AMD saves money by not having to design separate chips, losing silicon real-estate to disabled cores is nothing compared to what it would cost to design and manufacture separate triple/dual core chips just to save a few mm squared of silicon. If anything, the unused silicon might help transfer heat to the heat spreader.
- Status
one korean o.c do the test b4..with acc control with phenom-x3 8450
it was working to quadcore..
but,unfountenely he sold his cpu b4..
it has some possibility from triple to quad..
but don't forget it..it's not a 100%..
it depands on cpu,and mobo,also need ur luck
lol
it was working to quadcore..
but,unfountenely he sold his cpu b4..
it has some possibility from triple to quad..
but don't forget it..it's not a 100%..
it depands on cpu,and mobo,also need ur luck
lol
smalltime0
Distinguished
- Apr 29, 2008
- 309
- 0
- 18,780
[citation][nom]oicw[/nom]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...)[/citation]
They dont, they recycle the silicon instead.
Except that costs money and time and effort
They dont, they recycle the silicon instead.
Except that costs money and time and effort
elbert
Champion
- Jan 11, 2005
- 9,704
- 6
- 51,165
Most 3X's have their 4th core disabled due to tdp. The 95W 3X's, with the 4th core, will require more than 95W. The question then becomes what voltage increase price do you get the 4th core. Killing an otherwise good 3X in under 2 years could turn the bargain into a nightmare for some. I strongly suggest anyone attempting this to first note your 3X's voltage. If the 4th core's voltage causes an increases by more than 10% without question set the CPU back to a 3X. On a stock heatsink and fan 6% isn't advised.
AMD is NOT selling processors with one core that is broken. They are selling a perfectly healthy quadcore with one core disabled. Do you guys really think that you can massproduce 3 core processor by pointing out 4 core processors with one bad core??? That is as stupid as it is expensive. All that the 3 core processor is, is a marketing bluff.
My sources are telling me that AMD will pullback the 3 core processor with in a week "due to technical problems and a weak demand..."
My sources are telling me that AMD will pullback the 3 core processor with in a week "due to technical problems and a weak demand..."
neiroatopelcc
Distinguished
- Oct 3, 2006
- 3,078
- 0
- 20,810
[citation][nom]oicw[/nom]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.[/citation]
Exactly. The majority of people buying AMD chips are people that don't really want to waste/spend time toying around. So they'll just buy the chip that is rated at what they need. So if someone makes an x3 into an x4 just means they'll have sold one more cpu to someone who otherwise would've gone intel...
[citation][nom]jerreece[/nom]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.[/citation]
For the reason stated above. Also, I may want to add that no big OEM (hp, dell etc.) I'm aware of supplies the user with enough bios options to actually accomplish this. So many people, and almost all companies of a decent size, will not even have the option to unlock their x3 processors.
Exactly. The majority of people buying AMD chips are people that don't really want to waste/spend time toying around. So they'll just buy the chip that is rated at what they need. So if someone makes an x3 into an x4 just means they'll have sold one more cpu to someone who otherwise would've gone intel...
[citation][nom]jerreece[/nom]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.[/citation]
For the reason stated above. Also, I may want to add that no big OEM (hp, dell etc.) I'm aware of supplies the user with enough bios options to actually accomplish this. So many people, and almost all companies of a decent size, will not even have the option to unlock their x3 processors.
WheelsOfConfusion
Distinguished
- Aug 18, 2008
- 705
- 0
- 18,980
Wonder if this works with the 720. I'd love to be able to turn that into a 940 at will, though honestly I'd be satisfied with the "stock" 3 cores and lower power usage.
nekatreven
Distinguished
- Feb 20, 2007
- 415
- 0
- 18,780
[citation][nom]uiuiu[/nom]AMD is NOT selling processors with one core that is broken. They are selling a perfectly healthy quadcore with one core disabled. Do you guys really think that you can massproduce 3 core processor by pointing out 4 core processors with one bad core??? That is as stupid as it is expensive. All that the 3 core processor is, is a marketing bluff.My sources are telling me that AMD will pullback the 3 core processor with in a week "due to technical problems and a weak demand..."[/citation]
Well, I don't imagine anyone thought a 9700 with half of the pixel pipelines defective could be a 9500 either. Or that an Athlon XP could be an MP with a little scotch tape and some conductive goop.
