt McWapt]

Ummm, yes, that's an interesting point you raise:

When AMD produces a batch of chips they test them all out. The marketing guy obviously wants to sell them for the most money they can get.

The Reliability guy wants them to last long enough that systems don't blow up all over the market, giving AMD chips a bad name. So, they're as stable as advertised. The problem comes into being when people use them in ways they're not advertized as being stable for.

It's 'miracle' enough that they're yielding chips as fast as they are these days. Consider that if there's a 'clever trick' that can be used to squeeze more out of the chips, AMD is already doing so.

Personally I am just tired of the local ma-and-pop hardware shops refusing to warranty a CPU chip after it leaves the shop. They're forced to do so, because otherwise they'd rapidly become the free spare parts bin for all the overclockers. I'm tired of a small segment of the market creating an excuse for the owners of the shops to just wash their hands of their responsiblity to stand behind what they sell. And I suspect I'm not alone in this.

I overclocked, by mistake, years ago when I set the AT-bus multiplier wrong on my 486 motherboard (back when 486 motherboards were $600 affairs). I ran the AT-bus at 12 MHz for a long time, getting much zippier performance from the video subsystem. It also inadvertantly caused the machine to be unstable under certain conditions. But I never pretended when I discovered my mistake that I was doing anything particularly clever.

Re:Overclocking Denied

[god_of_the_machine]

It's 'miracle' enough that they're yielding chips as fast as they are these days. Consider that if there's a 'clever trick' that can be used to squeeze more out of the chips, AMD is already doing so.

True... but not at the low end. I've read many an article telling how the low-end Slot-A Athlons (550-700 MHz) are really re-marked 750's, because they aren't much harder to make, and it's simpler to make them in large batches and mark 'em down as needed.

So while it's true that it's a 'miracle' that they can make the high-end clocks... the low end sometimes has a lot of room. That's why 1GHz t-bird only overclock to around 1.1GHz... and even that probably isn't altogether stable. Wherease a Slot-A 550 can be easily overclocked to 750Mhz, sometimes higher, depending on the batch.

-rt-

Re:Overclocking Denied

[cmstremi]

The "...ar[e]n't as stable as advertized[sic]..." argument could only fly if they actually advertised the chips overclocking ability as a feature. AMD has never boasted OC as a feature for consumers. I doubt their decision has anything to do with Intel. It's just a natural decision they were forced into because (probably) they get a bad rap when some poor schlep burns up his new computer. The consumer doesn't/shouldn't HAVE to know about overclocking. His first reaction to a fried AMD chip will be "AMD sucks. Their chips don't work." This hurts AMD.

thees ees boolsheet

[gtx]

what difference does it really make to AMD if people overclock or not? if they break the chip, they will have to buy a new one! Perhaps if there have been alot of people returning overclocked burnt out chips, but that's AMD's fault for not enforcing it's warranty terms well enough. some people will say that AMD wants people to have to pay for higher clock speeds. they will. sure, you can get 950 MHz from a 700 MHz duron, but how high do you think you could get a 1GHz T-bird up to? people will still pay for the higher clock speeds if that is what they want. AMD isn't making any sense

I support AMD on this one

[Tet]

There is a genuine problem with people reselling overclocked systems without the buyer knowing that they're getting a less reliable system. AMD have to protect their brand name, and having hordes of people claiming that AMD chips are unreliable because they've been sold overclocked systems is not a good way to stay in business long. Yes, it will prevent the hobbyist that knows the risks and accepts them. That said, most overclockers I've met don't fall into that category -- they tend to have the mentality "wow, I can make my machine go faster" without knowing how it works, and where the extra speed is coming from (hint: your safety margin before things start breaking).

Re:I support AMD on this one

[dagoalieman]

Everyone is bringing up a good point when they say that the resold overclocked systems are unstable/unreliable yada yada. I have a question slightly off topic:

Why doesn't AMD get some overclockable motherboards, and sell those under a contract? Say something to the effect (and in legaleese) "We understand these can be overclocked, and in fact, expect you too. By signing this, you're saying you will, and thus void the warranty- no more responsibility for us"? And when they sell these, sell ONE PER PERSON AT A TIME- avoids a person buying em up for his company.

doing something like that would allow overclocks to be sold, avoid tarnishing their reputation, and even allows those damn lawyers to be pleased.

I'm sure I begged for some flames in there. Give em to me.

Re:I support AMD on this one

[Tet]

Why doesn't AMD get some overclockable motherboards, and sell those under a contract? Say something to the effect (and in legaleese) "We understand these can be overclocked, and in fact, expect you too. By signing this, you're saying you will, and thus void the warranty- no more responsibility for us"?

Nice idea. But that would mean AMD getting into the motherboard business (which, AFAIK, they don't do now). They could come to some arrangement with an existing company (e.g., Asus, Abit, GVP etc.), but there would still be the problem of people using the chips on motherboards other than the ones with which they were sold. Yes, they could make a physically different connector for overclockable chips to prevent this, but the market size wouldn't justify it.

Re:I support AMD on this one

[Spasemunki]

The "sell one person at a time" really throws a wrench into things. For one thing, this means that AMD can't sell bulk anywhere. They have to handle direct, individual level sales, which is a significant administrative burden. Secondly, they're never going to make enough money on individual Mobo sales. Manufacturers make their money selling to retailers and wholesalers in bulk. They cut out admin and other costs. Turning any profit for AMD would be very unlikely, once they've gone to the trouble of developing a board (especially since that isn't their area) and building the infrastructure to sell and move the things.

"Sweet creeping zombie Jesus!"

Why do computer companies dislike enthusiasts?

[w00ly_mammoth]

This is a bit of a rant beyond the AMD thingie, so bear with me.

There's a trend running through the industry. When linux programmers write *free* drivers for new hardware, oftentimes the manufacturer is very reluctant to support the OS. Sometimes the developers have to struggle to wrench out closed specs to write the drivers, and still the corporation sees it fit to at best ignore it. Oh, and by the way, this results in more of their products being sold.

Enthusiasts are the forerunners to new technology. Generally they are the first to embrace it and forecast where the industry is going. At the very least, they provide valuable feedback. Yet for some reason, the history of the computer industry has seen established companies simply ignore enthusiasts. This goes for mainframe makers who ignored the PC, *nix vendors which dismissed linux as a toy, and MS which dissed the internet as a useless fad. It may also be that the music industry is on this track by opposing mp3 fans instead of seeing where they are headed.

As a larger trend, when companies which started out in the garage lose touch with their roots and ignore enthusiasts, it might mean they crumble under their own weight. But in specific cases, I simply fail to understand why companies don't support them. For instance, linux today has millions of users, and yet when I go to logitech's page to see if their cordless mouse works with linux, there is NO info on it at all. I have to dredge thru deja.com to see if anyone has posted it. Why? Does logitech not see the benefit of spending a few thousand $ to hire someone to update their web site with info about linux? Or even if drivers are needed, can't they hire a couple of guys to write them? Even if a small fraction of the linux base buys their mice, they have made a good profit.

What am I missing here?

w/m

Re:Why do computer companies dislike enthusiasts?

[dirk]

This is a bit of a rant beyond the AMD thingie, so bear with me.

There's a trend running through the industry. When linux programmers write *free* drivers for new hardware, oftentimes the manufacturer is very reluctant to support the OS. Sometimes the developers have to struggle to wrench out closed specs to write the drivers, and still the corporation sees it fit to at best ignore it. Oh, and by the way, this results in more of their products being sold.

Enthusiasts are the forerunners to new technology. Generally they are the first to embrace it and forecast where the industry is going. At the very least, they provide valuable feedback. Yet for some reason, the history of the computer industry has seen established companies simply ignore enthusiasts. This goes for mainframe makers who ignored the PC, *nix vendors which dismissed linux as a toy, and MS which dissed the internet as a useless fad. It may also be that the music industry is on this track by opposing mp3 fans instead of seeing where they are headed.

As a larger trend, when companies which started out in the garage lose touch with their roots and ignore enthusiasts, it might mean they crumble under their own weight. But in specific cases, I simply fail to understand why companies don't support them. For instance, linux today has millions of users, and yet when I go to logitech's page to see if their cordless mouse works with linux, there is NO info on it at all. I have to dredge thru deja.com to see if anyone has posted it. Why? Does logitech not see the benefit of spending a few thousand $ to hire someone to update their web site with info about linux? Or even if drivers are needed, can't they hire a couple of guys to write them? Even if a small fraction of the linux base buys their mice, they have made a good profit.

What am I missing here?

I agree that enthusiats really help decide where the market goes, but you have to remember that until something reaches the rest of the market, it is a very small niche product, that doesn't really count as far as revenue goes. Let's take the Logitech mouse example. Let's say there 500,000,000 computers in the US today. Of those, maybe 5% need a new mouse per year (and that's being REALLY generous). That's 25 million mice sold. Linux has maybe 5% (that's probably a little high, but it works)of the PC market, so that's 1,250,000 Linux users who need mice. OF these, maybe 40% will buy a Logitech mouse. That's 50,000 mice. And probably half of these people would be influenced by whether they are using drivers from Logitech, or ones that someone else has written and uploaded to FreshMeat. So that's 25,000 mice purchases that could be affect, out of 25 million mice (that's .1% of total mouse sales). If they make $5 in profit from each mouse, that is a total of $125,000 per year, which is nothing and probably just about what they would have spent to make the drivers.

People can talk about how companies support causes, but in the end, it all comes down to money. There is usually not any money to be made on the small enthusiasts market, until it is no longer an enthusiast's market.

And don't forget that the enthusiast market is wrong as often as it is right, so it's not really a sure gamble.

Re:Why do computer companies dislike enthusiasts?

[www]

Let's say there 500,000,000 computers in the US today.

What about other countries

Of those, maybe 5% need a new mouse per year (and that's being REALLY generous). That's 25 million mice sold.

Let's assume that every new computer needs a new mouse, I would assume that more than 25,000,000 new computers are sold each year

Linux has maybe 5% (that's probably a little high, but it works)of the PC market, so that's 1,250,000 Linux users who need mice. OF these, maybe 40% will buy a Logitech mouse. That's 50,000 mice.

1,250,000 * 0.4 = 500,000 Linux users who need new mice.

And probably half of these people would be influenced by whether they are using drivers from Logitech, or ones that someone else has written and uploaded to FreshMeat. So that's 25,000

250,000

mice purchases that could be affect, out of 25 million mice (that's .1%

1%

of total mouse sales). If they make $5 in profit from each mouse, that is a total of $125,000

$1,250,000

per year, which is nothing

nothing?!?

and probably just about what they would have spent to make the drivers.

not any more.

As well, they will sell mice more than one year, a new/different mouse model would also likely be able to borrow mouse driver code from the other drivers as they all use very similar interfaces.

Re:Why do computer companies dislike enthusiasts?

[inburito]

I know this is not the point of your post, but still..

