AMD prepares three-core processors

[safan80]

http://www.theinquirer.net/?article=42369

Tri and tri again

By Charlie Demerjian: Saturday 15 September 2007, 12:58
THERE HAVE BEEN back channel stories floating around for weeks about tri-core AMD CPUs, but nothing solid. The rumours have been picking up more steam, and now they are finally solid: AMD is doing a three-core chip.

The first thing that comes to mind is bad yields, but I am told this is not the case for the three-core product lines. There may be some salvage done on bad quads, but that is not the overriding reason to do this.

The main reason is marketing, it seems AMD is learning from ATI. Most non-top SKU GPUs are simply top SKU die with features turned off, and if you look at the success of people unlocking that, you will see that it is far more than salvage.

AMD is probably doing this for two reasons; the lesser being salvage, the more important one being that Intel can't do it. Intel would have a far harder time making a tri-core part until Nehalem next September - it is easy to fuse off a core, far harder to MCM disparate cores.

This will allow AMD to come out with a lot of mid-range SKUs, having a complete 1-4 core range servicing every market. It also allows for complete market differentiation with a year or so's window into a place where Intel is not.

On the technical side, this is pretty trivial to do: three to core four is just a fuse to blow. What it gets you is a whole lot of choices. Remember the smooth run of SKUs, that was the beginning. If your clocks are thermally constrained, having three instead of four cores gives you a bin or two of speed. Given how few games use a second core fully, this might be a real win.

As far as money goes, assuming there is no salvage, three cores could still be a profit win, but it could be a loss. The selling price of a three core is greater than the price of a dual, and if that difference is greater than the manufacturing cost difference between a dual and a quad, AMD wins. If it is not, or people who would buy a quad buy a tri, then they lose.

Overall, it ends up with greater flexibility for AMD. How the firm uses it will determine whether or not this is a win, loss or draw. In any case, look for it on the consumer side, not the server first, and possibly moving over if it works out. µ

[rogueagent6]

Already posted here:

http://www.xtremesystems.org/forums/...d.php?t=158940

and then locked.

[Starscream]

Already posted here:

http://www.xtremesystems.org/forums/...d.php?t=158940

and then locked.

tbh i wonder why it was locked as there are many news topics that have the same value that arent locked.

but besides that.I think i remember reading a AMD news bit a good while ago were AMD said that triple core could be the next step. that it would be a good mix of die size and power.

[adamsleath]

ok;
well since this is a topic of interest...
if it is a product that is possible and is priced between dual and quad core and, and , and

it makes good sense to have as broad a product spectrum as possible.

i still think it's a quad with a bad core that is disabled...

[adamsleath]

thread
CLOSED

:ROFL:
i cant do that....

cant they just blank off one quarter of a native quad design and add more cache like some1 said in the first thread?

[metro.cl]

More flexibility yes but with the price of current CPUs is hard to see how they will fit 1,2,3 & 4 core CPUs on a range from USD 50 to USD 1000

[Vapor]

More flexibility yes but with the price of current CPUs is hard to see how they will fit 1,2,3 & 4 core CPUs on a range from USD 50 to USD 1000

They're gonna have a $1000 CPU?

[perkam]

THERE HAVE BEEN back channel stories floating around for weeks about tri-core AMD CPUs, but nothing solid.

That's why I unlocked it.

A little "too wild" meant that much lack of reliability is too much for xs news.

I will let this stand though, but only if it dies away quickly without any bickering.

Perkam

[Kobalt]

Whoever starts an AMD vs Intel conversation I will personally fly to your home and kill you. Interesting idea though. I wonder what the dies would look like...unless its just a quad with 1 core disabled?

[adamsleath]

of course it's amd vs intel

but it would be better to think of it as amd & intel.

but only if it dies away quickly without any bickering.

- so is there a reason why a tri core is not possible/practical/feasible?
i know nothing about the details of cpu design.

personally i cant see any reason to bother with it unless it's to utilise quads that have a bad core.

...there should be enuf variety in the product lineup just by speed binning.

...and shouldnt there be a whole range of quad core varieties (with various speed bins)....like with the c2d lineup for example...or amd's x2 line for that matter?

i wunda whether we will see a q6400 or an equivalent.