Same thing with RAM. A lot of the slow RAM you can buy is the same as the fast RAM. It just doesn't look as good under a microscope, and those imperfections dictate it's abilities at certain voltages.
This is not that far fetched. If AMD did find enough chips that only had 3 good cores, its cheaper to add artificially limited 3 core chips to them and sell an X3 line than it is to throw away the ones with a bad (or partially hindered) core. In fact I think the voltage explanation above your post is pretty plausible.
Well, I don't imagine anyone thought a 9700 with half of the pixel pipelines defective could be a 9500 either. Or that an Athlon XP could be an MP with a little scotch tape and some conductive goop.
Same thing with RAM. A lot of the slow RAM you can buy is the same as the fast RAM. It just doesn't look as good under a microscope, and those imperfections dictate it's abilities at certain voltages.
This is not that far fetched. If AMD did find enough chips that only had 3 good cores, its cheaper to add artificially limited 3 core chips to them and sell an X3 line than it is to throw away the ones with a bad (or partially hindered) core. In fact I think the voltage explanation above your post is pretty plausible.
WheelsOfConfusion
Distinguished
- Aug 18, 2008
- 705
- 0
- 18,980
I have no trouble believing that triple-core processors are simply quads that didn't pass QA on all cores or cache for some reason. I also have no issue with the company selling them as triple cores instead of "Defective Quads." I don't think they'd deliberately disable x4 quality quads just to sell more x3's, though. That would be a waste money for them, right? It'd be cannibalizing the profit of their higher-margin items when it doesn't cost anything less to produce. And x3's haven't traditionally sold as well as x4's.
Some of the 7xx will have defective cores. Both intel and AMD have sold cpus with defective parts disabled as a different model number.
Even their high end parts will have bad parts disabled and still sold as high end because there is redundancy for some things like cache built into every cpu. If their cpu has 4 meg of cache, you can bet it probably really has 4.1 megs or so of cache so they can disable some bad blocks that are bound to happen. The same is true of most complex icus.
Both companies have also disabled good cpus to make those cheaper part numbers when demand outstripped their supply of the bad ones. They have done it and they will do it if they don't have enough bad ones to fill the demand. Its a balancing act when using higher end parts to fill lower end part numbers by disabling parts of the chip. You dont want to undercut your higher markup part, but you also want to move product, they will balance the prices and supply to maximize profit.
Neither company says what their yield ratios are. So its a complete guess at how many of the cheaper parts will be bad parts with the bad parts disabled, and how many will be good parts with good parts disabled to fill demand.
If AMDs yield is really high you'll have a good shot at a 4th working core....if its not so good you'll have a bad shot at a working core. At a guess id say their yield on the phenom 2 is good.....for the older phenom id say their yield was not good.
Even their high end parts will have bad parts disabled and still sold as high end because there is redundancy for some things like cache built into every cpu. If their cpu has 4 meg of cache, you can bet it probably really has 4.1 megs or so of cache so they can disable some bad blocks that are bound to happen. The same is true of most complex icus.
Both companies have also disabled good cpus to make those cheaper part numbers when demand outstripped their supply of the bad ones. They have done it and they will do it if they don't have enough bad ones to fill the demand. Its a balancing act when using higher end parts to fill lower end part numbers by disabling parts of the chip. You dont want to undercut your higher markup part, but you also want to move product, they will balance the prices and supply to maximize profit.
Neither company says what their yield ratios are. So its a complete guess at how many of the cheaper parts will be bad parts with the bad parts disabled, and how many will be good parts with good parts disabled to fill demand.
If AMDs yield is really high you'll have a good shot at a 4th working core....if its not so good you'll have a bad shot at a working core. At a guess id say their yield on the phenom 2 is good.....for the older phenom id say their yield was not good.