500,000,000 computers in US? According to Census the population of US is approximately 275,214,186 as of today. This means that everyone would have about two computers (can you imagine a two year old baby having two computers?). Yup, big corporations have lots of computers and their employees have computers at home but still at least over a third of US households don't have a computer.

According to the information on that Census-site in 97 (yup, things change) less than half of US population had used a computer! Now we'd probably get more like 70% have used a computer and maybe 40-50% own one.. Have to wait for the latest census information..

Re:Why do computer companies dislike enthusiasts?

[Shadow99_1]

Well I think it's because to people who are mostly marketing (I mean look at most companies the most common occupation is something with selling what the other 25% of the company made) the 'fringe' enthusiasts don't represent their amrket. You can see thsi with how companies like Ati, 3dfx, & Nvidia have worked.

Ati in the early 3d days ruled the OEM systems that were seen as the 'mainstream' & so made more money than 3dfx by far. Now the moeny didn't really go back into R&D for Ati so obviously it went to pay all the marketing people who sold the stuff. On teh other hand 3dfx saw the profits weren't in the high performance enthusiast market and they created the banshee & V3 cards to try to steal OEM market share. Nvidia has kept up the image of being the top performer whiel still goign after OEM's, but even they release chips like the TNT2 M64 that fit the costs OEM's want in their components.

Now... As for linux & why it doesn't get what attention it should, just take a look a where I live in the eastern part of the US. In the last year within 100 miles of me (which includes Cleveland, Pittsburgh, & Buffalo) their has been 2 Unix jobs available (yes Unix not linux) & absolutely no linux jobs. Sure a few more places say they like it if you know linux, but they want you to be microsoft certified as a requirement. Until linux gets seen as a real soltuion in business, business will ignore it.

Re:Why do computer companies dislike enthusiasts?

[scotch]

I don't have any figures to back this up, but I would guess that 500,000,000 computers may be close. There are probably even more than that, but the number actually being used (as opposed to collecting dust in a garage) is undoubtedly less. I own 4, have another 4 at work that I exclusively use, so that's 8 (and only 1 is collecting dust). I'm sure the story is similar with other slasdotters/computer professionals/enthusiasts. So I make up for 7 two yr olds not having any.

As far as logitech goes, I always buy their mice because I find they have the best value for 3 button mice, which before the advent of the wheel mice were becoming hard to find, and are moreso now. Logitech should be aware that since X is still pretty much a 3-button mouse affair (I know wheel mice support is there, shut up already), they could easily capture most of the linux-new-mouse market with some good marketing.

Sick of 10? Try 11!!!!!

Re:Why do computer companies dislike enthusiasts?

[scotch]

Now... As for linux & why it doesn't get what attention it should, just take a look a where I live in the eastern part of the US. In the last year within 100 miles of me (which includes Cleveland, Pittsburgh, & Buffalo) their has been 2 Unix jobs available (yes Unix not linux) & absolutely no linux jobs.

You should move - last time I checked, there were tons of Unix related jobs out here in the West. Hell, my company would hire you (if you know your shit).

Sick of 11? Try 12!!!!

Tom's Hardware already has it beat.

[Saurentine]

For those of us who would like to continue overclocking our chips, Tom's Hardware already has a guide to set your silicon to whatever speed you prefer.

This link will show you how:
http://www.tomshardware.com/cpu/00q3/000711/inde x.html

Very simply. If it could work.

[Forge]

Very simply. You make your CPUs so they announce to the world at bootup

AMD-K7-Athelon-500MHz Running at 800MHz.

There are 2 reasons AMD and iNTEL don't and won't do that however.

Reason Number 1. It would mean having different CPU dies for each chip they sell. They aren't going to do that. They prefer to simply build a batch of chips and depending on how clean they come out you put a label on to claim a specific clock speed. Yes. Specific clock speeds are determined after the fact before labeling is done, not before.

Reason Number 2. Any Information the CPU issues about itself must go through the BIOS 1st. The problem is that someone with the resources of a 2 bit 10 box a day CPU manufacturer can arrange to have the BIOS altered so the quote above would say nice things like.

AMD-K7-Athelon-800MHz Running at 800MHz.

Of course being crummy and moronic corporations they just refuse to level with the customer and will continue to mislead you all as to what exactly they sell and why. Frankly, I think people would be nicer to them if they leveled with us.

PS: As for Locking the BIOS. fat FSCKing chance. We have grown accustomed to adding new features and fixing old bugs by flashing the BIOS. We aren't even going back to the old days of swapping chips, let alone having no BIOS flexibility at all.

Re:Very simply. If it could work.

[-brazil-]

Reason Number 1. It would mean having different CPU dies for each chip they sell. They aren't going to do that. They prefer to simply build a batch of chips and depending on how clean they come out you put a label on to claim a specific clock speed. Yes. Specific clock speeds are determined after the fact before labeling is done, not before.

And there's a reason for this. If you didn't do it, you'd have to throw away a lot more chips, and the OK ones would get about 5 times as expensive.

Re:Very simply. If it could work.

[dingbat_hp]

I've seen AUTOEXEC.BAT files with an
ECHO Really Fast 486 at 50MHz
in the first line (a few years back, when clock relabelling first started to be an issue for dodgy retailers). A few weeks ago I even saw a similar message in the boot.ini of an NT box ! Back in the days of Turbo buttons, some used to swear blind that the little LED displays on the front of the case (remember them ?) were an accurate measure of clock speed.

If you want to stop customers being ripped off by dubious retailers, you need to start with smarter, or better informed, customers. Telling them that clock speed doesn't matter(*) and that bundled WinModems are evil would be a good start. PC buyers are still naive and falling for bad deals that only look impressive in an advert.

(*) If it really matters to you, then it's your problem to find out how to check it. Don't be a label K1DD13 who thinks it's K00L just because it has the equivalent of a Tommy Hilfiger label on it.

Re:Very simply. If it could work.

[bmacy]

Actually it can be done fairly easily through PR. The same sort of idea but make available a program that will look the CPU Serial# up in the AMD database on-line. Require AMD authorized resellers to include some funky floppy/CD with the program on it (of course it works in Win32 only but you can cut/paste the serial number into an on-line form)... call it the TrueMHz(tm) program... "Does you AMD processor have TrueMHz?"(tm).

Ooo... ahhh... a new buzz word.

Re:Very simply. If it could work.

[glen]

So this person suggests a solution and then explains how his solution won't and can't work.

And then people moderate it up as insightful.

I don't get it.

Re:Very simply. If it could work.

[-brazil-]

Yeah, but labeling the chips beforehand would increase the rejection rates a big deal. Now, they make a platter of 500 PIIIs, and get maybe 2% that work at 1GHz, 10% that can do 900MHz, and so on till the 10% that fail completely. With prelabelled chips, you'd have to throw away all but those 2% that pass all the tests!

(Note: I have no idea whether my percentages are accurate or total garbage, but the process does work like that)

Re:Very simply. If it could work.

[ncc74656]

Very simply. You make your CPUs so they announce to the world at bootup

AMD-K7-Athelon-500MHz Running at 800MHz.

There are 2 reasons AMD and iNTEL don't and won't do that however.

Reason Number 1. It would mean having different CPU dies for each chip they sell. They aren't going to do that. They prefer to simply build a batch of chips and depending on how clean they come out you put a label on to claim a specific clock speed. Yes. Specific clock speeds are determined after the fact before labeling is done, not before.

Not necessarily...you either provide some extra connections at the edge of the die that either (1) allow for some configuration PROM space to be accessed to write in the correct speed or (2) hardwire the desired speed before the die is packaged. (Isn't Intel already using approach #1 to burn serial #s into the P!!!? I think they have some kind of CPU-id utility that's supposed to be able to read out the "advertised speed" of the processor. I thought they did this so that (l)users could check and make sure they got what they thought they bought.)

Reason Number 2. Any Information the CPU issues about itself must go through the BIOS 1st. The problem is that someone with the resources of a 2 bit 10 box a day CPU manufacturer can arrange to have the BIOS altered so the quote above would say nice things like.

AMD-K7-Athelon-800MHz Running at 800MHz.

This is probably beyond the capabilities of the average screwdriver shop, but I'll admit that PCChips did something similar in the past with 486 motherboards with fake L2 cache. The possibility is there...but since motherboard makers usually don't ship processors with their products, what incentive would they have to do this? (The fake cache on those PCChips motherboards was an integral part of the motherboard, so they h4x0red their BIOS to always display "256K L2 cache" (or something to that effect) at boot-up.)

_/_
/ v \
(IIGS( Scott Alfter (remove Voyager's hull # to send mail)
\_^_/

Re:Very simply. If it could work.

[barleyguy]

Reason Number 1. It would mean having different CPU dies for each chip they sell. They aren't going to do that. They prefer to simply build a batch of chips and depending on how clean they come out you put a label on to claim a specific clock speed. Yes. Specific clock speeds are determined after the fact before labeling is done, not before.

Actually, it wouldn't necessarily mean that at all. The way they will probably do it, which Intel is already talking about, is to install a small PLL (Phase Locked Loop) circuit into the processor at marking time. This locks the processor to one frequency, period. It would not only prevent multiplier changes, it would also prevent changing BUS SPEEDS. Intel seems to be ate up that people are buying Celerons and setting them at the 100 Mhz bus that the die was designed for.

I know - the article is about AMD. If you want to know how to overclock these processors, there is an article on Toms Hardware today that tells exactly how to do it. Multipliers and all.

Re:Very simply. If it could work.

[dynamitehack]

The argument that they wont do that because all the chips come from the same die is obviously invalid because they are doing something to the hardware to lock the clock. That same process can be leveraged to implement the "AMD xMHz" response.

If CPUs had a special instruction to identify their "factory speed rating" (like tires :)) then we could have utilities that check this out.

Re:Very simply. If it could work.

[Forrestina]

very, very good plan i belive. anyone wanna mod up?

-------

Re:Very simply. If it could work.

[Forge]

If I explain what's wrong
with other solutions maybe
I have explained why AMD
and iNTEL chose the
unpopular clock locking
solution.

the moderators seem to
agree.

Re:Very simply. If it could work.

[brettrix]

Actually chip fabs could do this pretty easily without multiple dies:

For example, Dallas Semiconductor has made a nice business of laser-etching unique 64 bit identifiers into many of their part lines -- each part is unique! (This takes the form of an 8 bit product and 48 bit serial no. the DS2401 Silicon Serial Number is the simplest in a neat series of such parts.)

Anyway, chips as expensive as CPUs end-up being tested on an automated test rig anyway, so that's a good time clock-rate them, and to then blast an id (with "rate-bits") into them. (Say wasn't Intel going to put an ID into each of their CPUs but this was shot down by privacy-advocates?)

Then this ID could be decoded by CMOS-setup-software and/or the OS boot sequence to display the rated clock speed of the device.

