Slashdot Mirror
Intel Puts the Lock on Overclocking
Patrick Schmid writes "Intel included an overclocking-prevention mechanism into the 915/925 chipsets. So far, only Asus and Gigabyte know how to override it. You can start from the beginning or jump to where we discuss the overclocking lockout."
..and 'only' some manufacturers ALREADY know how to get around it.
newsflash, some manufactures have not ever supported overclocking of any sort..
so there's a lock, but there isn't? that's the point of this? it's not like you could blindly choose what motherboard to get before if you were going to overclock it since some of them didn't really support it at..
world was created 5 seconds before this post as it is.
well now i don't even have to think about "intel or amd ?". If Intel want to prevent me overclocking a chip i paid for i will prevent giving them money.
It's not like anyone blaims Intel when someone kills an Intel processor by over-clocking it. I don't see any reason behind the prevention of over-clocking other than to try and make people have to upgrade more often or maybe because they want to lose marketshare.
because people overclock their systems and then try to claim warranty repair. sometimes, the overclocking is done by a middleman who re-labels chips. when the chip melts, the ball falls somewhere between intel and the innocent but bilked customer. this helps cut down on that.
Follow along here. I have a "2600" Athlon, which is really a 2.083 GHz chip, which supposedly takes a 166 MHz FSB. I have lowered the multiplier on the chip, but raised the FSB to 200, since I havd DDR400 memory. No stability issues whatsoever, and various benchmarks report about a 1/5 improvement in memory bandwidth, etc.
I have no real desire to rev the chip higher than spec, in fact, its so damn hot now, I'm thinking about dropping the overall GHz. But its useful to be able to twiddle the multipliers to suit your needs. Thank you AMD, fsck you Intel.
I want to delete my account but Slashdot doesn't allow it.
But rather against merchants that overclock and then sell machines as the next-highest processor. I remember back when Intel first started doing this the company said it wasn't targeting the actions of the end user but rather shady mercahnts.
Dumb question, but does this also affect _under_clocking?
DNA just wants to be free...
It's probably to prevent re-marking gray-market CPUs as faster chips, which burn out, get returned, and cause problems for the reseller and for Intel. Not every move by a corporate is evil, ya know. It would be more difficult than it's worth to make an OC-friendly chip that is un-remark-able. This is a sizeable problem for AMD, but the laser-trim bridges are a decent solution, to which they probably have a patent so Intel has to come up with something else now.
Breakfast served all day!
Why do people overclock their PCs? The money you have to spend for extra cooling could also be spent for a faster CPU. And the manufactuerer does not specify a chip for a certain clock-speed without a reason.
Sometimes I take the other way und underclock my Athlon TB1333. With less speed you can also decrease the core voltage and save some energy. If you combine this with a tool like (L)Vcool, you get a really cool & quiet computer, even with the boil an egg on it Athlon.
But thats just my opinion.
With the advent of high end gaming, people are willing to spend more on a system. But these parts will rise in price accordingly. Building your own will no longer be a viable method of computer purchasing. Dell, HP and others will make sure of that.
With the help of the big two prices will go up, because of the need to make profits. You can't make a whole lot when you're charging $500 per box, but at $3K there's room.
Peace
Maybe Intel is just tired of accomodating people who burn their chip up trying to overclock it, then think to themselves, "Hey, I can just blame this on a faulty heatsink." God forbid they actually admit that they blew a hundred bucks on a chip and proceded to burn it to a crisp for a piddly 5% performance boost.
If you burn up your chip and then lie to Intel in order to get a replacement, you're a loser, you're ripping Intel off, and you're fucking it up for everyone else by forcing Intel to implement anti-overclocking provisions such as this one.
Having said that, people always figure a way around the limitations anyway.
