Shades of win and fail at the same time...

Gotta love Gigabyte.

+1 Gigabyte

Gigabyte is a bad boy.

I think you guys are missing the big picture on this one. There is a real reason the core is locked, it doesn't work correctly. It comes down to some people who will unlock the core, have bad experiences and blast AMD chips. They will have a freeze in the final deathmatch of the uber core unlocker tournament...

Second reason is that people will unlock the cores, and later sell or give the computers to others whom may not have knowledge that a defective core is in use. Hopefully a hospital doesn't buy the computer and use it to run your life support system?

^+1 lol

haha Gigabyte lol +1

While that is true. Your example of a hospital is not going to happen how they purchase through legitimate suppliers.

Also it is not a certainty that you will in fact have problems related to the unlocked core/s.

I now know what brand to buy for my next build.

i don't know why AMD would want to block such feature...

if people wants to take risk in getting extra performance out of their CPU, then AMD should let them...

it's not hurting AMD as people who unlocks disabled cores knows the risks to such hacks...

Gotta say... this is a bad idea for people to be doing this and especially bad for Gigabyte to be endorsing it and providing the means to do it. It is called respecting another company. The cores aren't locked because AMD wants you to buy a quad core instead of a dual core, they are locked because they don't function properly. What happens when someone unlocks the core and fry's their chip? Does that person return the chip to Best Buy and say it is defective? RTM? Bad business for AMD.

To icepick314, the reason is that people may corrupt their own Windows installation and lose data, etc..

While I agree they should keep that feature, I guess that 90% of the people unlocking their 4th core don't realize what it means losing a very important file that is due for tomorrow...

I wonder if this will play havoc on software activations...

You people make it sound like unlocking a core WILL result in complete failure.. And while that might happen, just because amd disabled it doesn't mean it's gonna be completely bad.. There could be numerous reasons it got disabled.. 100% full load it crashes.. I don't know about amd, but my core i7 920 only got up to 100% once, and that's only because I prime tested it. It could've simply not have passed their QA, but still be alright for normal day use.

Ya it's a risk, but I doubt anyways I doubt it's gonna take your whole computer with you if it does go. Max I'd say is it takes the motherboard out to.

Not all CPUs with locked cores are faulty. Do you think that AMD would have enough faulty CPUs to keep up with the demands for X2 and X3 Phenom II's? If you get some stable cores then good for you, if not then too bad. No one said that it is 100% guaranteed success and unlockers know it. To avoid errors simply do hours of testing before you decide on using the CPU at the particular setting.

Unlocking a Phenom II X2 550 $109.99 to a Phenom II X4 955 $245.00 sounds like a pretty good risk to take. I doubt that it will damage anything if the unlocking fails. You wouldn't try something risky on a computer with important files without assurance of stability first right?

Exactly, When AMD cannot meet the demand with faulty processors they will simply cripple ones that work. So by allowing people to do this, they are giving them a unit they didn't pay for, and they would have sold for more.

I'm even more surprised that AMD is disabling the cores via software (or microcode) than doing a physical processor cutting to eliminate electrical flow to the "faulty" cores. I kinda think that AMD wants people to do this as a marketing stunt? I'm quite happy with my AMD Phenom II 810, but that's food for thought?

In a batch of cpus if one is defective they all get discarded or bumped down to a lower chip. . . . I think.

My 965P-DS3 died 3 times in less than a year... for that matter, I will never buy another Gigabyte product even if the moon was included with it.

Granted, AMD may sell some x4-qualified chips as x3 or x4 parts. But do we have any reason to believe that this happens often? AMD has some reason to NOT do this. They already have processors that fill those (or similar) product niches. The Athlon II, for example, or older production chips still in retail channels. It's been mentioned here and elsewhere that the Phenom II is an expensive chip to produce. I don't have the numbers conveniently on hand, but isn't the die size/performance worse compared to i7? AMD is running lower margins. Given that, wouldn't it be in their best interest to keep the parts selling for as high as price as possible (again, given thar retail channels haven't exhausted supplies of Phenom I and older Athlon64 x2 parts)?

AMD probably does have some perfectly legitimate reasons to block people from enabling those cores, though. What if they do in fact get a reputation for unlockable cores that fail later on, or can't be unlocked at all? People will buy those x2 parts, hoping to turn them into an x4. Even people on this board (probably about as tech-savvy an audience as one can find) are saying, "Oh, they're just x4s that AMD is locking because they don't make enough faulty chips." If THIS BOARD is pish-poshing and thinking they can get an x4 part for the cost of an x2, what are other people going to think?

I suppose it's logical to say that AMD shouldn't care, because everybody knows that they're taking a risk, and they're not guaranteed anything... but is that really true? Somebody on this very board is going to be disappointed, because somebody's going to buy that x2 thinking it's a perfectly good x4 that AMD sold as an x2. No matter how much somebody can say, "Oh, I won't be disappointed. I know the risks" can we ever really say that? It's watching my friends' kids beg to play games with the older kids and say, "I won't cry if I lose. I just want to play." Then they lose, and... Well, you do the math. I hate to say it, but sometimes it's better to not let the little kids play. It upsets them too much to lose.