So, barring the politics, this is very feasible. I think it's a great idea; thanks for sharing it Forge! :-)

-Brette

Links.

[Black Parrot]

> he fact is, the new t-bird athlons will be locked to those who don't know the secrets.

You can find a couple of quick links by visiting this article at The Register.

--

Underclocking.

[Anonymous Coward]

Underclocking is the art of running a modern high speed processor far below it's rated clock speed. As an example I have a AMD Athlon 800 MHz processor running at an amazing 200 MHz*. My goal to run it at 4.77 MHz so I can actually win at all those video games I have. By the way, OS is Windows 2000 which runs great at 200 MHz I can't wait to try it at 4.77 MHz. *I have a resistor on top to add additional heat so the processor doesn't get to cold.

Re:Underclocking.

[wiliano]

I underclock my awesome work machine machine by running Windows NT.

Re:Underclocking.

[Anonymous Coward]

In most modern chips you can only underclock down to some minimum rate. Below that rate the dynamic registers within the CPU and other associated chips doesn't get refreshed often enough and the whole system crashes.

I know, because I progressively underclocked a 286 system years back. It started out with a 16 MHz oscillator block (8 MHz system) and worked when I ran it with a 2 MHz oscillator block. I have some 32 KHz blocks, however, that it just wouldn't work with.

It's painful watching a 8 MHz machine go through a POST running at 1 MHz, mind you. You end up counting through each step of the floppy drive stepper motor during the floppy drive seek. If you tolerate waiting through the Memory check, of course.

Re:Underclocking.

[Benley]

Believe it or not, this is not unheardof! Several years back, Fujitsu sold a Lifebook laptop with a 120mHz Pentium MMX. As any self-respecting geek knows, Intel never sold an MMX processor slower than 166mHz! Upon close inspection, this chip was revealed to be a thin-film-package 166mHz chip underclocked to 120mHz for some bizarre reason. We're currently working on "overclocking" this little wonder of engineering to its rated speed. Or would that be "anticlocking" or just "clocking"? The world may never know.

Re:Underclocking.

[timmyd]

maybe they underclocked it to reduce power consumption for the laptop.

Re:Underclocking.

[jafuser]

maybe they underclocked it to reduce power consumption for the laptop

Or heat output... be careful not to cook the 'book...

reputable overclockers (step and company...)

[drenehtsral]

It seemed that AMD was quite happy with overclockers when they were competing for headlines in the Mhz race, infact i seem to remember they quitle lavishly praised Cryotech for building a vastly overclocked highly cooled system... It seems that now that they have some solid market share and a succesful brand, they want to close that door. I personally have been buying exclusively AMD cpu's since midway through the 80486 days, and probably will continue to do so, and i haven't bothered much with overclocking, but i still think it's a little tricky to walk that line between protecting consumers and confining hobbiests...

BULLSHIT

[Lan-Z]

This is bullshit! What's next a speed regulator on my car or maybe even my shoes?

My friend Tom Grant of Contamination will be devistated by this. Tom once overclocked a 486 DX4/100 and the damn thing ran faster than my PII. Props to Tom, but he soon will be fucked in the bootyhole if you ask me.

:)

Re:BULLSHIT

[rlk]

Many cars do have speed limiters. The Chrysler 300M (without the sport package does). Apparently even the BMW M5 has a limiter, somewhere around 155 MPH. Those are both the highest performance cars produced by their respective manufacturers (avoiding issues of who really owns Chrysler, and the Dodge Viper).

Re:BULLSHIT

[British]

Don't high-end foreign cars have speed/horsepower regulator chips on them? I know there's a scene of people modding their cars up with chips as if it was a Playastation for better performance.

Re:BULLSHIT

[Frank T. Lofaro Jr.]

For example, most Galants have built in protection against going fast enough to harm the engine or accelerating fast enough to cause physical harm to the drivers

Umm, I don't think any car could even possibly have fast enough acceleration (even with all governors, etc overridden) to physically harm a driver with G-forces. The car, yes, that could get hurt. But medically can you be injured by G-forces? Yes, but the threshold is way up there. From zero to 100 MPH flat will not hurt you. You car ripping itself apart or you crashing into a tree subsequent to that will.

Yes I admit this is nitpicking.

Does anyone know of a car that could even go from 0 to 55 MPH in 2 seconds? Anything even close to that? I've been on an amusement park ride that did that. Pretty cool. So that has to be a medically safe acceleration. Now imagine the (huge) power output needed to make a car do that. Imagine the likelyhood the tires won't just slip in the attempt (probably zero).

Re:BULLSHIT

[Frank T. Lofaro Jr.]

Sorry for the typo. From 0 to 100 MPH in 1 sec flat is what I meant.

Isn't your sig from Smokey and the Bandit?

[GlitchZ]

I beleive it is.

Re:Isn't your sig from Smokey and the Bandit?

[subuser]

its a sample it could also be credited to Pop Will Eat Itself because it was a joint song but yeah Smokey and the Bandit is the originator.

Consumer Protection

[Brave Little Toaster]

What people should remember here is that this is not intended to hurt the overclocking community as such. It is intended to protect Average Joe consumers who don't know squat about overclocking, but still want a fast machine.

AMD doesn't want its vendors to overclock a chip and sell that product to the consumer as something it isn't. Basically, protecting the consumers from fraud.

Not very nice for the more technically advanced user, but good for Average Joe.

--

Re:Consumer Protection

[13Echo]

We need to realize that there are more "average joes" out there than "Technically advanced users." Besides... There will be workarounds.

Chips have no set speed internally

[Phredrick Dobbs]

The problem is, there is no such thing as an Athlon 600 or 700 or whatever. There is simply a chip with resistors determining which speed the motherboard should run it at.

Some processors turn out better than others, and those are sold as higher speeds. To make a different configuration for each incremental step of a chip would be insanely expensive.

Overclockers are a niche market, there is no reason that AMD should take that much financial damage for their sake.

-Phredrick Dobbs
Emperor of the Universe
Grand and High Protector of Everything

Re:Chips have no set speed internally

[inburito]

Actually on latest Amd processors there are no resistors but rather the multiplier is locked by burning jumpers with laser after the chip tested. Tom's Hardware has a little more on this..

Re:Chips have no set speed internally

[Rumble]

Some processors turn out better than others, and those are sold as higher speeds. To make a different configuration for each incremental step of a chip would be insanely expensive.

This might be the case to a degree, however, I doubt it is the case that there are no changes in configuration from processor to processor (say a k7-600 to a k7-[78]00 or something). Sure the entire design isn't reworked, but the minor optimizations are what makes the difference at those high speeds...

Besides, in CMOS, aside from the difficulty in achieving quality silicon wafers, the actual production of the chips would be relatively constant, and would not lend itself to the amount of variation that you seem to suggest. But of course, as always, I could be wrong.

-Ryan

What about the Abit Board?

[Gandalf_007]

This says the Asus board will not have the clock-unlocking features, but what about the Abit KT7 which was so proudly hailed on slashdot a few days ago? Does it still have these features?

We'll live

[Belgarion]

You can probably still fiddle the FSB. But that pretty limits the options to a few percentages, as the FSB is already at 200.

I've been upgrading the FSB since my 10MHz-8088 machine. I replaced the 20MHz crystal by a 26MHz from my CB radio. Worked great, but the floppy controller couldn't take it. Had to build in a switch, but that crashed the machine due to spikes and stuff. So you had to turn it off, set it back to 20, and boot to use the floppy. Not too much trouble, because the boot took only 10 seconds in those dayz.

Darn. My FSB is now 112, btw.

Why don't the chips know how fast they are?

[Dave Fiddes]

Why don't guys like Intel and AMD put a teenie weanie bit of PROM in their chips (like they had for the serial number) that gets burned with the bus speed/multiplier? Then software (like the BIOS) could read this back when it does its usual CPUID trick and verify that the chip was running in spec.

The BIOS (and/or OS) could report that the system was being over/under clocked. The OCers would love this (especially if it came up in large red and orange letters ;) and innocents wouldn't get burned by dodgy resellers.

This strikes me as being pretty easy to implement particularly if you consider that Intel has already done this with its CPU serial number.

Re:Why don't the chips know how fast they are?

[Chep]

Actually, you don't need a PROM (or, exactly, you don't need a specific PROM : the Duron/Tbird surface copper contact rows can be seen as a special-purpose PROM in its own right).
All you'd need is a way to read the copper contact values (some mechanism like CPUID) into the processor's registers.

The problem is, the BIOS is waaaay out of control from AMD (or Intel). There could be "black hat" motherboard sellers which could quite easily sell "mute" motherboards (all motherboards sold right now are "mute"), so that the luse^Wconsumer still get screwed.

Re:Why don't the chips know how fast they are?

[radish]

The real reason Intel/AMD don't like overclocking chips is because it allows people to buy cheaper chips rather than the more expensive, faster ones. Seeing as the actual production cost of all the chips of the same type is basically the same, the margins are MUCH higher for units which pass testing at higher speeds. If they put in your (very sensible) idea they would basically be saying "hey guys - we don't mind if you spend $100 and instead of $400!", which is obviously against their best interest. By bringing up the spectre of "evil overclocking dealers" they can excuse putting in the multiplier locks etc.

Re:Why don't the chips know how fast they are?

[joshsisk]

By bringing up the spectre of "evil overclocking dealers" they can excuse putting in the multiplier locks etc

I know at least 4 people who have bought overclocked or mismarked machines. The most absurd example of this was a friend who had me install a new video card for them into their PII-266... which, upon opening the case was revealed to not even be a PII! I don't think the company that sound it to him even bother to OC it... They just lied to him about what it was and he didn't know any better. I had a "PII-266" that was marked as a PII-233, myself...

Josh Sisk

Re:Why don't the chips know how fast they are?

[gehrehmee]

This is illegal, yes. Now AMD should be trying to prosecute people who are mismarking these CPU's. Instead, they implement a locking mechanism which places restrictions on _all_ it's customers, when other options (CPU id'ing, or in this case, just an obvious override to the laser-marked settings) exist. There's simply too many unanswered questions to presume that they're doing this for legitimate reasons.

Re:Why don't the chips know how fast they are?

[Axel23]

Prosecution by AMD isn't a realistic expectation.

First that assumes that AMD is notified by the victims in hte first place (and only a small proportion of end users are going to do that).

Then they have to find Fly-By-Night-Computers Inc (who have probably disappeared by then anyway).

Even if it hasn't, by the time a trial date is set, the company will more than likely have declared bankruptcy.

The damage to AMD's reputation is already done anyway.

Then theres the cost of the lawywers.

Preventing overclocking in the first place is a far more effective solution from a time, financial and public image perspective.

Re:Why don't the chips know how fast they are?

[nerdusa]

Here is a concrete suggestion to accomplish this made in a letter to AMD.

Is chip speed determined before final packaging?

