Turn a 3-Core AMD Phenom 2 into a 4-Core

[gxsolace]

We received word that a Korean hardware review site has managed to turn a triple-core Phenom II X3 710 into a 4 core CPU.Apparently, the quick switcheroo requires a Biostar motherboard says the site, but customers boards that have the same BIOS op

Turn a 3-Core AMD Phenom 2 into a 4-Core : Read more

 

[bustapr]

I first thought of gatting a budget core 2 quad but this really sounds like more bang for the buck. That is if it has minimal bugs.

 

[kyeana]

hmm interesting. I wonder how that effects stability and overclocking?

 

[scryer_360]

It worked! Just loaded a 710, and just like the article says, switching ACC to "Auto" on a biostar board now has me reading 4 cores on the bios loading screen.

The thing is, in order to make the 3 core processor, AMD has to test a processor bad. With millions of the things running down the line, they are probably testing a few samples from a batch (this is industry practice right?). So if one is testing bad, they just take the whole batch down to three cores. That means others in a batch are still good. I wonder if the same isn't achievable with non-biostar boards?

 

[scryer_360]

Though, this has to happen to Intel as well. Why does AMD intentionally harm the bottom line by creating a product that fills no market, unless the product is priced way below the desired finish (that is, the only reason to buy a triple core is that it'd be cheaper than a four core, but AMD is making four cores into triple cores).

 

[jsloan]

like with video cards, because a manufacturer sells it with 3 cores actives does not mean the 4th won't work, it may and it may not.

sometime they because of demand they are forced to ship what would have been 4 cores as 3 cores, but they don't want to lower their 4 core prices, but can't produce enough of them for the demand so they will quietly ship the good 4 core ones as 3 core ones. this happens.

so people should try and then run extensive testing of the core. maybe it had a small flaw the results in periodic errors or errors only with certain operations. one just has to test, test, test.

 

[leo2kp]

[citation][nom]scryer_360[/nom]Though, this has to happen to Intel as well. Why does AMD intentionally harm the bottom line by creating a product that fills no market, unless the product is priced way below the desired finish (that is, the only reason to buy a triple core is that it'd be cheaper than a four core, but AMD is making four cores into triple cores).[/citation]

That would be true unless the 4th core is actually failing their tests. So instead of throwing the whole thing away, why not sell it as a 3-core for less? They end up making a lot more money that way because they're selling what they would otherwise call "garbage".

 

[AndyYankee17]

yeah, you take a $200 hunk of junk and salvage it for cheap, my bet is that they're taking a loss on the X3's

 

[scryer_360]

leo2kp is right though: if the 4 core doesn't work, what would otherwise be trash you can at least get some money out of as a 3 core. Still, in order to satisfy 3 core demand, AMD must be taking batches that had a small failure rate (and remember, if the few samples per batch have a fail, that doesn't mean all the processors in the batch do), they would have to canabalize good 4 cores to make some 3 core chips as nessecary. Unless sales of three core chips are low enough to just use "scraps."

 

[hellwig]

How do you take a loss on something that you would otherwise throw away? At a minimum, selling X3's recoups some of that loss, assuming they don't actually make a profit off it, which I'm sure they do.

Its the same way the Radeon 4870 and 4850 are the same chip, but the 4870 is higher-clocked and uses much faster GDDR5 memory. Re-tooling the production line is where the cost in incurred. The more products you can make out of a single process, the better.

Chip makers have for ages been using a single chip and making minor mods to create different products. The old Intel Celeron was just a Pentium with some cache disabled. A Core i7 920 and 975 Extreme are essentially the same chip, the 920 just has a small, locked multiplier while the much more expensive extreme just has an unlocked multiplier.

 

[Kent_Diego]

This is an obvious hoax. Enabling a disabled core from a motherboard setting is silly. You at least need to draw a pencil line on the chip substrate.

 

[Silluete]

If this new true aren't AMD gonna do something? or at least asking biostar to change the bios

 

[jerreece]

Assuming this is true, it could be bad for AMD's quad core sales. Why pay more for a quad core, if you can buy a tri-core and potentially "unlock" the 4th hidden core?

LOL!

Sort of amusing. But I guess it does sort of make sense. But the fact that simply setting "Auto" on your motherboard causes the 4th core to activate would suggest AMD didn't do a very good job in disabling the potentially bad core.

