AMD tri core

[Hatman]

Yes tri's are useless maybe we should dump duals then?

Seriously, get a grip guys, itll be anotehr price range that AMD will have filled and that intel willl only be able to answer with high clocked dual cores.

"The only real reason I can see tri core benefiting AMD, is using the bad quad cores."

Erm.. how about to make more money from poeple who want better then dual but cannot afford quad?

 

[theaxemaster]

Yet another instance where the technically minded say "bah, asymmetric=bad" and the consumer says "3 > 2" but in this case the consumer is correct. Though, honestly, I would be surprised if the yield is bad enough that they would have a lot of them to sell to OEMs in sufficient quantity for them to justify sticking them into their product lines. I expect them more in the custom builder channel. If Intel were building monolithic quad cores, they would be doing the exact same thing.

Though I'm intrigued by the prospect of a 6 core system, oddly enough. If you could find a 2P board that took AM2/2+ chips that would be sweet. Should be a hell of a lot cheaper than the socket 1207 stuff.

 

[yomamafor1]

Yes tri's are useless maybe we should dump duals then?

Seriously, get a grip guys, itll be anotehr price range that AMD will have filled and that intel willl only be able to answer with high clocked dual cores.

"The only real reason I can see tri core benefiting AMD, is using the bad quad cores."

Erm.. how about to make more money from poeple who want better then dual but cannot afford quad?

I believe Intel is going to answer this by dropping prices on the quad cores, or even under-clocked quads.

 

[chunkymonster]

No one is slamming AMD for doing it. We're just uncertain what caused AMD to come up with a "tri-core", when they have a native quad core. It was never on the roadmap. Seems like "Phenom X3" just jumped out of nowhere, which sparked some speculation.

As I said, it could possibly be true that AMD have horrible yield on the Barcelonas. From the technical information leaked during the Technology Analyst Day by AMD, Barcelona currently have about 30% yield.

So it would make sense for them to disable a defect core, and sell them for revenue, then wasting a core to mark them down to dual core.

Hmm, my bad on the tone of my reply. Type faster than I think...

The news reports about tri-core stated that it is a native (Barcelona/Phenom) quad core with one core disabled. I would speculate that AMD did not realize the quantity of tri-cores until they started ramping up production on Barcelona. It would be safe to say that's the reason for AMD coming up with a tri-core proc.

Given that monolithic quad core is a relatively new design and manufacturing process, until they get the kinks worked out of the fab, I'm not surprised with low yields. Really just verifies the issues that AMD has had and offers some insight as to why Barcelona has been delayed. Given the low yields and low clock speeds, and even with the delayed release, it still sounds like AMD rushed Barlceona into production.

 

[TechnologyCoordinator]

Some have said, AMD IS BEING INNOVATIVE! And I say, to a certain extent.


Tri-core is a solution to a manufacturing issue, not a solution to a customer demand.

I'm not saying AMD shouldn't do it, in fact, I think it's the right move aside from fixing their manufacturing process.


Also, I wonder if Barcelona dual-cores will be quad-cores with two defective cores?

 

[BaronMatrix]

Does AMD's Tri Core not make sense to anyone else? I don't get it. There doesn't seem to be a niche market for a tri core, when a quad core from intel is only 240. The only reason I can think AMD would be making a tri core is to sell some crappy silicon? So many programs are optimized for 2,4,8,16 cores, not 3! They're selling these just to take it at a loss and get some money back for the crappy silicon. Maybe... Thoughts? It just doesn't make sense. I'm not saying I know what they're doing, I'm not saying that this is what they're doing, but the tri core area isn't going to have much of a niche market, maybe from 200-220?

It allows them to charge $175 for a 2GHz X2, $235 for an X3 and $290 for an X4.SO for people who want a little more horsepower they can spend the extra $60 for an extra core. Not a bad deal. Also, multithreading isn't in multiples of two.

There are conflicting stories but what I've heard is that these ARE NOT quads that don't have four working cores but actual 3 core procs.

 

[yomamafor1]

It allows them to charge $175 for a 2GHz X2, $235 for an X3 and $290 for an X4.SO for people who want a little more horsepower they can spend the extra $60 for an extra core. Not a bad deal. Also, multithreading isn't in multiples of two.