[Phredrick Dobbs]

Well, the problem is, for any process that AMD can do after the production of the chip (which is when they determine the speed) which sets the REAL speed, some remarker can do it elsewhere.

If the chip speed is determined before final packaging, which I doubt, then I guess they could test the chip, then write something, then seal the chip off.

AMD does have a pretty interesting system, as you can see at the Tom's Hardware site linked in the story. They apparently use lasers to burn the REAL speed into the chip. I suppose they could have used this burned speed as the REAL speed, and had a softer overclocker method so that it would work as you suggested. Maybe that just hadn't occurred to them.

-Phredrick Dobbs
Emperor of the Universe
Grand and High Protector of Everything

Very simple solution..

[BLKMGK]

I overclock everything I own. Currently Intel Celeron 533's are clocking to over 850mhz. An AMD CPU with a comparable price will NOT beat a Celeron at that speed - period. IF AMD prevents overclocking then they will simply not be price competitive for my dollar and NOT receive it - simple!

I'll be interested to see how this effects the new ABIT MB. It's supposed to allow multiplier modification in the BIOS and was to be my next purchase. Guess what AMD - ABIT and YOU will not be receiving my money if you've clock locked. For that matter even multiplier locking your CPUs may be enough to turn me away since your CPUs cann apparently not handle FSBs much over about 110mhz.

The "issue" here is that AMD supposedly doesn't want remarked PROCESSORS. If that's really what they're whining about and not overclocked SYSTEMS then their current setup of locking the multiplier but allowing external logic to change it is fine ala ABIT. If they've decided this isn't good enough then this crap about wanting to stop remarkers is just that - crap. For that matter even with a solid multiplier lock you can still overclock with the FSB unless they've locked that too - not likely. So they've not actually stopped overclocked "systems" either.

What EXACTLY is it that AMD is trying to accomplish? Or perhaps Tom is full of crap yet again?

I'm voting with my wallet and so should everyone else. If AMD is going to pull the rug out and change their price\performance ratio against the Celeron then they won't be getting my money nor that of any of the friends I advise on computer purchases...

Re:Very simple solution..

[gmhowell]

I think AMD no longer gives a rat's patoot what overclockers do. When they needed every sale they could, of course they glommed on and 'allowed' hackers to OC their chips. Now that they have just enough respectibility (again:) and just enough OEM contracts (again:) they no longer need you. Don't worry: should the sales swing to Intel as a result of this or Merced, or whatever, AMD will once again be trying to court the OC crowd.

(BTW, for those who care: I tend to be chip on the processor and spend money on the graphics, network, and disk subsystems. Works for me.)

Doh! Correction

[BLKMGK]

"I overclock everything I own. Currently Intel Celeron 533's are clocking to over 850mhz. An AMD CPU with a comparable price will NOT beat a Celeron at that speed - period. IF AMD prevents overclocking then they will simply not be price competitive for my dollar and NOT receive it - simple!"

What I mean is - if they prevent overclocking they will NOT be able to compete. Clock for clock the AMD and Celeron "race" isn't - the AMD is a better CPU but it's amazing what an extra couple of hundred mhz will give to a lesser CPU :-)

The Reality of Chip Speed

[tarsi210]

From the mine-runs-hotter-than-yours-does dept.

There is an interesting point that people should start to realize, and it will be a point of education to get them to do so: CPUs do not have a static speed setting.

Believe it or not, it took me several years to get this stamped into my head. What? Overclock? You can do that? Huh.....um...isn't that bad for it? No, and we know that now. It's the amount of people who don't know it that are causing the problems. If the attitude was that every chip can be flexed in terms of speed, the problem would lessen. Then you run monitors to check the CPU speed and protect yourself against having a burnout. Simple. (supposedly)

If you run a chip slower than rated, it lasts longer because it's cooler. If you run it AT rated, it lasts for about its specified lifetime. Run it faster than rated, it'll most likely burn out before its rated lifetime. But then again, in most instances, rated lifetime far exceeds the practical usage of CPUs in the marketplace. So run 'em fast! Sure, they'll burn out, but not likely before you get a new one anyway.

Really, I do see the point that AMD is trying to make. I understand the concern for dork-shops overclocking computers to make that extra buck, and I applaud that effort. However, seems to me that there has to be a better way as opposed to limiting those who want to turbo-charge their CPUs.

Re:The Reality of Chip Speed

[cajun603]

Hey, this could be really cool! Re-enable the "Turbo" switch that is still on most available PC cases. "For when you absolutely, positively, must have that spreadsheet crunched in 200milliseconds!" :-) -cajun

Re:The Reality of Chip Speed

[Asic Eng]

Hm...

In order for a flipflop to switch correctly
you need to maintain a setup time (data valid
before clock edge) and a hold time (data valid
after clock edge). The actually available setup
time is determined by the time it takes data to
propagate through the logic in front of the flipflop.

Lets take a simple setup (an AND gate with
input and output flipflops).

FF ---
AND -- FF -- Y
FF ---

(Wish I could use a decent ASCII graphic here...)

Now the actual setup time on flipflop "y" is:

clock_period - (output_delay + wiring_delay + logic_delay)

This needs to be bigger than the required input
setup time.

Also you need to find the critical path for this
analysis - the wiring delay coming from the upper
flipflop could easily be 10 times longer than
from the lower flipflop. Relevant obviously is
the max.

Now the problem with measuring whether the circuit
works reliably is that depending on the data the
critical path contributes to the result or not.
In the above example, if the lower ff has the
longer wiring delay associated, this will only
show up if the upper ff has a '1' as an output.
(because 0 AND x = 0; 1 AND x = x)

You can not realiable test the device without
having the designer's test patterns. You can't
calculate these patterns without knowing the
exact structure of the device.

If you overclock you ought to be aware that you
do not and can not know exactly what you are doing.

Re:The Reality of Chip Speed

[dgb2n]

The main problem with this is that the lifecycle of CPU's greatly exceeds their likely usefulness.

For example, the rated lifecycle of most CPU's is about 10 years. My Athlon 500 that's overclocked to 700 runs slightly hotter and the lifecycle will be reduced, probably less than half. 5 years from now, I'll be breaking in my K9-3.5 Ghz.

AMD and Intel know that few users (particularly those who will overclock) really care about reducing the lifespan of their processors. We trade them out far faster than they wear out.

r/

Dave

Re:The Reality of Chip Speed

[Jonavin]

That's actually not a bad idea. CeleronII 566 FSB 66 for doing "work"... turn on TURBO for some gaming at 850 FSB 100 or when you don't care if you crash and burn.

.. where's those mod points when you need them...

Overclocking can be dangerous

[RayChuang]

Let's face it.

The majority of computer users out there have reservations about overclocking your CPU because trying it can be a very dangerous thing to do.

Between running the risk of melting down the CPU, causing some peripheral cards to not work properly because you had to increase the FSB speed in your overclocking attempt, and causing general system failures because your power supply can't keep up, it's not really worth the effort unless you're willing to spend the time and effort to get top-notch system cases, extra cooling fans, extra big CPU heatsink/fans, and top-quality 300 watt or larger power supplies.

Besides, nowadays the real bottleneck isn't the CPU. You get much more immediate benefits by getting as much system RAM as you can afford and buying a 7200 rpm or faster hard drive.

Re:Overclocking can be dangerous

[Nargler]

It's ill-informed attitudes like this that spread such fud about overclocking.

People who have reservations and are not informed need to learn the facts so they can make an educated decision. All the people I know who overclock their systems know exactly what they're doing with their systems and the risks involved. If you don't know what you're doing then yes it's possible to damage the system but more often then not it'll just not work and it'll function fine once returned to its default.

The whole great thing about not being multiplier locked is that you don't need to change the fsb out of spec. Also your power supply has nothing to do with overclocking, Athlons need a good power supply regardless since they are more power hungry anyway.

In games cpu, memory and graphic card performance are still the limiting factors.

Re:Overclocking can be dangerous

[Rico_Suave]

I've overclocked every system I've built in the past three years (over a half-dozen boxes). NONE have suffered any problems due to overclocking. When I was building them, as soon as I had any kind of software glitches, I lowered the speed to see if that was the cause. Problem solved.

--

We have the right....

[Dest]

to overclock our systems day in and day out. Right? Yes it breaks the warranty, but if your OC'ing your gonna know what your doing for the love of god! I just had to say what i think. Though if some store does try to sell CPU faster than they are that is kind of cheap. The only way to solve this is examine the comp yourself. Get a program which checks the clock speed or check your BIOS, if you have a BIOS multiplier option or some other horse. -Dest

Bull!

[BLKMGK]

If they've changed this from the way it WAS then they are indeed hurting the overclocker. Lock the multiplier on the chip such that the chip can't be modified and remarked so that it looks like a faster chip - fine. But changing it such that motherboards like the ABIT can't change the multiplier using external logic? Bullpucky!

Even with the way they've supposedly done it now I can overclock with the FSB so how does this protect anyone from buying an overclocked "system"? It doesn't! If the FSB could be jacked up over say about 110mhz I might not be so unhappy but I've not seen many reports that can do it. What's that give me against say a Celeron 533 that runs 850mhz and costs a few dollars more than a DURON?

What exactly are they trying to accomplish?

AMD should embrace the GFD's

[Kondoor]

Gold Finger Devices just make sense. AMD should put a set of pins or some way to connect a GFD to its chip, but to do so make it so you have to remove a cover and then addon a circuit board that is to big to hide inside the cover of the CPU. This way you allow people to overclock there processor but you can just stick a GFD on a processor overclock it, and remark it. Since most people will notice a big chunk of pcb sticking out there processor.

Re:AMD should embrace the GFD's

[CptLogic]

>>Since most people will notice a big chunk of pcb sticking out there processor. Interesting idea but it won't work for the following reason: The reason AMD has gained ground on Intel more than any other, is because companies like Mesh, protek, Simply, Time, Evesham, SSC and other big name direct sales kids, opted for AMD solutions instead of Intel. The majority of people that buy these systems are never going to look inside the case. They come as a ready set up package in the majority of instances, with all the hardware you want. People won't see the evidence. And if they did, if AMD "supported" overclocking, when the consumer complained and said "Oi, I paid 1500 currency units for a 950Mhz AMD Athlon and what i've got is actually a 700Mhz Athlon overclocked to 950, I'm not happy!" The retailer can say "hey, it's an Athlon, it's running at 950Mhz, AMD are cool with us doing this to their processors so tough doo doo!" AMD have to make sure that if someone pays for a 950Mhz processor, that's what they get. I, however, beleive that if I can sucessfully hack something to make it better for me, then I should be allowed to do it freely. I don't see how any clock locking mechanism cannot be circumvented by hardware hacking. I have seen an overclocked Mobo that was hacked to run at voltages it was not originally designed to support, hence a new CPU could be added without needing to upgrade the old Socket7 Mobo to a Super7. 'Course, the FSB was still limited to 66Mhz but that's another story. Chris.

