Everyone who has any interest in overclocking knows about multiplier locking by now. Many have also read the ‘scare’ regarding bus locking, which Intel has denied and debunked recently. Many have also seen the comment from one of Intel’s support reps about how CPU locking is intended to thwart the remarkers, not the overclockers. As has been said many times before, overclocking is a fairly small sample of the Intel customer base, however remarking is a very big problem. This same support rep indicated that overclockers would not appreciate Intel’s next move, while Tom’s Hardware Guide and other overclocking advocacy sites have sent out a call for all overclockers to voice their displeasure at having to pay for what the remarkers are doing. The thinking is that Intel should be able to stop remarking, without affecting overclocking.
This argument sounds very appealing, especially to overclocking advocates. Intel should put safeguards in place that will allow easy identification of remarked chips, while allowing overclockers to continue to push their chips however far they want. This is such an obvious and simple thing to do… or is it? In order to determine whether it is possible to stop remarking, but allow overclocking, we have to look at what the similarities and differences are.
The argument for overclocking is that it is the chip owners right to do whatever he wants with his property. Overclocking affects nobody but the individual who engages in the practice, and allows him to potentially get more life out of his existing system. With manufacturers making technology obsolete ever more quickly, this is the consumers defense. On the other hand, it may damage components, which will then potentially be returned under warranty service and it affects Intel (or other manufacturers) sales by some degree. The first of the ‘against’ issues is a valid issue, but since processors are very rarely damaged by overclocking it really isn’t of any great concern. As for the second argument, it has already been established that overclocking isn’t a widespread enough phenomenon for Intel to really worry about.
On the subject of remarking – there isn’t really much of a ‘for’ argument from either the manufacturer or consumer. Unscrupulous business people purchase (or even steal) large quantities of processors, spend a large amount of money professionally remarking the chips, and then sell them to vendors and end users at prices below what the equivalent ‘official’ processor sells for yet still make a huge profit. Unsuspecting customers think they have gotten a great deal, and end up running overclocked without even knowing it. The only ones who make out in this scenario are the crooks, and most remarks go completely unnoticed by the customer. The ones who actually do notice are usually those who then attempt to overclock their new processor, only to find out it has already been overclocked to its maximum speed!
Intel currently marks each processor with it’s rated speed on the bottom of the chip, along with a ‘lot code’, which can be cross referenced on Intel’s web site. It would not be feasible to embed the speed into the chip itself, since Intel and other manufacturers regularly ‘re-bin’ their chips to make up for shortages, or to reduce over supplies. What remarkers actually do then, is to either remove or cover up Intel’s markings, and replace it with a ‘fake’ one so it looks like a faster spec’d chip. Various schemes have been proposed which would allow someone to more easily identify a remarked chip, but what if you never get to actually look at it? Since the vast majority of processors are sold installed into a machine, and relatively few users open the system up to look inside, how will any external visual indicators help these people recognize a remarked chip? Any ‘internal’ indicators will hamper the manufacturers ability to re-bin their own processors (in effect, doing their own remarking – though it may be down rather than up)?
In fact, isn’t a remarked processor merely an overclocked processor with a different coat on? If Intel allows ‘some’ people to overclock, how do they prevent others from doing it and fooling the naive public? They might have a special ‘outlet’ that sells directly to end users, but with an admittedly small overclocking population, how cost effective would that be, and if the population isn’t that small, they would have to open up *many* outlets at an even greater cost (and since these would obviously be relatively low end processors, the profit margins would be lower). If they sold these in quantity through distributors, there would be no way to actually control who gets them, and the remarkers are back in business. Another idea of selling ‘overclockable’ processors at a higher price is defeating the whole purpose of overclocking!!
When Intel limited the multiplier options, the overclocking community was fairly upset, however there was still the option of using faster bus speeds, so the limitations were at least tolerable. With future chipsets, certain processors will again be limited to 66MHz as well as being multiplier locked (which is now a single multipler, instead of a limited range). At this time, I fully expect the overclocking community to become very upset and make numerous threats towards Intel. Unfortunately, the issue of remarking is a very serious problem that Intel (and others) must address. It is almost impossible to allow some people to overclock, while preventing others unless the company is willing to invest a very large amount of money to implement and monitor it. Since the percentage of customers who overclock is fairly small, it just does not seem smart business to try to satisfy this minority.
Both Cyrix and IDT already have some form of overclocking protection, by virtue of the fact that their designs only allow certain multipliers already. It also appears that AMD may decide to lock the multiplier on some (or perhaps all) future processors to help thwart the remarkers as well. It seems to me that no matter *what* scheme is used to externally identify processors, it will still not stop unscrupulous vendors and businesses from putting an overclocked processor into a machine, and sell it ‘under covers’ as long as the opportunity is there. Let’s face it, the majority of users shop based upon price, therefore many vendors also shop based upon price (via the gray market). There are actually some ‘vendors’ who specifically build overclocked systems for people – either overtly or covertly. This is exactly the practice that Intel is trying to eliminate, because who knows whether the customer actually realizes what he is getting, or if he will inform the next person he may sell that system to (after all, he wasn’t able to set it up himself, so he probably won’t be able to ‘fix’ it before selling it). As appealing as it sounds to keep overclocking while eliminating remarking, I just don’t think it is a reasonable expectation.
Be the first to discuss this article!