There are conflicting stories but what I've heard is that these ARE NOT quads that don't have four working cores but actual 3 core procs.

First of all, AMD wouldn't go out and design a native three core from ground up. It would take them too much time and resources, which AMD is lacking both of them.

Secondly, Phenom X3 was never on the roadmap. If AMD actually designed a three-core processor, it will be on the roadmap. It seems like Phenom X3 was jumped out of nowhere.

Thirdly, where did you "hear" this rumor? Care to provide information?

Also, in case you forgot, let me remind you to answer these questions for me:

1. Where, and when, did Intel pay 1 billion dollar to AMD for infringing their license agreement?
2. What's your take on my "30% yield" analysis?

 

[BaronMatrix]

Thats actually the first thing i though dragon...

AMD is making a "true" quad core they are bound to have stuff ups as it has too be a pretty complex die arrangement. And if they dont quite get it right and they stuff one core up... they dont have to throw it away now they can just say its a tri-core and sell it off. Just a way to increase there yield which in turn improves the big p word

PROFIT

So you think that a large number fo the original "native" dual core Opterons and X2 ended up as A64? AMD made a killing (share wise) with their native dual when Intel still had MCM.

I think the Sematech rated NUMBER ONE AMD will get yields up to 65%, perhaps 80%. It's not like Hector's sitting there with an electron paint brush. They actually have design engineers that graduated college and everything. Plenty of places made the point that nVidia does it all the time. One design, several models. Some have deactivated parts. the thing is that this could be the precursor to Fusion, where a 45nm IGP goes into the blank.

It will probably also clock higher than quads and use less power.

 

[yomamafor1]

So you think that a large number fo the original "native" dual core Opterons and X2 ended up as A64? AMD made a killing (share wise) with their native dual when Intel still had MCM.

I think the Sematech rated NUMBER ONE AMD will get yields up to 65%, perhaps 80%. It's not like Hector's sitting there with an electron paint brush. They actually have design engineers that graduated college and everything. Plenty of places made the point that nVidia does it all the time. One design, several models. Some have deactivated parts. the thing is that this could be the precursor to Fusion, where a 45nm IGP goes into the blank.

It will probably also clock higher than quads and use less power.

Again, please provide information regarding this?

On the side note, up to 65%, perhaps 80% doesn't make any logical sense. "up to 65%" means 65% is the best yield, yet there is a higher yield rate of 80%?

Did you come up with that too?

 

[BaronMatrix]

First of all, AMD wouldn't go out and design a native three core from ground up. It would take them too much time and resources, which AMD is lacking both of them.

Secondly, Phenom X3 was never on the roadmap. If AMD actually designed a three-core processor, it will be on the roadmap. It seems like Phenom X3 was jumped out of nowhere.

Thirdly, where did you "hear" this rumor? Care to provide information?

Also, in case you forgot, let me remind you to answer these questions for me:

1. Where, and when, did Intel pay 1 billion dollar to AMD for infringing their license agreement?
2. What's your take on my "30% yield" analysis?

This is possible because of Direct Connect. Sure maybe some said "let's sell defective quads as tris" but I would say that it's a pricing strategy. They wouldn't have to do anything but make a mask with three cores. It's still the same design just like Kuma X2. Of course they could also setup the tools to only etch three cores. I'm not a CPU manuf so I'm just spitballing. The North Bridge won't care how many cores there are.


1. Intel lost in 1992, I believe. Ask around.

2. I didn't see an analysis. You took AMDs defect density and said 30%. Perhaps you should redo the math for the slow amongst us.

 

[BaronMatrix]

Again, please provide information regarding this?

On the side note, up to 65%, perhaps 80% doesn't make any logical sense. "up to 65%" means 65% is the best yield, yet there is a higher yield rate of 80%?

Did you come up with that too?

Don't you know how to use Google? Search for Sematech AMD. Fab 30 was the number one fab in the world for a few years if I'm not mistaken. Perhaps the duals will yield higher. Since there is no data and AMD has been known for high yields, it's not too far-fetched.