Re:AMD should embrace the GFD's

[theFool]

Wait a minute...

The point is to protect the computer idiots! How many people do you expect to open the case and look at the processor? Especially when they probabally don't know what the processor looks like, how can they know that GFD isn't supposted to be there?
How could this help at all?

It's kind of understandable though

[archmedes5]

I can think of two reasons for this, sales is one, since unlike intel, they can't afford to lose money on this venture, and it's position in the market place isn't quite as secure as intels.

I think however one of the most important reasons is to prevent the remarking of processors, which seems to have moved to AMD chips recently. It's a dissapointing turn of event, but AMD may not have had a whole lot of choice.

An alternative

[gunner800]

Much of the wild-ass speculation (also known as Slashdot groupthink) is that AMD is doing this because unstable overclocked systems are damaging their brand name. I think this is a plausible explanation, although none of us really know the motiviation.

If this is the reason, I'd suggest an alternative. Don't impose restrictions on motherboard designs, impose restrictions on vendors. They could make their licensing such that in order to sell AMD products, overclocked systems must be clearly labeled.

Selling an overclocked system that's not labeled could be penalized simply by making the vendor liable for repairs, or by having to pay AMD a crapload of $$.

Poof, no more (especially) unstable systems with the AMD name on them, and no more angry overclockers.

Or am I smoking crack? I'm basing this on the way Smith & Wesson is handling their vendors after the lawsuit, but processors are not handguns.


My mom is not a Karma whore!

Re:An alternative

[Shadow99_1]

Well the biggest problem here is that most people that sell oced systems don't do it with direct contact with AMD, but through resellers who aren't that picky & really can't be. The things that will hold for the distrubiters can't be applied to those below them because they aren't buying the chips with the license. Heck I've seen low budget oc shops buy a few hundred chips from someone at a computer show (though thsoe were celery chips from intel not AMD chips). The purchaser came back the next weekend with their own booth selling systems which were just oced chips (because they were celey chips they were simply FSB oced). I don't think their is a legal way to avoid it.

Re:An alternative

[rtscts]

impose restrictions on vendors

absolutely. i dunno what things are like in the USA, but on my side of the pond, less than honourable vendors do not use $400 ABit motherboards in their crappy boxen.

if they do, the eeprom idea that someone else came up with to ID CPU's that are running at a different speed than the mfr's rating, plus a motherboard vendor approval scheme, will ensure that it's damn obvious if a vendor isn't on the level, with no way to engineer around it.

Re:An alternative

[Brian Knotts]

I'm basing this on the way Smith & Wesson is handling their vendors after the lawsuit, but processors are not handguns.

What vendors? Hardly anyone is buying S&W since they bent over for Clintigula.

--

There is still light at the end...

[Mr_Escher]

Sure enough, Tom's Hardware has done it again. Although it may seem messy, removing fine lines from the cpu packaging itself with etching solution, it sure beats cracking open the Athlon casing and constructing (or buying) a hack method of setting clock multipliers.

The thing that gets me is: there is no (obvious) reason that a motherboard manufacturer couldn't produce a board that 'ignores' AMDs settings. It could provide a display such as: CPU: AMD Duron 700 @ 950Mhz because it could get the manufacturers specification from the links on the cpu and then use its own settings for the actual clock. Flashing something like an omnious red "CPU Overclocked!" message on startup would dismay would-be reseller from selling cpus out of spec to unsuspecting users and tarnishing AMDs well earned brandname. Is this a viable solution for hardcore overclockers and AMD alike?

Just my 2c but at the current abismal $AU exchange rate that comes to about 1.2c US! Utopia isn't all it's cracked up to be.

Re:There is still light at the end...

[Rico_Suave]

Yes, there is such a thing on most MB's, but can be disabled with a simple BIOS tweak.

--

Because remarkers don't warn users

[Phredrick Dobbs]

AMD is looking out for its less savvy customers, namely those who buy whole systems at once. A remarker could just buy several of those motherboards, build systems, overclock the chip, and then sell them for a great profit. Then, when he starts getting sufficiently paranoid, he disappears. It's not that hard, really.

AMD was attempting to, but has failed, to prevent a remarker from having the technology to cost-effectively remark their chips. The chips are overclockable, as you can see here.

-Phredrick Dobbs
Emperor of the Universe
Grand and High Protector of Everything

This DOESN'T prevent overclocking!

[SpinyNorman]

You can still overclock Athlon/Duron the same way you do Celeron/PIII by increasing FSB speed.

There is no difference between AMD and Intel in this regard - they BOTH lock the multiplier. While it might have been cool to allow it to be adjusted, there is no way to prevent remarkers from abusing this.

Given that AMD has been selling 900MHz Thunderbirds downbinned to 700MHz, and that Duron appears equally overclockable (same way as Celeron), I'd still rate AMD as very overclocker friendly!

Don't hack on AMD over this

[HardCase]

I guess I liked the idea of a non-locked processor when I first saw the review, although I'm not sure why. But I also didn't believe that AMD would really produce the new chips without locking the multiplier. Obviously they are concerned about remarking...we've seen plenty of evidence that it happens.

But what about a less sinister reason? It doesn't make much economic sense for AMD to sell a 500MHz processor that any old Joe can overclock to 700 or 800MHz just by changing the multiplier. And what happens when somebody really juices the chip and it fries? Is AMD supposed to replace it under warranty? Obviously not, but how are they to know?

Of course, the determined hacker is going to overcome this, I'm sure. You hear about all sorts of tweaks, from add-in interfaces to a soldering glob job to conductive ink on the traces. And that's OK...because it's pretty apparent when the chip fails what caused it.

I'd sure like to have a non-locked processor, but I think that they are a thing of the past...and given that chip prices tend to fall rapidly, I think that overclocking is less of a way to get a faster chip cheaper and more of a way to get bragging rights.

My take? It's a non-issue.

=h=

Re:Don't hack on AMD over this

[Mr_Escher]

...and given that chip prices tend to fall rapidly, I thinkthat overclocking is less of a way to get a faster chip cheaper and more of a way to get bragging rights.

I agree with the most part, but I think that the price gaps between high and low end processors can be a little puzzling. Look at any table of speed/price for Pentium IIIs or Athlons and there is a point in the table where prices suddenly rise by hundreds of dollars for a puny 50Mhz increase. A 650 Mhz Athlon (we don't have Duron here in oz yet) is 70% the speed of a 950, but less than 30% of the price. The money I'd save by purchasing a cpu for a few hundred dollars less (and overclocking for the same performance) would buy a nice stick of ram, a big hard disc etc.

Overclocked Motorscooters =:-)

[drenehtsral]

I was lurking in alt.scooter (a discussion group for the dweebs among us who own motorscooters (including me =:-) ) and somebody with really bad english asked how to overclock his 50cc scooter. People do that sort of stuff a lot. a 50cc scooter is actually quite happily capable (on flat terrain) of going 40mph, but by law to be sold as a Class B moped in new york state they are governed to 30, which sucks.

On the other hand my Vespa 90 goes 45 so i had to get a full motorcycle license to ride it. On the plus side, i can now legally ride my motorcycle too =:-)

Sounds like the Guns issue

[Malachi]

The Government wants to reign in control so that bad people don't have them. Bad people will still have them and now I won't.. Why.. cause I'm a good person.. and good people get fucked every single day because of a stupid people. I just wish I lived in a world where when stupid people did something, people realized it was stupid people and not the good people. Dammit, don't tell me how to live! I'm an adult dammit.

-Malachi

What you are missing

[Life Blood]

Many computer companies never were enthusiasts. Apple started out that way but IBM never did and many other startups stopped being enthusiasts very early when they realized they could make a buck.

Enthusiasts are not an accurate predictor. While linux seems to be taking off, tons of other enthusiast projects have gone nowhere. Tons of enthusiasts still write text adventures on the net, but I doubt a video game company is really going to start developing them commercially again.

Companies develop for a platform for several reasons. One, it has the most marketshare so they can make more profit. Two, its easiest to develop for so they can make lots of profit. Three, its what the company itself uses so they know there is a market. Linux faces many problems because its not the marketshare leader and its not what the company uses either. It can easily slip under the non-technical management radar even though the techs all use it at home. All things being equal Linux is probably easier to develop for, but a company may not have many/any professional linux programmers, so all things may not be equal.

BTW you have no "right" to overclock. So while AMD may be hurting its enthusiast market, its also going to make more money because many of those enthusiasts are going to have to buy more expensive chips now.

Re:What you are missing

[grahamm]

Why do you have not have the "right" to overclock? Should the hardware designer not be at liberty to apply whatever clock (s)he likes to any chips in the design? The manufacturers publish datasheets which show the timing graphs, but what right do the manufacturers have to mandate the clock rate which is applied to their chips?

Re:What you are missing

[vrt3]

Depending on what industry a developer works in, all sorts of horrible things can happen if chips are designed in out of spec. In the Medical Device Manufacturing industry, you'd damn well better be using chips within the specification of the manufacturer. I suppose in more trivial markets, i.e. game consoles, designers can get away with whatever they feel like doing.

No offense, but I think you are missing the point. I totally agree that engineers aren't supposed to design out of spec, but I think the original poster was talking about the end user. If I buy a CPU, I would think I absolutely have the right to do with it whatever I want. I can fry it in the microwave if I want to, I can overclock if I want to. All on my own risk, naturally.

Re:What you are missing

[philg]

I don't think the original poster's point was that companies should start major development projects based on what the enthusiasts were doing, just to be "ahead of the curve". He's just looking for customer service, and he sees no reason why the customers most excited about a company's products are the ones the company seems least interested in keeping.

My take on it is that they figure they have to do less to keep you as a customer, because you're already really psyched about their product. Of course, if the companies were more encouraging to the enthusiasts, they might get a competing company's "enthusiasts" to jump ship. And if customers paid attention to the way these companies treated loyal fans, they might be a little leery of committing their business to a company whose level of service varies inversely with how intersted the customer is in the product.

Ah, well. One more reason to get on the clue train.

phil

Re:What you are missing

[grahamm]

Except that multiplier locking also prevents underclocking. If the chip is set for x6 multiplier, then for a given FSB the clock frequency is fixed. So you cannot run the system at 500MHz with a 100MHz FSB, you have to run it at 600MHz. Designers often want to run at lower clock rates than maximum

Re:What you are missing

[Tungz10]

You're still free to do whatever you want with it. There's no law forbidding you from overclocking. AMD just exercised their right to add whatever THEY want to their cpus, which is overclock protection.

BTW, if it was easy to overclock, AMD would have to raise their prices. People would buy lower marked cpus, but AMD still has to cover its costs.

It also complicates support for everyone involved if a computer starts randomly crashing because a user overclocked. How would you like to be the tech guy on the other end when some guy thinks YOUR software is crashing his computer when in reality it is his overclocked cpu?

Re:What you are missing

[GreatUnknown]

Do you want the CPU in your pacemaker overclocked so that it's 40% less reliable?

