Early Tri Core look

So, whats so good about the tri core?

It lets AMD sell Quads with a dead core. I assume they will be priced halfway between dual and quad, so they may make sense to some.

Nothing wrong with triple, Nothing that great either.
 
Sure it is - there is a whopping 520% increase (or $210) between the cheapest dual core ($50) and the cheapest quad core ($260) :p
You also have to account for performance. The range will narrow down a lot because AMD isn't going to be selling Kuma for $50. :p

Between dual core K10 and quad core K10, there's probably a relatively small price range to cram a tri core. But we really won't know how small until March. Unless K10 remains hard to yield, I think X3 will have a short life span.

My official guess is there will only be a $25-$35 difference between dual and tri/tri and quad cores by then for K10 CPUs closest in performance (i.e. for a model X3 where X2 < X3 < X4, and there are no other X2s faster and no X4s slower to satisfy that).
 
Was this not debunked multiple times already ? :)

No...? How do you propose to debunk that? I just double-checked my numbers, and as far as I can tell they are still correct.

@pxc: I don't have to account for performance when talking about price :p
 
No...? How do you propose to debunk that? I just double-checked my numbers, and as far as I can tell they are still correct.

@pxc: I don't have to account for performance when talking about price :p
I see the problem...you're comparing the cheapest quad core K10 to the cheapest dual core K8, no shit there's a huge price difference between it and the last generation of processors. Now if you want to do a fair comparison, you need to take the price of the cheapest dual core K10 and the cheapest quad core K10, and considering we have no clue what the former might be at this point, all comparisons are moot until its release.
 
I see the problem...you're comparing the cheapest quad core K10 to the cheapest dual core K8, no shit there's a huge price difference between it and the last generation of processors. Now if you want to do a fair comparison, you need to take the price of the cheapest dual core K10 and the cheapest quad core K10, and considering we have no clue what the former might be at this point, all comparisons are moot until its release.


ok how about a C2 Q6600 vs a C2D E4400? or Pentium E2140? :p all the same gen of chips
 
Yes, compare those all you like, those are the same generation. K8 and K10 are NOT, and I believe I alluded to that int the very post you quoted.

point was, there is a huge difference between dual and quad core chips on the market currently, same generation(intel) or not (amd) so a tri core at lets say $150 wouldn't be bad
 
If you are comparing the cost of Quads vs Duals you should use the same cores. Not the more expensive cores in the Quad to the most cost reduced cheapo cores on the market.

Q6600 ($265) vs E6600 ($229).

In which case the gap is small and the quad is great deal. Intel is clearly not leaving room in it's lineup for a triple. AMD may differ.

Say $200 for the quad, $150 for the Triple and $100 for the dual. I am sure some will buy either.

But I don't think having a triple is going to be a big advantage that will have people chosing AMD over Intel.
 
If you are comparing the cost of Quads vs Duals you should use the same cores. Not the more expensive cores in the Quad to the most cost reduced cheapo cores on the market.

Q6600 ($265) vs E6600 ($229).

In which case the gap is small and the quad is great deal. Intel is clearly not leaving room in it's lineup for a triple. AMD may differ.

Say $200 for the quad, $150 for the Triple and $100 for the dual. I am sure some will buy either.

But I don't think having a triple is going to be a big advantage that will have people chosing AMD over Intel.


you don't need to use the same cores for this comparison, we're not comparing the performance of dual to quad, we're comparing the price of the CHEAPEST dual to the CHEAPEST quad, it's a market, not a performance measure in this case.

Chaapest Intel Core 2 Gen dual core is the E2140
Cheapest Intel Core 2 Gen Quad core the Q6600

when you're talking about an oem market, most people won't care what gen the Athlon X2 is vs the Phenom X4, they just know oh man thats a Quad core but it costs $200 more, and same goes for intel. Only people who are a bit more computer savy would know and care, and that is a minority =p

so there is definetly a market to fill with tri cores, for AMD it = profit in 2 ways, 1 it can use a core with 3 good cores and 1 not so good, without binning it as a dual core, and 2, it creates a whole new product to offer the market.

plus chances are, if someone who coud utilize a quad core, 3d rendering server etc, a quad core 2ghz chip would out perform a 3ghz dual core chip.so using the same core as a comparison doesn't really apply unless you're talking about single threaded performance.
 
you don't need to use the same cores for this comparison, we're not comparing the performance of dual to quad, we're comparing the price of the CHEAPEST dual to the CHEAPEST quad, it's a market, not a performance measure in this case.