 

[yomamafor1]

This is possible because of Direct Connect. Sure maybe some said "let's sell defective quads as tris" but I would say that it's a pricing strategy. They wouldn't have to do anything but make a mask with three cores. It's still the same design just like Kuma X2. Of course they could also setup the tools to only etch three cores. I'm not a CPU manuf so I'm just spitballing. The North Bridge won't care how many cores there are.

So you're basically admitting you're not sure if tri-core is from defective quad core or not. Yet, you spoke as if it is the fact.

1. Intel lost in 1992, I believe. Ask around.

Actually in 1993. Intel paid 10 million in damages to AMD. So do you want to explain where your "1 billion" comes from?

2. I didn't see an analysis. You took AMDs defect density and said 30%. Perhaps you should redo the math for the slow amongst us.

I not only quoted AMD's defect density, I also provided a technical document explaining yield management. Within that document is the function for calculating yield.

So I guess you're not going to contest my "30% yield" statement?

 

[yomamafor1]

Don't you know how to use Google? Search for Sematech AMD. Fab 30 was the number one fab in the world for a few years if I'm not mistaken. Perhaps the duals will yield higher. Since there is no data and AMD has been known for high yields, it's not too far-fetched.

Humm

I searched for "Sematech AMD" yielded nothing
"Sematech AMD 65nm" yielded nothing
"Sematech AMD 65nm yield" yielded a powerpoint by AMD detailing how advanced their 90nm process is.

Maybe you can help me a little bit out here?

 

[Martell77]

Makes perfect sense to me. They are drowning in the price war with Intel and if they price these right and actually advertise them, they could make a good profit. Honestly, if you can't afford a quad, whats the next best option a 3x or 2x?

With this consumers can get the best of both worlds. More power than a dual and cheaper than a quad. With people wanting best bang for buck, this could open a market for AMD that Intel could only compete by dropping the Quads price. Intel started a price war and AMD is gonna smack them around a bit now with it. If Intel lowers the price of quads to compete, their profit from them will take a nose dive.

Disabling the core, for whatever reason, does not bother me. This has been done be various companies for years. If it were due to a defective core, as long as the rest works and the warranty is the same as their quads there is no harm.

Good idea AMD. Now try some advertising, make yourselves a brand name that people will recongnize and trust.

 

[NMDante]

Makes perfect sense to me. They are drowning in the price war with Intel and if they price these right and actually advertise them, they could make a good profit. Honestly, if you can't afford a quad, whats the next best option a 3x or 2x?

With this consumers can get the best of both worlds. More power than a dual and cheaper than a quad. With people wanting best bang for buck, this could open a market for AMD that Intel could only compete by dropping the Quads price. Intel started a price war and AMD is gonna smack them around a bit now with it. If Intel lowers the price of quads to compete, their profit from them will take a nose dive.

Disabling the core, for whatever reason, does not bother me. This has been done be various companies for years. If it were due to a defective core, as long as the rest works and the warranty is the same as their quads there is no harm.

Good idea AMD. Now try some advertising, make yourselves a brand name that people will recongnize and trust.

You do realize that it will also affect AMD's quad core prices as well, which will in turn, make their already non-existent profit margins get only smaller.
Let's see, Intel has a quad selling for under $300, so AMD's 3 core would have to be lower than that, let's say $190. Where would that price AMD's quads? $250? $299? Then AMD will have to not price them so low, as to not compete with their own dual core products.
Not much profit there.
Add to the mix is that Intel will introduce Penryn in November, which would likely lower quads again a bit, so more profit margin is gone from AMD. How much money has Intel loss during this price war? How much as AMD? Who's smacking who around again?

Yeah, so let's undercut our own quad core sales with our 3 cores, and make even less per CPU. That's brilliant.

 

[Stimpy]

Better to make less profit per CPU than make a loss on a CPU you have to throw in the bin

 

[Martell77]

You do realize that it will also affect AMD's quad core prices as well, which will in turn, make their already non-existent profit margins get only smaller.
Let's see, Intel has a quad selling for under $300, so AMD's 3 core would have to be lower than that, let's say $190. Where would that price AMD's quads? $250? $299? Then AMD will have to not price them so low, as to not compete with their own dual core products.
Not much profit there.
Add to the mix is that Intel will introduce Penryn in November, which would likely lower quads again a bit, so more profit margin is gone from AMD. How much money has Intel loss during this price war? How much as AMD? Who's smacking who around again?