 

[eddieroolz]

That's neat! I'd so get a Phenom II right now..

 

[Greg_77]

If these processors turn out to be stable, you will see many budget machines built on this processor in the hope they get a quad core capable processor. Lets hope AMD doesn't disable this new "functionality".

 

[68vistacruiser]

It's interesting to see that AMD makes so many bad quads that they have a plan for selling them as triples. Kind of makes me wonder about their manufacturing quality. At least I won't have to worry about it, I don't fall for the AMD hype anymore.

 

[blackpanther26]

I think AMD did this on porpose. Maybe this is AMD's ace? But you are taking a risk on enabling the 3rd CPU core. You may get one that would work fine but then again, you might get one that can't be stable at all even at 1GHz. But for $125.99 you are getting a bargan if you can get it to 4 cores but if not you still have a good CPU. I hope AMD keeps it this way and have a option in their AM3 boards to enable the 4th core. But as for me. I'm set with my Q9650. But I may try this out sence I have a 790GX motherboard (Foxconn). And see if it works.

 

[oicw]

[citation][nom]68vistacruiser[/nom]It's interesting to see that AMD makes so many bad quads that they have a plan for selling them as triples. Kind of makes me wonder about their manufacturing quality. At least I won't have to worry about it, I don't fall for the AMD hype anymore.[/citation]

Or, you could turn around and say, wonder how much cash Intel is throwing down the drain by junking all the chips that don't have 100% stability on all 4 cores. (though right now it seems that Intel just might have that kind of money to waste...)

I also agree w/ blackpanther that AMD probably don't care whether or not we try to enable the 4th core. At best this may actually help their sales from people willing to experiment with things. People who actually NEED a quad core will still buy quad cores, as an X3 has absolutely no gurantee the 4th core would work.

 

[nekatreven]

reminds me of the days of turning the radeon 9500 pro into a regular 9700.

AMD probably wouldn't have made the X3 line if none of the chips had an actual bad core. So be aware that this may be a great way to destroy an X3 if it really did have a bad core.

More than a few people turned a good, fast 9500 into a garbled, artifact spitting, unrecognizable 9700 by enabling 4 bad pixel pipelines with a softhack.

 

[tipoo]

Kind of reminds me of the whole turning a thorton into a barton thing.

 

[Guest]

guys, AMD do a lot more in testing than just chuck Prime on each core for a few hours to see if its all good. Its very likely that a defective core could work fine in 99.99% of use, but AMD has disabled it because it failed in that one small special case that most people won't ever experience. Also reminds me of the Applebred Durons a few years ago that had 64kB of L2 cache. The cores themselves were simply Thortons with a large chunk of the L2 cache disabled. Pencil modding most of the time gave you a full Thorton processor, but sometimes it;d give you a brain dead CPU. Making money out of what would otherwise be garbage makes financial sense.

 

[Guest]

Alot of people assume that there's going to be tonnes of CPUs that work just fine except that one core is broken. Statistically, cache has the most transistors, and therefore potentially the most flaws, while some triple cores may have a bad core, AMD is probably getting a pretty decent yield, and is disabling plenty of healthy quads to sell as triple cores. Yield only improves throughout the lifetime of a stepping, do you think the triple-cores will go away when AMDs yield hits 99%? No, of course not, the X3 competes directly with Intels dualcores, they're just going to disable a healthy quad.

 

[armistitiu]

is this real? can someone confirm this? i find it hard to believe that AMD would let this happen. Also if you don't have a American Megatrends BIOS can you still do this? because i was planning to get a Gigabyte mobo (with my phenom II 720)and as far as i know they use Award.

 

[chy18]

We r not sure it's 100%. I'm from Korea..That happens begin from korea..
We gathering more info about it..
It depands on cpu&mobo.. lol
also lots of korean guys bought heca+mobo..
some people works well into quad core with stability..
some people can't use it into quad core..
Time after time..day by day..we could reconize
which one is needed..it's variable..
lol

 

[Guest]

AMD makes a profit selling it with 3 cores as well as 4. The reason they might have purposely disabled a good core could be as simple as marketing. Some people are willing to pay a hundred bucks for a processor but not 150. If AMD don't have anything to sell to that customer at that price/performance range, someone else will. A lower profit margin is from this market segment is better than none. Time will tell how many of these processors actually have a working fourth core though.