This isn't talking about multiplier lock - with very few exceptions, almost ALL of the pentium 4 line were multiplier locked - Intel never stopped doing that once they started. The only new processors which are completely unlocked are the Athlon 64 FXs, although the regular 64s have all their lower multipliers unlocked (so you can still reach higher FSB even after the raw mhz limit has been reached)>
If you have to ask, you'll never know.
When you have a 400Mhz CPU and you goose it to 600Mhz you have a 50% improvement. To get a 50% improvement out of 3Ghz CPU you have to juice it to 4.5Ghz. And let's face it anything less than a 25% improvement, or in this example a 750Mhz improvment - the actual perceived improvement is practically ZERO. So it seems to be fairly useless in the big scheme of things.
Why not build a machine instead that can boot in 2 seconds or has a 100% disk I/O performance improvement?
Oh wait I forgot - having 0.0054% better FPS playing some 1337 shooter game is da Shit. All hail me and my enormous ferrite testicles.
Intel is also looking to cut down on RMAs. There are a suprising number of asshats that overclock components then attempt to RMA components that fail. Read any overclocking forum or bbs and you find a fair amount of people that have no problem trying to defraud vendors when they burn their chip out from overclocking/overvolting.
Is there any other reason to "put the lock on overclocking"? Since overclocking voids the warranty, support concerns are not valid.
Preventing overclocking makes very good sense for Intel.
This move is NOT designed to prevent end-users from overclocking; that is an unfortunate side effect.
The real reason is because often, shady resellers will be Intel chips, overclock and sell them as faster than they really are. When the chips fail (which, if overclocking is widespread, they inevitbly will in some cases), it looks to the end user like Intel makes crappy chips; obviously this is bad for business.
Now accusations of intentionally marking chips down from what their capable of may or may not be true. In some cases that's justified; better safe than sorry for Intel: they'd rather have chips that aren't performing as fast as they possibly could then chips failing because they weren't capable of the level they were marked at.
Then again sometimes this is a sketchy practice.
However, standard pratice is not to round up when labelling speeds. Witness the P2 266, DDR266, etc. Shying away from DDR666 is no more logical than skipping the 13th floor in tall buildings.
You wish. Tom's Hardware points out that the Intel-supplied fan-heatsink is inadequate.
Contribute to civilization: ari.aynrand.org/donate
The relavent quote: "Obtaining overclocks in the 4GHz range were not an issue though."
Bottom line: wait for retail boards to become available before jumping to conclusions.
"intel blocks overclock yada yada yada"
Does history never stop repeating itself?
whatever their plan with this no-overclock policy is, they seems to make profit with it, because every now and them they release something like this.
I've strugled with an old Pmmx chip that cruchs my packets back home to downclock it and make it live long... wait.. maybe their goal is to stop downcloack. Maybe their chips ALREADY come overclocked and hence, with a live so short that you will change computers faster then the moore law can count hertz units increments! (i.e. you will have two computers market as gigaHertz before you buy any teraHertz)
In the good old days, processor makers would usually sell 3 different speeds of chip -
1. The base speed
2. A little better, for $100 more
3. TEH ALPHA AND OMEGA, for $500 more.
The only real difference between the base speed chips and TEH ALPHA AND OMEGA are clock limitations - I've never seen a chip that couldn't be stably clocked up to at least the next model's worth without remaining stable, and occasionally you get lucky. On paper what's supposed to happen is that processors will yield a few different speeds, with most failing to be stable at the maximum speed - but that's not really how it works when the chip reaches retail.
The big problem with this practice is as follows:
A. The performance difference between, say, a P4 2.53 and a P4 2.80 is almost zero
B. Nearly all P4 2.53's can overclock to 2.8 without any problems.
C. The price difference between a P4 at 2.26 ghz and a P4 at 2.53 ghz is 30 bucks; the price difference between a P4 3.4 Extreme and a P4 3.2 extreme is about $100 - and don't even get me started about the Pentium-M. Overclocking can really save quite a bit of money.
Most cheap people just buy the slowest processor in a family and overclock it to the next in the family - and it very often works flawlessly.