Yeah, so let's undercut our own quad core sales with our 3 cores, and make even less per CPU. That's brilliant.

So did/does the GF8800GTS cut into or drag down the prices of the GTX?

Unless Intel drops the price of their quads again, this shouldn't effect the 4x's much at all. These could be priced to compete directly with the higher end C2D's and the 2x's could be in with the lower-mid range of C2D's.

Basically it will redefine the market in a way that Intel will only be able to drop prices on the quads to compete as a 3 core processor isn't really a good option for them.

Now don't get me wrong, for this to work, AMD will have to do alot of things right. I just would like to see them execute this properly and return to being competitive.

Note: I have not owned an AMD product since my K6-2 300mhz. I am not any companies fanboy. I just see the potential of this and think it could help AMD. My computer runs on a P4 3.2ghz processor. As I have said in other threads, I plan to build a new computer around tax season and there is no telling what processor brand its gonna have. I just want the best for my money, and if that means a tri-core, so be it.

Will love to see how these bench against the C2D's. The could go head to head with Penryn if the performance is good enough.

 

[NMDante]

So did/does the GF8800GTS cut into or drag down the prices of the GTX?

Unless Intel drops the price of their quads again, this shouldn't effect the 4x's much at all. These could be priced to compete directly with the higher end C2D's and the 2x's could be in with the lower-mid range of C2D's.

Basically it will redefine the market in a way that Intel will only be able to drop prices on the quads to compete as a 3 core processor isn't really a good option for them.

Now don't get me wrong, for this to work, AMD will have to do alot of things right. I just would like to see them execute this properly and return to being competitive.

Note: I have not owned an AMD product since my K6-2 300mhz. I am not any companies fanboy. I just see the potential of this and think it could help AMD. My computer runs on a P4 3.2ghz processor. As I have said in other threads, I plan to build a new computer around tax season and there is no telling what processor brand its gonna have. I just want the best for my money, and if that means a tri-core, so be it.

Will love to see how these bench against the C2D's. The could go head to head with Penryn if the performance is good enough.

No, but the GTS is somewhere between $250-300 dollar difference from the GTX. Not $75-100. Huge difference there. If the GTS was priced $75-100 difference from the GTX, do you think anyone would buy a GTS over the GTX? (Mind you, I'm using the GTS 320MB, not the 640MB. The difference is a little closer with the 640MB).

AMD cannot price these 3X CPUs to compete with their own line, so it will be difficult to find a good range for them. Also, Intel has already priced their Quads under $300, before AMD even had a quad. Nothing to stop them from lowering them even more now.

Look at today's retail market - Intel Quad is $279 for Q6600 (2.4GHz). AMD's top dual 6400+X2 is $244, and the 6000+X2 is $170. So, for $30 more, you can get a Quad over the 6400+, and for about $100 you can get a Quad over the 6000+. You squeeze the X3 between both, and you are now eliminating the top dual core product from AMD, and possibly undercutting all the other dual cores (which could lead to price cuts, which won't help AMD). Even if the 3X is priced at $200, that's still only $80 more for a Quad.

AMD selling these things is fine, but I am just trying to see where they would be placed to be both competitive and not undercut themselves. And if rumors of bad yields is true, they cannot price these too low, or they will cost more to make than to sell.


@Stimpy

Better to make less profit per CPU than make a loss on a CPU you have to throw in the bin

No profit if sold for less than what it cost to make.

 

[jaywalker256]

at least this way they will reduce lost profits.

lets just say (i have no idea how much it costs to make a cpu, and these prices arent anywhere near right, but this is just an example) it costs $100 to make each quad core cpu. one core is bad, rebrand it as a tri-core, and sell it for $80. losing $20 is much better than the alternative of just dumping the chip and losing the total $100.

 

[Martell77]

No, but the GTS is somewhere between $250-300 dollar difference from the GTX. Not $75-100. Huge difference there. If the GTS was priced $75-100 difference from the GTX, do you think anyone would buy a GTS over the GTX? (Mind you, I'm using the GTS 320MB, not the 640MB. The difference is a little closer with the 640MB).