Not really, but would you really want to overclock the CPU in your pacemaker so it goes twice as fast as it's supposed to?

Ultimate Underclocking.

[S�gnal ll]

with my incredible power switch, i can underclock my awesome machine to 0 MHz.

Re:Ultimate Underclocking.

[dillon_rinker]

See, the funny thing is, you think you're being funny but you're actually mistaking the "no power" state for the "no clocks" state. Running a system at 0MHz is a sort of demanding task; it requires that NO timing signals be sent to the CPU (thus maintaining the CPU indefinitely in whatever state it's in), but that the CPU be ready and waiting to start working again the instant the clock starts again.

Re:The answer is

[hawk]

On-die fuses.

The same die is used for multiple speeds, as noted elsewhere, and chips get sold at the speed for which they test reliably. There's no way of knowing which chip will be a 700 and which a 750, and thus the chip can't be set to directly reveal its identity.

However, the *range* of speeds at chich a chip might test are known ahead of time. With this, it is possible to put something destructable to indicate speed. Have fuses inside for the range of 600-900, in increments of whatever, and blow those for speeds in excess of the rating. A similar method for the physcial chips would be break-off tabs indicating hte speed; break off distance beyond which the chip isn't verified.

hawk

DOn't Despare

[Meenky]

Tom tells you how to over clock it anyway. It's just a little harder. goto http://www.tomshardware.com/cpu/00q3/000711/index. html and Tom will tell all.

Actually...

[DebtAngel]

AMD has a horrible problem with people remarking their Slot1 chips. The problem is most distinct in Australia. Read the Monday Blurn on Tom's Hardware for the scoop (including the contents of correspondence from AMD. However, this morning Tom turned around and disclosed how to get around the "lock" and overclock the chip anyway. The overclock involves burining away and recreating the contacts on the chip. The advantage to that is that chip alterations will be very, very, very obvious (or at least I believe that to be the case - we'll see soon).

Re:Actually...

[barleyguy]

The chip alterations probably will be obvious, but the "boxed" processors are sold with a fan attached. It should be possible to overclock a processor, put it into a counterfeit box, move the holograms from the original box, reattach the factory fan, and sell it as a higher clocked processor. When the fan is permanently attached to the processor, nobody ever looks underneath it.

Re:Actually...

[DebtAngel]

Okay you got me there.

Mind you, if you're buying the whole box from a less than reputable source, what are the odds you're going to look very hard at even the fan?

So yeah, either way the hanest customer loses out.

One problem

[DebtAngel]

Most classic Athlons don't overclock much past 105 FSB (before you go spouting the 200 FSB number at me, that's DDR, and most BIOSs use the smaller number, as far as I know).

I don't see the newer chips ones being much different, and with BIOS settings going to 105, 110, and to the stratosphere, I don't see you getting much of a boost just by playing with the FSB speed. If you want the goods, you have to play with the multiplier.

Re:One problem

[SpinyNorman]

That depends on what clock speed you're talking about. The lower speed parts overclock better because there's more headroom in the core voltage.

The Duron, which is cheaper and only exists in "slower" (still 600MHz+) speed grades is much more overclockable though.

Re:The Big Lie: Overclocking does not "blow up CPU

[CptLogic]

I seem to have spotted your deliberate mistake. To run a CPU faster than it's standard rating you need to increase the clock multiplier, yes. However you also need to feed it extra energy to run faster, i.e. you have to up the voltage. Chris.

Not different CPU dies...

[PinkyAndThaBrain]

Intel already has some kind of PROM in their chip for the microcode/ID stuff and AMD could put something simular in too, in fact they only need some fuses which they can very very easily incorperate.

Test/rate the chip and program speed in PROM and blow fuse to make reprogramming physically impossible, or just use a set of fuses and blow the right ones to encode the speed if you dont want to use EEPROM or some such.

Re:Not different CPU dies...

[Fishstick]

This would work if the blown fuse scheme only lowered the stated speed only with blowing further fuses.

000000000000 = 1000 Mhz
100000000000 = 950
110000000000 = 900

etc..

so altering the chip by blowing additional fuses could only lower the stated speed.

Now you have to make sure that you leave enough fuses for future speeds, what do you do... start at 2000 MHz and blow fuses down to your current TOL at 1000?

Re:Not different CPU dies...

[Chris Pimlott]

Um... if you just use normally binary 12 bits is enough to get to 4096 Mhz. 16 bits is enough for up to 65536 Mhz.

Still doesn't solve the issue of just fscking with the BIOS to make it report whatever you want.

Re:Not different CPU dies...

[Fishstick]

well, kind of my point actually. If you use binary, someone could come along and burn the next high-oder fuse to up the clock

0001 = 1Mhz

could be altered by burning the next higher order digit

0011 = 3Mhz

My question was about coming up with a encoding scheme so that any tampering to burn additional fuses would result in a lower stated speed.

Re:Not different CPU dies...

[curri]

What about negating the bits, so blowing a fuse actually makes a 1 to be a 0 ? That way you could only go down by blowing fuses, right ??

Re:Not different CPU dies...

[Detritus]

You can add a security fuse that is blown after the chip's ROM is programmed. Once the security fuse is blown, the ROM can't be reprogrammed.

Re:Not different CPU dies...

[Forge]

I thoght of that. And mentioned it.

People want upgradeble BIOSs. You have to
block end users and enthusists before you
afect two bit box makers and criminal CPU
retailers.

Re:Not different CPU dies...

[Tungz10]

It would be more efficient to code them the normal way in binary. For example: 000=700, 001=750, 010=800, 011 = 850.

Here's the catch. For each bit, you have two fuses. You always blow one fuse. If the first fuse is blown, the bit is 1. If the second fuse is blown, the bit is 0. If they're both blown, that's invalid, and the chip shouldn't operate.

Re:Not different CPU dies...

[MentlFlos]

The only problem I can think of with this is those scummy retailers who are really good with a soldering iron and have a small pile of $.03 fuses.
$.12 worth of fuses, a little time and vola, instant 1ghz

Other than that, great idea :)
---------------------------------------
The art of flying is throwing yourself at the ground...
... and missing.

Re:Not different CPU dies...

[Helge Hafting]

The only problem I can think of with this is those scummy retailers who are really good with a soldering iron and have a small pile of $.03 fuses. These aren't fuses on the circuit board, it is fuses on the silicon chip itself. You can't buy that kind of fuses. Of course, you can't stop the scum retailer who is really good with a electron microscope, clean-room and chip manufacturing equipment. But his fake chips will be a hell of a lot more expensive than the genuine ones.

How bout

[aliens]

This sux big fat oprah titties. I have on order, a motherboard and CPU, that I know COULD do 950Mhz+ but I WON'T be able to. So let's see my options now...

1) Duron 600 with an Asus A7V with the VIA KT133 chipset, which will probably have just as many little problems as the KX133 did, seeing how they're just about the same. tentative price $250+

2) Celery 2 566 overclocked to 873Mhz+ on the very excellent stable 440BX chipset price $215

I was all for getting a Duron, but now after waiting and waiting for a decent mobo, they're coming out overpriced and castrated, just what I wanted. Bah, if no one was allowed to know about the clock multipliers before hand I'd have nothing to complain about, but it's kinda like hearing from a friend that he test drove a kick ass car. Then you go to buy one, and they stuck something behind the gas pedal so you can't push it all the way down, but the price remains the same.

Re:Overclockers are a weird bunch

[RayChuang]

Given that most motherboards that can accept the Pentium III already have at least ATA-33 IDE hard drive support, the best thing to do is get a 7200 rpm or faster hard drive. That makes a big difference if you're reading big files.

And with the price of CD-ROM drives so darned cheap nowadays, you should also consider getting a 48X to 52X drive for US$50-US$60, too. ;-)

When your system has enough system RAM, it will run quite fast because the OS doesn't have to use the hard drive as virtual memory, which speeds things up quite a bit.

This was already tried and failed

[idot]

The serial number which is printed onto the chip package can be rubbed down and a new serial number can be added. It just makes remarking a little more difficult.

This happened with the Pentium MMX and with DRAM modules some years ago. And Intel decided correctly that the only solution is a multiplier lock.

Re:Why not **PAY** for more MHz, like everyone els

[shuffler]

Actually, overclockers are more like people who modify their cars from factory specs; such as boring out their engines, or putting different size tires on. It's illegal in some countries, but not in the United States.

its all about money

[ArchieBunker]

What percentage of the desktop market does linux take up compared with windows? 1 or 2, maybe even as high as 4. The cost is not justified just for a small fraction of your customers.

Re:Why not **PAY** for more MHz, like everyone els

[joshsisk]

Just like everyone else. You overclockers are as bad as pirates who crack time-locked "trial" versions of software to get the full version for free. Ya want the full version? THEN PAY FOR IT K1DD13Z.

Not at all the same thing. If I buy a computer, I can do any damn-fool thing I want to with it: overclock it, not overclock it, convert it into a water fountain, whatever. What's the problem with that?

Josh Sisk

Re:Why not **PAY** for more MHz, like everyone els

[Zebbers]

mod this up! that is the exact point. And just like cars, performance modifying can cause serious irreversible damage, especially for the unskilled. There are serious, very expensive, risks involved with overclocking. And as far as paying for more MHZ, thats not really the point. Anyone can buy a gigahertz cpu, but to take a lesser cpu, modify and spend $$ to get it to work even better is a show of skill and brings pride. And along the car analogy...anyone can go out and buy a corvette/ferrari/aston martin(fill with preferred sports car), but to take a stock production car and turn it into a performance car takes labor and skill.

But they could ....

[taniwha]

Reason Number 1. It would mean having different CPU dies for each chip they sell. They aren't going to do that. They prefer to simply build a batch of chips and depending on how clean they come out you put a label on to claim a specific clock speed. Yes. Specific clock speeds are determined after the fact before labeling is done, not before.

Not quite - if they are locking frequencies then what they're probably doing now is bonding out the frequency programming pads when the package the die - what this means is that the robot that solders the tiny wires between the chip carrier and the die wires up different wires depending on whether wafer sort decided that the dies were fast enough ie they decided how fast the die ought to be before they packaged it, not after - most chips get 2 sorts of tests - before and after in order to weed out bad die early - packaging is an expensive step - also all the die on a wafer tend to run at the same speed because they've all received the same processing - sometimes extra circuitry is added to a wafer to allow it to be easily characterised.

What they could do instead is to add an extra set of pins saying what speed they think the chip should run at and make those available in an internal register for the OS to print at boot time. This way you could have overclocking and a CPU that announced how fast its manufacturer thinks it should run.

Re:Eulogizing?

[Lord Kano]

eulogize (yl-jz)
v. tr. eulogized, eulogizing, eulogizes.
To praise highly in speech or writing.

http://www.dictionary.com/cgi-bin/dict.pl?term=e ulogizing

Sounds like an accurate enough use of the word to me.