Is there any point to that comparison. To me if you are doing a price comparison, you compare like items.

2140 cores are much smaller and cost less to fabricate. I suppose Intel could build a cheaper quad out of these as well but performance would likely suffer with the smaller cache.

What has this got to do with triples. The triples will be severely undercut by ultra-cheap dual cores as well.

We will have to wait and see what the market looks like and where triples sit when they actually ship. By then smaller Penryn quads will be available. Intel lowered the price of quads to under $300 with no competition. I am sure that next year with competition and new smaller die (less expensive to produce) Penryn quads we will see Intel Quads under $200.

Regardless, Triples are just a curiosity, a way for AMD to sell Quads with a Dead core. They will never be produced in enough volume to be worth talking about unless AMDs process defect level is really screwed. It is pretty hard to get excited about 3/4 of mediocre PhenomX4. I guess for AMD fans that aren't sure they want more than two cores and don't want to spend much money this might make sense.
 
Wow, just wow.

You must be kidding or...

Yes, I was kidding, hence the ":p" ;)

But my point was more people will go to the store, look at the computers and think "dual < triple < quad". However, the jump from dual to quad will be $200 (even enthusiasts around here go for the cheaper c2ds). If the jump from dual to triple is, say, $75-100, it could be well positioned to pick up some of the midrange market.
 
But I don't think having a triple is going to be a big advantage that will have people chosing AMD over Intel.
Not enthusiasts, but the OEMs might like the X3 for greater stratification within their product range. Even a couple of bucks difference for them can equate to a significant advantage when budgeting and marketing their systems. For them it's all about marketing and product differentiation. I've seen this kind of strategy employed countless times in a number of industries, and to a large extent it succeeds.
 
Not enthusiasts, but the OEMs might like the X3 for greater stratification within their product range. Even a couple of bucks difference for them can equate to a significant advantage when budgeting and marketing their systems. For them it's all about marketing and product differentiation. I've seen this kind of strategy employed countless times in a number of industries, and to a large extent it succeeds.

You talk about this like AMD will actually build X3s. These are quads with a bad core they shouldn't be available in large numbers unless something is wrong. AMDs quads have huge die so they won't want to sell them as triples for a lower price unless they absolutely have to. The X3s effect on the market will be nil.
 
Tri-Core will be great for people too cheap for Quad Core but don't want a "shitty dual core"
 
You talk about this like AMD will actually build X3s. These are quads with a bad core they shouldn't be available in large numbers unless something is wrong. AMDs quads have huge die so they won't want to sell them as triples for a lower price unless they absolutely have to. The X3s effect on the market will be nil.
The overall effect they have on the industry may be negligible, but that in itself doesn't preclude the OEMs from purchasing them and configuring systems with these processors. That's all I meant with my use of the term 'marketing' in sharp contrast to anything related to market share, which appears to be your mistaken apprehension of my post.
 
Not enthusiasts, but the OEMs might like the X3 for greater stratification within their product range. Even a couple of bucks difference for them can equate to a significant advantage when budgeting and marketing their systems. For them it's all about marketing and product differentiation. I've seen this kind of strategy employed countless times in a number of industries, and to a large extent it succeeds.

As you said marketing marketing marketing. It's not a difficult concept to grasp especially considering marketing is thrown at you from the time you wake up in the morning until you go to sleep at night.

Another example, video cards. How often have we seen video cards which are nothing but stripped down versions of higher end cards? How many of these have people been able to "unlock" before and turn into the regular version?

How often are CPUs set at a lower speed even though they can be binned at higher speeds? Very often because while the upper end of the market has a higher profit margin, it's the midrange and lower ends of the market that make the real money due to volume and demand.