AMD cannot price these 3X CPUs to compete with their own line, so it will be difficult to find a good range for them. Also, Intel has already priced their Quads under $300, before AMD even had a quad. Nothing to stop them from lowering them even more now.

Look at today's retail market - Intel Quad is $279 for Q6600 (2.4GHz). AMD's top dual 6400+X2 is $244, and the 6000+X2 is $170. So, for $30 more, you can get a Quad over the 6400+, and for about $100 you can get a Quad over the 6000+. You squeeze the X3 between both, and you are now eliminating the top dual core product from AMD, and possibly undercutting all the other dual cores (which could lead to price cuts, which won't help AMD). Even if the 3X is priced at $200, that's still only $80 more for a Quad.

AMD selling these things is fine, but I am just trying to see where they would be placed to be both competitive and not undercut themselves. And if rumors of bad yields is true, they cannot price these too low, or they will cost more to make than to sell.


@Stimpy

Better to make less profit per CPU than make a loss on a CPU you have to throw in the bin

No profit if sold for less than what it cost to make.

Maybe not at a profit, but they could reduce the loss from the production of the quads. (If they are in fact just defective quads)

As for the rest, you basically made the point I was going for. They will have to restructure the market to fit the new processors, and being a tri-core, your average person buying a computer in Best Buy will want the tri-core over the dual. The 6000 and 6400 will likely drop drastically in price when the 2x phenoms come out as they wont compete with the new arch (I would think). The 6000 and 6400 would (again I would think) make up the low end where the E4300 resides. Basically, positioned right, the could take the performance crown at all but possibly the top levels.

I don't see Intel dropping the price on the quads yet as they are already really low, as you pointed out. So it potentially puts Intel in a hard spot. They will be beaten in the low, lower-mid, and mid and upper-mid ranges.

AMD could change the game on Intel and if done properly, they will give Intel a bit of a run for their money.

AMD has a real chance to make an impressive come back, but it will take solid quality, competitive performance, and smooth execution of production and marketing. If they don't start advertising like Intel does, I can't see them making a strong comeback.

 

[NMDante]

Maybe not at a profit, but they could reduce the loss from the production of the quads. (If they are in fact just defective quads)

As for the rest, you basically made the point I was going for. They will have to restructure the market to fit the new processors, and being a tri-core, your average person buying a computer in Best Buy will want the tri-core over the dual. The 6000 and 6400 will likely drop drastically in price when the 2x phenoms come out as they wont compete with the new arch (I would think). The 6000 and 6400 would (again I would think) make up the low end where the E4300 resides. Basically, positioned right, the could take the performance crown at all but possibly the top levels.

I don't see Intel dropping the price on the quads yet as they are already really low, as you pointed out. So it potentially puts Intel in a hard spot. They will be beaten in the low, lower-mid, and mid and upper-mid ranges.

AMD could change the game on Intel and if done properly, they will give Intel a bit of a run for their money.

AMD has a real chance to make an impressive come back, but it will take solid quality, competitive performance, and smooth execution of production and marketing. If they don't start advertising like Intel does, I can't see them making a strong comeback.

Not really. Intel can afford to drop prices on their Quads, if they want to push Quads as the mainstream CPU. I really don't see them doing it, but they can. Also, you can look at their present Core 2 Duo prices. The E6600 is priced more than the E6750. Why? To push the newer CPUs, with the newer features? Possibly.

If AMD puts the price of 3X close to Intel's Quads, I see Intel dropping prices on Quads to make the 3X CPUs almost a non-factor. Again, you just have to look at the prices today for their own dual core CPUs. The newer E6x50 series CPUs are priced lower than the E6x00 series CPUs. If Intel intros a new Q6x50 series, I see them priced the same way as the E6x50 series CPUs.