LK

Puhleeeze

[BLKMGK]

Troll, go home. "Out Quake"?! Spare me - I get as many FPS as I need and run Distributed.net in the background just fine thanks. The fastest system I've got, not counting SMP systems, is cranking darned near 3million keys per sec! The CPU cost me all of about $80.

Slap your benchmarks on the table for comparison if you must play the "mine's bigger" game but I'll bet my bang for the buck makes you look pretty silly. Stop trying to justify spending all that money with floating point benchmark jibberish - if it doesn't get me something in the real world I'm not interested.

"No nothings" from an anon poster? People in glass houses shouldn't throw stones. I've been at this a good long time and benchmarks hardly rule the world. Run a Celeron at 850+ and it performs just fine. Go away troll, far away....

Sad thing is I think the AMD CPU is awesome and own stock! But if they can't meet my performance bang for the buck rule they don't get my cash. I've been waiting for the DURON for my next upgrade but if this thing is locked down I'll save my money. I can just slap a Celeron in those systems instead for an outlay of less than $100. (sigh)

Re:Puhleeeze

[Robert S Gormley]

Hey, spanky, anyone can run distributed.net in the background. Because it's a LOW PRIORITY application. When someone else wants CPU time, it idles. I can do that with a 386sx33 too and still get full usefulness out of it.

No it doesn't

[DebtAngel]

The FSB has nothing to do with the chip, really. When you mess with the FSB you mess with most of the components on the mobo. Classic Athlons can't get much past 105/107 no matter what you do. Granted, I have no idea how the core voltage will affect the Duron (which runs at what? 1.5V normally?), but I don't see it cracking 110/112.

The reasons Celerons can overclock so high is because they run at a measly 66 MHz FSB, and they are designed to run at 133. They are, after all, just a P3 with half the cache turned off. :)

Re:No it doesn't

[Denjiro]

I've got a classic Athlon 500 running at 770 currently. 7.0 multiplier and 110mhz FSB. 110 seems to be the max that you'll get with any frequency, and large number of people less.

Here's why.

[joshsisk]

what difference does it really make to AMD if people overclock or not? if they break the chip, they will have to buy a new one!

If they make it easy for you to overclock, then they also make it easy for "DiscountComputersByMail.com" to overclock, and when your grandmother burns out the 950MHz machine that was actually a 700MHz, then she thinks that she broke it and buys another. Or maybe she thinks "I should have listened to my friend Bea and not gotten an Athlon."

Either way, Grandma's getting screwed, and that's why companies make it hard to overclock. If we lived in a magical world where only hobbyists and geeks wanted to overclock their machines, they wouldn't care. But there is always someone willing to make a buck off of people who don't know any better.

Josh Sisk

Why is it that....

[BLKMGK]

everyone who claims that they can prove that overclocking causes "data corruption" always does it with super secret code they can't give out and can only demonstrate in their own home?!

So long as you've not got the CPU glowing in the dark and the peripherals so far out of spec the bus rings like a bell it's NOT a problem! I do my taxes on an overclocked machine, I play games on an overclocked machine, I do my mail and WEB on an overclocked machine.

The only reason I ever see instability is because I push my main machine right to the bleeding edge. The other 6 machines aren't clocked right to the edge and run 24X7 for months on end with NO problems. That includes NT4, WIN2K, WIN98, and RH 6.2. All of them run Distributed.net's code too and you'd better belive it bumps the CPU temp up a few extra degrees too :-) One CPU, a cachless Celeron 266, has run 24X7 at 448mhz since it was brand new. RIPs MP3 and burns CDs just fine - cache doesn't effect it for those uses and it cranks keys with the best. Where's this data corruption problem and how come it hasn't "melted down"?

OC fraud

[sik puppy]

Several threads have commented on unscroupulous resellers selling oc'd systems and charging premius for it - there is already as system for dealing with this - its called fraud - if I pay for a 950 mHz Athalon, and you sell me a 700 oc'd to 950 and charge the 950 price, you have committed a crime

If the bios just identifies the cpu and clock speed, and flashes a warning on bootup if the cpu is overclocked. That should defuse the scumbags.

This is good

[Animats]

They're complaining because a motherboard manufacturer had some DIP switches on a pre-release board but got rid of them for the production product. Big deal. That's good engineering; any diddle knobs, DIP switches, or jumper pins you can eliminate should be eliminated; you get rid of a manufacturing step and simplify installation.

I'd be more impressed with the overclocking community if they reported uptime. "It works" does not mean "it stayed up long enough to run the CPU speed meter program". "It works" means "we ran the AMI CPU diagnostics for 72 hours with the machine in the burn-in oven at the max rated system temperature and there were no errors".

If they worked on cars, the overclocking crowd would be at the side of the freeway waiting for a tow truck.

Underclocking

[adnt]

I think that AMD is actually wanting to stop also those overclockers at home, not only OCs fooling their customers.

If AMD is going to put the correct multiplier on the chip, they could as well make it so that it was readable using some code instead of blocking the multiplier to be exactly that. If it was done that way, it would be easy for the BIOS to tell at what speed the chip is supposed to run. It would be also easy to code a program to check for tampered BIOSes, which report wrong MHz.

It's funny that many slashdotters are supporting AMD when they do this, they would be much angrier if it was Intel would start locking the multipliers now (I know, Intel has been doing this for ages already.)

To my most important point: The processor speed is determined so that the expected life time will be specific. By overclocking it'll be decreased and by underclocking increased accordingly. They are not only stopping overclocking but also underclocking, which also tends to be useful sometimes. You might want to assure longer life time for your box. Other possible case is that you want to build a silent box (for e.g. mp3 playing, whatever where you don't want disturbing noise). There it would be good to underclock the CPU so you could use just heat sinks instead of fans to keep it cool. Also in some cases there is the kind of circumstances that underclocking is needed.

One more scenario where locked multiplier is bad: Future motherboards with higher FSB. You can't the chip to work with those. Big advantage for AMD was that even some of their AMD K6 CPUs worked with 100 MHz as theis multiplier could be decreased.

Even though underclocking/overclocking still works via FSB, it brings the other problems related to that - some devices being more unstable because of weird speed etc.

You didn't get what I said

[RayChuang]

Look, I said originally that more or less that overclocking can be dangerous -if you don't know what you are doing-.

Overclocking is more than just cranking up the speed of the CPU. Because higher speeds can heat up the CPU quite a bit, cause bus speeds to be beyond the safe limits of many peripheral cards, and in many cases cause a bigger draw on the power supply, knowledgeable people will do things like get a high-quality CPU heatsink/fan, extra cooling fans, a beefier power supply and check around to get peripheral cards that works at high-than-normal bus speeds.

In short, if you want to overclock your CPU, you better do your homework or you'll be wondering why the CPU literally melted down and you're getting strange OS crashes.

Re:You've got it backwards...

[Fishstick]

K, that makes sense, I guess. Still, someone pointed out that you still rely on the BIOS to faithfully report the manufacturer's rated speed. Not sure how you ever get past that anyway.

Why block it out?

[d2htornado]

While I've never done it, I personally don't see what's wrong with overclocking a chip. If you take it upon yourself to mess with the chip speed settings, etc. and you F-it up, it's your own fault. Just don't go crying to AMD/Intel/whatever saying that their chip sucks. They designed it to run at that speed and the designed speed was surpassed, making any kind of warranty, etc. void. So why did AMD block out the ability to overclock these new chips? I don't know. But they are taking away a freedom enjoyed by many people in the overclocking community-the ability to tinker and learn from tinkering. This is one of the best ways to learn-by doing. That's how I got my start working with computers. I think that AMD is making a big mistake by locking out the ability to overclock their chips.

Possible?

[F0rlorn]

I'm talking mostly from ignorance here, but hell, if you can do this I'm guessing you can find a way to overclock an AMD chip.

Re:Possible?

[jkovach]

It works... I performed a similar modification on my TI-85, but it isn't 4X faster. It's more like 1.5 times faster, but I added a switch to select between normal and overclocked modes that probably added additional capacitance to the circuit. But it certainly is noticeable. I recall reading that this modification undoes some slowdowns that TI put in to increase battery life.

Acronym of the day : FUD

[Rico_Suave]

Sorry, while the K7 line *does* have a better FPU, it's not *THAT* good. I have two systems - an Athlon 650 and a P3/600E that's oc'ed to 744... and the P3 is quite a bit faster in Quake.

--

Idiot.

[Rico_Suave]

I have overclocked chips that have been running for two years straight, 24/7. Just 'cause *you* can't do it properly doesn't mean others can't.

--

Good thing!

[gelfling]

Take all the money you spend on liquid nitrogen heat exchangers, Peltier plates and go buy an SSD array. Since most of the performance problem for most applications in the real world is I/O eliminate all the mechanical parts altogether. If you want better game fps go buy a high end video card.

OTOH we should be able to sue vendors and recover if they sell us a grey market machine that's relabeled - just like if we bought a used car sold as new.

How about PowKet?

[Plastix]

Socket 462 Male -> Socket 462 Female, in between we have a PCB with the ability to change the voltage going to the chip... It works for them old 100 Mhz Mother boards with 500mhz Chip, why not a new Duron/Thunderbird?

It can still be done, though.

[electricmonk]

The tecnique is detailed in this article over at Tom's Hardware.

This development will tend to weed out all but the most hardcore overclockers, though, as far as modifying the clock multiplier goes.

Re:Troll?

[Robert S Gormley]

No. He's a fucking moron. However, I'll let you off on the assumption you're insulting the moderators, not making a serious claim :)

Re:There might be a very good reason for it...

[Forge]

Overclocking or runing without
a fan voids your waranty.
Enthusiasts like me accept that
at the bigining.

AMD is not responsible for
obviusly cooked chips.

Re:BIOS lying won't happen

[Forge]

It's not the BIOS manufacturer.
It's the 2 bit vendor with the
hex editor and the BIOS flashing
software.

Stoping that goy means blocking
users from upgrades. _I_ won't
buy a Motherboard whithout a
flasheble BIOS these days. I
have needed that feature too
often.

Then it is agreed.

[coli]

People, wake up, AMD is more evil then Intel.

Re:Why not **PAY** for more MHz, like everyone els

[AntiNorm]

You're not really paying for an Athlon (example) at XXX MHz, you're really just paying for an Athlon, period. The real reason for the sharp price differences is that in the "faster" versions, the silicon is better. So if you have a piece of silicon that is factory-set at, say, 700 MHz, what is wrong with pushing it to 800 MHz if it is capable of doing so?


=================================

Re:Underclocking, for stability

[lanner]

Good reason for underclocking = router. See linux router project.

Obviously an AC troll, but...

[StarKruzr]

I still feel it necessary to address this for those of us who don't really understand the situation.