Niva
Distinguished
- Jul 20, 2006
- 383
- 1
- 18,785
Hah, I still have a ATI 9700 Pro in my main computer for daily usage at home! That card rocks, love seeing it mentioned still. I remember the 9500 to 9700 fiasco.
I do want to add a bit of my thoughts here though. I highly doubt AMD sacrifices an entire batch of quad cores by turning them into tripple cores simply by testing a small number of samples in the batch. They have to be runnning some sort of burn-in benchmark and noting errors in cores and disabling them on individual basis for good reasons. I'd never risk binning my CPU in this way unless I knew I could go back somehow.
I do want to add a bit of my thoughts here though. I highly doubt AMD sacrifices an entire batch of quad cores by turning them into tripple cores simply by testing a small number of samples in the batch. They have to be runnning some sort of burn-in benchmark and noting errors in cores and disabling them on individual basis for good reasons. I'd never risk binning my CPU in this way unless I knew I could go back somehow.
neiroatopelcc
Distinguished
- Oct 3, 2006
- 3,078
- 0
- 20,810
Well if it's just a matter of bios I don't see the real risk. I suppose one could just reset the bios and it'll be x3 again - unless enabling a core that is broken enough to break something else - but if that's the case I'd assume their own testing would've broken the rest.
So finally full story about 2 months later on Tom's. Interesting stuff. What I can't believe is people so out of touch with reality commenting about Intel and why they haven't released a triple core. Amazing. People don't realize that Core 2 Quad's are not quads but two dual core's mated up. AMD quad-cores are true, native, quad cores so any one of the core's can be bad and still have 3 good one's working fine. With Intel they would have to do a Kuma-style job and disable one the two defective dual cores and have a chip running only 50% enabled. Perhaps if Intel was smart they could do this too, but you don't get the benefit of the larger shared cache like you do on the AMD chips that are partially disabled..
neiroatopelcc
Distinguished
- Oct 3, 2006
- 3,078
- 0
- 20,810
[citation][nom]danbfree[/nom]So finally full story about 2 months later on Tom's. Interesting stuff. What I can't believe is people so out of touch with reality commenting about Intel and why they haven't released a triple core. Amazing. People don't realize that Core 2 Quad's are not quads but two dual core's mated up. AMD quad-cores are true, native, quad cores so any one of the core's can be bad and still have 3 good one's working fine. With Intel they would have to do a Kuma-style job and disable one the two defective dual cores and have a chip running only 50% enabled. Perhaps if Intel was smart they could do this too, but you don't get the benefit of the larger shared cache like you do on the AMD chips that are partially disabled..[/citation]
Ehh ? what are you replying to ? who said anything about intel tricore, or an intel quad with two disabled cores? Do you read the comments you're replying to? I assume you do reply to comments, as the story certainly doesn't mention either of these two... although the thread is so old I possibly could have forgotten about it...
Ehh ? what are you replying to ? who said anything about intel tricore, or an intel quad with two disabled cores? Do you read the comments you're replying to? I assume you do reply to comments, as the story certainly doesn't mention either of these two... although the thread is so old I possibly could have forgotten about it...
neiroatopelcc
Distinguished
- Oct 3, 2006
- 3,078
- 0
- 20,810
[citation][nom]jesuscheung[/nom]doesnt work for me... why?? my motherboard is magic pro..[/citation]
it isn't guaranteed to work. in fact there are more boards known not to work than there are that do.
it isn't guaranteed to work. in fact there are more boards known not to work than there are that do.
- Status
TRENDING THREADS
-
-
-
News TSMC founder says Intel should focus on AI, not advanced process technologies- Started by Admin
- Replies: 27
-
Discussion What's your favourite video game you've been playing?- Started by amdfangirl
- Replies: 4K
-
News Intel CEO Pat Gelsinger retires, effective immediately — also steps down from BOD, two co-CEOs step in
- Started by Admin
- Replies: 124
Facebook
Twitter
Instagram