 
You talk about this like AMD will actually build X3s. These are quads with a bad core they shouldn't be available in large numbers unless something is wrong. AMDs quads have huge die so they won't want to sell them as triples for a lower price unless they absolutely have to. The X3s effect on the market will be nil.

I take it you've never heard of Celerons? Plenty of Celerons were PIIIs or P4's which had bad cache or something else wrong. Part of it was disabled and then sold off as another part. AMD has done this previously before also.

It's called using your resources to make more money. It's possible that a lot of perfectly fine quad cores could be sold as Tri cores due to demand and a lower price.

Also, a tri core doesn't have to be a quad with a nonworking core. It may just be a core that doesn't clock very high and the other three do. At that point it's better to sell it has a higher clocked tri core than it is to sell it as a lower clocked quad. There are Core2Duos out which cost more than Core2Quads just because they are clocked higher. AMD could do the same thing with the tri cores.

 
If you are comparing the cost of Quads vs Duals you should use the same cores. Not the more expensive cores in the Quad to the most cost reduced cheapo cores on the market.

Q6600 ($265) vs E6600 ($229).

I agree with this theory and thats what I've been thinking along. I have no idea what everyone else is thinking or talking about. Never mind Digital, I dont know where he got his business degree but it sure as hell wasnt from an a creditd school.

Also to add. It would probably serve AMD far more purpose to have triple cores compete directly with Intel's dual core. Hopefully in this comparison amd could actually win something hands down, thus giving them a price to performance advantage. It also alleviates a serious congestion issue with their current multi-core marketing strategy. With the ability to clock cores at different speeds I dont know why they're even bothering to hold onto dual cores. If they could develop a piece of software that could save your preference settings for certain applications then they really wouldnt need dual core and they really outta be able to beat c2d's.

Just my 2 cents on that.
 
Regardless, Triples are just a curiosity, a way for AMD to sell Quads with a Dead core. They will never be produced in enough volume to be worth talking about unless AMDs process defect level is really screwed. It is pretty hard to get excited about 3/4 of mediocre PhenomX4. I guess for AMD fans that aren't sure they want more than two cores and don't want to spend much money this might make sense.


I think it could give us a better idea of defects for sure.If AMD floods the market with this chips in 4 to 6 months,that says a lot to me.
 
I think it could give us a better idea of defects for sure.If AMD floods the market with this chips in 4 to 6 months,that says a lot to me.

It means they are pushing tri core over dual core which should give AMD the more realistic mass of the market.
 
It means they are pushing tri core over dual core which should give AMD the more realistic mass of the market.

I'm guessing that he was saying joe average consumer or desktop workstations, as opposed to enthusiasts.

Although I don't agree with the terminology that they're pushing tri-core. I think it's just a smart way to recoupe defects. Lemonade and all that.
 
I agree with this theory and thats what I've been thinking along. I have no idea what everyone else is thinking or talking about. Never mind Digital, I dont know where he got his business degree but it sure as hell wasnt from an a creditd school.

Also to add. It would probably serve AMD far more purpose to have triple cores compete directly with Intel's dual core. Hopefully in this comparison amd could actually win something hands down, thus giving them a price to performance advantage. It also alleviates a serious congestion issue with their current multi-core marketing strategy. With the ability to clock cores at different speeds I dont know why they're even bothering to hold onto dual cores. If they could develop a piece of software that could save your preference settings for certain applications then they really wouldnt need dual core and they really outta be able to beat c2d's.

Just my 2 cents on that.


thanks for that, but I don't have any sort of eduction in business :p and I don't need it to figure out where "Tri-Core" Cpus might be a GOOD addition to a companies product line.

if you really think that Tri-cores are meant for the enthusiast market, then I think your credited school really isn't that great.
 
thanks for that, but I don't have any sort of eduction in business :p and I don't need it to figure out where "Tri-Core" Cpus might be a GOOD addition to a companies product line.

Tri-cores are a great addition if you have quads with one bad core and they are worth a small premium over a dual core, but they are nothing to get excited about.
 