[perkam]

Actually, I now realize that its very feasible and now Fuad has a picture of the die layout too.

It's right here: http://av.rds.yahoo.com/_ylt=A9ibyKT...m/triforce.jpg

Perkam

[adamsleath]

triangular sockets

[Starscream]

Maybe Single core will die of for the desktop market.

that we will only see low speed single cores for embedded systems, laptops and developing countrys,

and that slow Dual Core will be the new budget market.
And the rest of the market filled with triple and quad core.

[adamsleath]

...it's all run by bean counters...whatver the result of yields in production ...& what is supplied to the market.

...but i suppose i see where that is going...if you are still selling single cores; then it makes sense that a triple core could entice more single core buyers to move to either a dual or tricore....

while phasing out single cores.

i think it is more a case of marketing what is produced rather than designing for a market.

i'd like to see a distribution of sales over all cpu's.

the whole concept of designing multicores i thought was to make it modular....just keep replicating essentially the same cores by 'printing' them onto one die.

AS a consumer i like to have choice (even if it is limited)
i think having lots of choices (for essentially the same product design)
like 2160/4400/6600/6850 for example

or 3600/ 5200/ 6000/ 6400...
is more betterer...consumers/mass market likes choice imo.

[Starscream]

...it's all run by bean counters...whatver the result of yields in production ...& what is supplied to the market.

...but i suppose i see where that is going...if you are still selling single cores; then it makes sense that a triple core could entice more single core buyers to move to either a dual or tricore....

while phasing out single cores.

i think it is more a case of marketing what is produced rather than designing for a market.

i'd like to see a distribution of sales over all cpu's.

the whole concept of designing multicores i thought was to make it modular....just keep replicating essentially the same cores by 'printing' them onto one die.

Current the single cores are like 30 to 40 euro were a AMD Dual Core is 50 to 55 euro.
I can see uses for Single core CPUs but the price difference to a dual core is just to small.

and like said maybe some guy at AMD did the math and they figured out that not alot of stuff out there can fully max out a good quad core.
And that a native tri core is so much smaller in die space compared to a quad core thus getting more CPus out of a waffer.

and that those 2 things combined make a Tri Core CPU feasible when it comes to profit.

[metro.cl]

They're gonna have a $1000 CPU?

For desktop probably normally the highest end part has the max price of around USD 1000.

If AMD wants to have a high end CPU the max the could sell it should be 1.000 + o - USD200 so there isnt a lot of space if you look at prices to put all the CPUs

[adamsleath]

regardless of space in the pricing structure...the desired result is more sales.

cheaper units sell more.

more expensive units sell less - and less often

[hollo]

i like this move. i suppose there would be price points where it would become worthwhile for AMD - eg if a 2.5ghz tri-core sold for more than a 2.1ghz quad core, when they have a quad core where 3 of the cores could be binned at 2.5ghz while one is weaker and would only make a 2.1ghz bin they could disable a core and sell it for more as a 2.5ghz tri-core.

i'm not sure how often that would happen, but for overclockers, who need reliable cores that can handle a lot of out of spec voltage and frequency, this is a very good thing imo. AMD

[AliG]

I agree this is a smart move for amd, because even though it may seem useless to servers, it means their yeilds should drastically improve as any cpus that have one bad core can just become a try core, or even two for a dual. I doubt they would produce any single cores this way, but it definitely would be a way to save money

[Kunaak]

the 360 is a example of a tri core in action.
its already being done.
just not for the desktop... yet.

[`danny]

I hope this turns out to be an awesome idea for AMD

[The_Beast]

"don't stick all your eggs in one basket" must be AMD motto

[awdrifter]

These tri-core cpus are probably failed quad-cores. If they price these to compete with the high end Intel dual-core, they would get quite a bit of markets from the non-gamers.

[Jimmer411]

Well if this is true, and they replace dualcores with these then I think that it would be pretty interesting. AMD Tricores vs Intel Dualcores, C2D has an obvious clock advantage, where AMD would multi task better.

[freeloader]

Tri core would be nice. A bit more powerful than a dual core, without the heating or power usage of a quad. It could work