 

[bliq]

I believe Intel is going to answer this by dropping prices on the quad cores, or even under-clocked quads.

acutally, my guess would be that Intel releases tri-cores also to go head to head with AMD. It's brilliant really. and video cards have been doing it for a decade. Really, the benches are going to speak volumes about whether the tricores are really worth it. I suspect that with more apps getting optimized for multiple cores, that it will. As a person who works in a enterprise dev environment, we'd never be stupid enough to limit ourselves to just two cores, as it would leave things like 8 way procs and 4 way procs wasted. Just like the linux kernel used to be, we'd only care about a single threaded app vs multiple threads, 2 cores, 3,4, 8, we wouldn't care. We'd take as many threads as available as our apps need to handle that level of load.

 

[BaronMatrix]

So you're basically admitting you're not sure if tri-core is from defective quad core or not. Yet, you spoke as if it is the fact.

1. Intel lost in 1992, I believe. Ask around.

Actually in 1993. Intel paid 10 million in damages to AMD. So do you want to explain where your "1 billion" comes from?


I not only quoted AMD's defect density, I also provided a technical document explaining yield management. Within that document is the function for calculating yield.

So I guess you're not going to contest my "30% yield" statement?

That's not the only thing from 1993 I don't exactly remember.
I said there were conflicting reports. XBit has a story, ExtremeTech has a story. Fudzilla has ten.

And no I don't believe your 30% yields. Like I said before the data in the link was 1996 data. I would think that the process has improved by now. I scanned through it but didn't see an algorithm.

 

[yomamafor1]

Maybe not at a profit, but they could reduce the loss from the production of the quads. (If they are in fact just defective quads)

As for the rest, you basically made the point I was going for. They will have to restructure the market to fit the new processors, and being a tri-core, your average person buying a computer in Best Buy will want the tri-core over the dual. The 6000 and 6400 will likely drop drastically in price when the 2x phenoms come out as they wont compete with the new arch (I would think). The 6000 and 6400 would (again I would think) make up the low end where the E4300 resides. Basically, positioned right, the could take the performance crown at all but possibly the top levels.

Actually, no. His point is completely different from yours. The problem with X3s is that, they don't fit in the existing price range. Even if AMD restructure the price range, it has to taken into account of profitability. This limited their options for adjusting price range. There is no way AMD can take several half a billion losts.

If Phenom X3s are positioned right, AMD can at least survive. Forget profitability; forget performance crown. Before AMD can crank up the speed, all the performance crown talk are pretty useless.

I don't see Intel dropping the price on the quads yet as they are already really low, as you pointed out. So it potentially puts Intel in a hard spot. They will be beaten in the low, lower-mid, and mid and upper-mid ranges.

Again, AMD needs to take into account profitability. Since it costs a lot more for AMD to produce a Barcelona then Intel to produce two Conroe cores, AMD can't price Barcelona like Intel price Conroe/ Kentsfield.

AMD could change the game on Intel and if done properly, they will give Intel a bit of a run for their money.

Intel has a lot more room to adjust accordingly to AMD's pricing, then vice versa. If AMD drops the price, Intel will drop the price to the point where AMD can't follow, or outperform them.

AMD has a real chance to make an impressive come back, but it will take solid quality, competitive performance, and smooth execution of production and marketing. If they don't start advertising like Intel does, I can't see them making a strong comeback.

AMD's real chance of making a come back is when Barcelona delivers the performance it was designed for, and that is when the clockspeed increases. At the moment, AMD does not need advertisement. They need solid product lineups, and precise execution. At the moment, I don't see any of them.

 

[yomamafor1]

acutally, my guess would be that Intel releases tri-cores also to go head to head with AMD. It's brilliant really. and video cards have been doing it for a decade. Really, the benches are going to speak volumes about whether the tricores are really worth it. I suspect that with more apps getting optimized for multiple cores, that it will. As a person who works in a enterprise dev environment, we'd never be stupid enough to limit ourselves to just two cores, as it would leave things like 8 way procs and 4 way procs wasted. Just like the linux kernel used to be, we'd only care about a single threaded app vs multiple threads, 2 cores, 3,4, 8, we wouldn't care. We'd take as many threads as available as our apps need to handle that level of load.

No, Intel will not likely to introduce a tri-core to compete. They just need to lower the price on their Quads, and in NMDante's word, "a tri-core becomes a non-factor".

If a Q6600 costs 250USD, while a Phenom X3 costs 280, which one will you buy?