It is illegal to crack software because you don't own the software you buy. In fact, when you buy software, you're not actually purchasing the software. If you did, legally, you'd be purchasing all rights to it, including the right to resell it, copy it and distribute it to your heart's content, claim it is your own product, crack it so that it does things other than what it was supposed to do, etc. Therefore, commercial software companies sell us very specific licenses that allow us to use the software in some manner (usually, install it and use it on one computer, etc.).

However, when you purchase a processor, you're not purchasing something that is easily reverse-engineered or copied. Therefore, you simply purchase the piece of hardware, warrantied to a specific, manufacturer-specified clock rating, after which you can do pretty much whatever the hell you choose with it, including use it as a butter knife - or overclock it. There ARE consequences when you use the product in a manner inconsistent with its labeling - but only insofar as the warranty is concerned (it is voided).

~'Kruzr

Re:Obviously an AC troll, but...

[Anonymous Coward]

It's really revealing when we start comparing software licensing to hardware owning. When you own a defective piece of hardware, you are fully entitled to a refund or repair if it doesn't work the way it's advertised. If a piece of software, however, doesn't work as advertised, the entire "licensing" agreement seems to be tilted sharply in the vendor's favor. You as the user of the software usually 'agree' to a disclaimer disavowing the vendor of all responsibility the minute you open the shrink-wrap. This is before you can even see if the shit works!

And most software has some sort of disclaimer that it is not "guaranteed suitable for a particular purpose" or some such thing. It's really quite amazing. The licensing agreement entitles the company to have your cash, soul, and first-born son while obligating them to absolutely nothing.

So when these companies start spouting off about how "piracy" is hurting their business I really have a hard time getting all emotional about it. If the software was sold as a product in the same fashion as a car, house, or even a book (you can write in a book you own all you want, you know) it would be different. But basically the scam is that once you buy the stuff you're done, whether it works or not. How many places will accept opened software packages as returns?

I often think that a far better scheme for both the consumer and the vendor would be that all software is freely available, at your own risk. What you pay for (and what can be far more easily tracked) is support from the company.

Would it encourage companies to write crap software just to make money off the support agreement? I don't think so. A company that achieved a reputation for that could be avoided by the savvy consumer. Would it make mega-millions for the software companies? Probably still. Most computer users still struggle through simple daily operations . . . a reasonable fee schedule for support would allow beginning end-users to afford to ask questions (stupid or otherwise) and high-level consulting would still be a booming business. I can't see the need for that going away anytime soon.

At any rate, I just find it very revealing that buying software is completely different than buying every other kind of product on the planet . . .

Re:Obviously an AC troll, but...

[Frank T. Lofaro Jr.]

How many places will accept opened software packages as returns?

Another question: How much does a shrinkwrapping device cost? :) There's a solution for you... It looks like it was never opened.

Re:Will you people now stop supporting that x86 cr

[Beatbyte]

Yeah. Makes sense to me. Everyone knows we're all Carmack's inside.... just tell that top my damn car!

Re:Why not **PAY** for more MHz, like everyone els

[hamshere]

Why not? It requires skill, knowledge, bravery.. :) There's a certain something in having a piece of equipment that you have modified. It makes it more 'yours'. You can go and buy a computer off the shelf that's just like thousands of other computers, or you can build one. You can modify it, make it just right. It's the same idea, even if the same amount of work is not involved.

Bologna

[Brian Knotts]

If you have have a clue, you know what the limits of your components are, and you don't go beyond those limits. Chips are not designed for a particular speed; they are simply sold by batch. Some will clock faster than others. If you *don't* clock your chip to the fastest speed that it will run at reliably, you're missing out on performance, and not getting your money's worth, IMO.

I've been quite happy with my OCd Celerys; the one I'm using now is an SMP system with two 333s @ 500 MHz @ 2.00v. Didn't even have to boost the voltage! :-)

And it's been extremely stable.

--

Oddly enough..

[BLKMGK]

Even playing Quake it still crunches keys albeit at a lower rate. Gosh, I guess I've got enough CPU left over after UT or Quake to do other things. Must be a pretty slow CPU huh?

The point is I didn't pay much for the CPU and it performs VERY well. AMD's price vs performance ratio was doing well against the Celeron so long as I could overclock it. Stop the overclocking and for me and other overclockers the price of AMD just shot straight up.

Why should I buy AMD when they cost more for less performance at that point? Just hating Intel isn't good enough reason for me to send my money to another company - especially one that's acting like AMD is right now.

Personally I'm waiting to see production CPUs on the shelf with production boards. If all this crap is true I'll be pissed and not upgrade to AMD. Hell, if Tom's site is correct what AMD has done isn't going to stop remarkers anyway - they'll just solder the jumpers on top of the CPU! If it's that easy let me do it in the BIOS...

FPU

[BLKMGK]

I'm hoping that as the AMD CPU gets more mainstream we'll see some optimizations for it and more performance as a result. The FPU on the AMD is REALLY good and their SMP plans awesome but when I'm playing "Joe user" I look at cost vs performance. Not benchmarks but things like how fast does it flip frames in my favorite game of the week. Heck these days not many games except the Quake line seem to be terribly FPU sensitive anyway - since I like UT better.... (shrug)

Heh, RC5 scores influence me too. Surprise surprise - my cachless Celeron cranks out numbers that are right there with my cached chips at the same speed and it played Quake just fine way back when too....

Except that he's right...

[BLKMGK]

Yeah - there's a certain amount of time needed etc. etc. However what he said is correct with regards to CPUs - they aren't built for a specific speed - a 600 is NO different than a 700 unless the stepping changed. In other words - some engineering guy didn't design a specific line for the 600, the 700, and so on.

That being said - push it until it's not stable, back it off, you're there. As a chip line matures it gets cleaner and cleaner as the manufacturer tunes it and gets the FAB running better for higher yields. When this happens most any of the chips will clock pretty high and the only difference is the speed rating stamped on the CPU. This is why the current Celeron line clocks to 800 almost as a rule and in some cases as high as 900mhz! This from a chip stamped 533. Obviously it's got plenty of on off time at say 800mhz or 750 if you want rock solid stability...

I expected to see the same thing happen with Duron and T-bird and unlocking the multiplier was just a bonus. Giving me that added flexability allows me to get my setup tuned just right - limiting the multiplier means I have to rely on the FSB and right now that just sux. I thought AMD was a litle more friendlier than that - hell they allowed the GFDs didn't they? Bad company - no dollar!

Re:Except that he's right...

[Asic Eng]

Guess it's pointless to reply to it now, but here it goes:

CPUs - they aren't built for a specific speed

Correct, but that's true for basically any chip. The issue is that they are tested for a specific speed.

push it until it's not stable

And that is exactly what you can not do.

The problem is that the point it will stop working correctly depends on the task the processor is executing. There is a different max speed for multiplying 3FF * 45C than for 803 * 3DD for example.

If you don't know the critical path of a design you can't be sure it will work at a certain speed - you can only make a guess.

When some company manufactures their chip they'll test it with the right patterns - they rate it for the highest speed grade at which it can pass the test.

If you get a 600 MHz rated devices, and it's possible to buy a 650 MHz device - well duh it means your chip has been tested for 650 MHz and failed. That's why they sell it cheaper...

I don't believe it...

[BLKMGK]

I've heard this often. As a result I've got one CPU that I've kept in a machine running hot as hell for months 24X7. It's a cachless Celeron 266 - running 448mhz. I've done just about everything I can short of really being abusive to kill that sucker - it won't die! I've had friends bring me systems with failed fans who's heatsink was so hot it removed a fingerprint from my finger yet when cooled they ran fine! I've got one PII 233 "in the field" - same one that zapped a fingerprint - that's running 300mhz to this day. It's got to be at least 3 years old - how old is the PII 233 line anyway? It was one of the earliest ones.

I just don't think these suckers die as easily as people think. Since 88 I've only ever had one CPU die on me. It was a Celeron 300A that died within 8 hours and wasn't overclocked - obviously defective and Intel replaced it. Hand that CPU down when you're done with it - it'll make someone else happy :-)

Yes and no.

[Godwin O'Hitler]

Yes there's a conspiracy to create a demand for processors that are far faster than people need, and overclocking is a nice way of saying up yours to the bloat people.
No, that's not the reason you're overclocking. You don't really need that extra power. You just want to brag about how fast you made your processor run. Processors-overclocked-to-1Ghz customers are not the same market as processors-claiming-to-be-1Ghz customers, so if you overclock, it's no skin off the manufacturer's nose.
Of course, there's also another, very small, not too conscientious market that might be interested in overclocking processors; I'll leave you to figure out which market. But I can't believe the manufacturers are trying to close that one up - that would imply they're concerned about our well-being, and you're not going to make me swallow that!

Re:Durons already had this

[taniwha]

Durons do have a set of pins to request the multiplier.

But, if they're external pins, this doesn't get around the 'fake chip' problem - better they are on-die in some form to make it hard for people to be selling overclocked chips.

All of this may be moot in the future - with multiplier-agile mobile parts appearing that can change their multipliers on the fly these pins are becoming simply a suggestion (what frequency you should come out of reset at - but this is probably going to be set low to save power at reset)

The question is

[drinkypoo]

Will this lead to a slocket-A setup, and continued sales of slot A boards?

I mean, those with Slot A boards can just pop the case off of the module and stick a goldfinger on there, but the upcoming PGA style AMD chips aren't going to work quite so easily...

BWAHAHAHA

[Rico_Suave]

Still buying into the cult of Jobs, eh? Sorry, throw away your pathetic Photoshop benchmarks and wake up - a P3/833 will blow the doors off a G4/450

--

chips MADE for overclocking!

[crazy nick]

Here's my proposal: The factories/companies that make these chips could have a separate division for making chips desigined specifically for overclocking.

because of the possiblilities of an exploding computer, these chips could come with a disclaimer, no warranty, no tech support, no this, some of that, etc, maybe a higher price, just so individuals dont have to circumvent the overclocking "lockout" being proposed here. I think this could even make money for processor companies.

cooling methods.

[crazy nick]

See if you can obtain a very small refrigator-type device and put your computer inside it. If you have the cash, try using lots and lots of dry ice. Fans are a classic solution. Move to Alaska or somewhere of comparable temperature. Encase the machine in a waterjacket. Icepacks may do well.

Re:The Big Lie: Overclocking does not "blow up CPU

[Frank T. Lofaro Jr.]

How would thermal damage occur if the chip shuts down before the point at that damage occurs? If the shutdown stops the gates from switching state (state switches and the time in between states create more heat in CMOS based chips than time spent statically in a '1' or '0' state) the heat dissipation goes down drastically. Most of the chip is now clocked and running at 0 MHz.

Am I missing something though?

P.S. I have heard electromigration damage can also occur. This is a separate issue, and thermal shutdown won't save you. Supposedly at the higher speeds (and voltages which go with it usually) the metal atoms can get knocked out of place by electrons and cause breaks and short-circuits in the internal connections.