If I were running a business, I'd want people to be happy with what they buy, and for them to buy what they intend to buy. I don't want people gambling on what they might or might not get. The last thing I'd want is somebody building their new machine, getting the x2 hoping for an x4, and for the next 2 years always thinking, "Well, it's okay. I mean... I guess I got what I paid for... but... well... I was hoping I was going to get a *good* chip."

It seems like none of you actually know how the semiconductor manufacturing process. It's really hard for them to make a quad core cpu with all cores working!(which is the reason intel makes dual cores and puts two of them in one package for a quad core). The more cores the larger the die size, the larger the die size, the higher the chance of it containing a random defect. AMD has 2x-5x the units with only 2/3 cores working compared to cpu's with all four cores working, which is understood in their business plan to sell the dual and tri core parts. I'll guarantee that they don't have a bunch of working quad core parts that they are shipping as 2 core or 3 core parts.

I think it's great when a bunch of enthusiasts discuss these topics, but it's clear that most of you don't understand the actual manufacturing and yield challenges that semiconductor manufacturers have.

The parts that are dual or tri cores DO contain some type of a problem in the de-activated cores. It may only appear in some situations, but you have no idea what or when that situation is.

That is incorrect. A batch of CPU's consist of 25 wafers. Each wafer contains anywhere from 50 to 500 CPU's, depending on the product version (size of cpu). Of all of the CPU's on said wafer, a percentage of them will fail. Let's take a typical example and say that 10% are GOOD quad cores, 20% are GOOD Tri-cores, and 40% are GOOD dual-cores. You would then have 30% of the die that fail for various things. You might have 15% that fail for logic or scan tests, each die failing in a different way as the defect just happened to randomly occur in the logic area. You might have another 10% that contain a non-repairable defect in the L1/L2 cache SRAM arrays, of course with each defect being at a different bitcell location. That would leave 5% of the remaining failures that may happen to have a defect in "other" circuitry.

Every wafer has defects on it, only a percentage of chips are good. There are teams of people that work all day long on solving these problems to try and get more good chips on each wafer (like myself).

If your unlocking a core to begin with, chances are you know the risks involved.

In reality I think this helps AMD. Sure they'll sell less of their higher end models, but the publicity will only help their sales for the budget chips. I actually put an PIIX2-550 in my htpc, knowing I can always unlock a core or two for some free performance down the road.
(tested all 4 cores @3.2ghz for ~24hrs with superpi)

I would hope that a hospital wouldn't buy mission critical hardware from Crazy Bob's garage sale. If you come to visit grandma and her life support system has flaming skull stickers and cathode lights, maybe you should consider having her moved to another hospital.
On a more serious note, I doubt those sorts of systems use desktop CPUs anyway.
I think it was a good move for AMD. They made it harder to unlock the cores so the average Joe wouldn't do it on a whim and potentially ruin something, yet still leaving it possible to unlock for those enthusiasts who are determined and know the risks.

To all those people who say the other cores are faulty.

I remember when ATi bought out the X800XT (I think) it was really an X850 but had some of the pipelines disabled so they could have a card at a lower price point. If you flashed the X800XT with the X850 BIOS it was exactly the was card.

Personally I would be surprised if most of the Dual/Tripple cores were not perfectly good quad cores otherwise to fill demand AMD must have serious yeild problems.

People seem to think that yields are horrible: They are not.

When a chip first goes into production, yields can be low to very low, but by the time production is ramped up, it is quite high, and by the end of production it is close to 100% defect free. If production was so full of defects, and so unreliable, then computers themselves would fail alot more often than they do, since there would surely be weak parts of the processor resulting from the inconsistencies.

AMD should shut up and ride the wave, it has been a catalyst for their sales.
And in the end only a few do the hack, they know they are taking a risk, and only some are successful and if failing they may buy a new one again and NOT do sh't with it, so what's to worry for AMD?

We don't really know if bugs on a disabled core can be responsible for virusses or bluescreens...
If so it would be better to just leave the core disabled.
But if gaming performance would increase with smaller issues (eg: now and then a glitch in the graphics), enabling cores could be very interesting!
However,with current prices, I'm not going to risk buying a 2core to enable the other 2 cores, if for less than $50 or $80 more I can buy a 4core!

In fact a dual core phenom 2 paired with a Radeon 49xx card, still is going to give me plenty of fps on most games on my LCD screen.
And if not, I could reduce the resolution to 1280x800 or so, where I'll still have plenty of resolution,quality and gaming performance.

I mean, it's no longer 640x480 where every pixel counts. Most games can be run quite comfortably on 800p screens!

Also note some good CPUs are actually have disabled core because AMD want's to meed the demand for (dual) core CPUs.

Say you have 5% failure rate. Each chip has four cores on it. So your chance of getting 4 passes in a row is (.95)^4 = 0.81 or 81%. What AMD has done is increase yield by saying only three cores need to pass. Your rate, in the same situation of getting AT LEAST three cores to pass is: (.95)^4 + (.95)^3*(.05)*4 = 0.99. The yield improves a lot if you're allowed to have one failure. Still, the chance is pretty good that the one core they disabled is actually a good, working core.

Note: If AMD is binning out the four-pass cores as a separate part number, the incidence of the disabled cores actually being bad would be higher.