Tri-cores are a great addition if you have quads with one bad core and they are worth a small premium over a dual core, but they are nothing to get excited about.
tri cores were never meant for you to get excited about them, its more for OEMs, AMD, and their shareholders to get excited about :p
 
tri cores were never meant for you to get excited about them, its more for OEMs, AMD, and their shareholders to get excited about :p

Again. Tris will ship in relatively small number unless AMDs process is riddled with defects. That doesn't seem to be something to get excited about.

Any shareholder excited about Phenom is a simpleton. Phenom launch is a disaster. This is a huge die and right from launch they are only able to get lowish end prices, because it is glitched. Selling anything other bad core quads as Triples makes no sense because it further lowers ASP on this large die cpu.

This is a way to make lemonade with the lemons. But really, too many lemons is not a good thing.
 
Again. Tris will ship in relatively small number unless AMDs process is riddled with defects. That doesn't seem to be something to get excited about.

Any shareholder excited about Phenom is a simpleton. Phenom launch is a disaster. This is a huge die and right from launch they are only able to get lowish end prices, because it is glitched. Selling anything other bad core quads as Triples makes no sense because it further lowers ASP on this large die cpu.

This is a way to make lemonade with the lemons. But really, too many lemons is not a good thing.
Do you think these share holders are all new? how about someone with a few hundred K invested already?

ofcourse they would be happy, rather then a quad core that can't perform with all 4 cores going to a dual core selling for $100, now it's going to be a tri core and sell for $150, thats 33% more profit from what would have been something less

according to scali2, amd yields are less then 50% :p someone posted info about it too, so thats a significant increase in profit, out of lets say
10,000 quad core dies, 5,000 live on to be quads, and now 5000 can be turned into tri cores intead of dual cores

using some simple data
lets say
Quad = $200
Tri = $150
dual = $100
(price)
now AMD can make 5000x50 more then before, thats $250,000 increase in profits from what would have been $500,000 is now $750,000 per 10,000 cpus
(i'm not saying my numbers are accurate) but put that to larger scale, say 100,000 cpus, and you can see a HUGE jump in profits.

Thats only the aspect of utilizing non yielding quads, how about the fact that consumers(ones who buy dell/hp/ETC) can now have 1 more choice in amd products?, its a whole new line actually.

"Introducing the AMD Phenom X3"
simple joe says, oh man X3 must be better then X2 and look its not as expensive as X4, so sure, X3 is good!. since most people who buy pcs, don't know much about whats inside, but we think, bigger number = beter. and you'll have all the salesmen tryin to upgrade someone to an X3, ooh sir, you can upgrade to an X3, thats an multi-tasking increase of 33% for only $49.99!

seriously, it's pretty simple, I don't understand why you keep saying it's nothing to be excited about, it was never meant for enthusiasts, so you won't be excited but other people will.
 
It is much simpler than that. If this is good news for shareholders, it is the time to buy AMD stock. Are you going to buy some?

It doesn't matter how much stock you have or when you bought it. If it is likely to go down, you are still better off selling.

The X3 is not the kind of thing that will significantly help the business. The only thing holding AMD above $10 is them finally coming back on the graphics front with the RV670 which is a compelling product. The CPU front is clearly dead at least until Bulldozer(and probably beyond).

If NVidia drops a killer 9 series graphics card in the new year, watch AMD stock crumble. Either way, the smart money has left this stock.
 
if you really think that Tri-cores are meant for the enthusiast market, then I think your credited school really isn't that great.

While I was a bit harsh you shouldnt have miss misconstrued my argument that much. I was talking about mainstream for tri-core. Get rid of the dual-core all together and replace it with tri-core.

I'm not going to reexplain my reasoning for it all over again either.
 
Any thing in the middle is a midrange chip.obviously the tri core is faster than the dual core depending on the application.Why would the make the extra process to kill just one bad or slow chip.It is noted clock for clock the Phenoms out are slower than what Intel has in its class and for the enthusiast.I'm quit sure that AMD knows the position that they are in and making steps to get out.tri-core is an enthusiast chip and in less than 6 months we will find out.http://www.wired.com/techbiz/it/news/2007/09/tricore .I believe that the native gives AMD a more flexable range for ther cpu's such as single,dual tri and quad.
 
I was talking about mainstream for tri-core. Get rid of the dual-core all together and replace it with tri-core.
Actually, the more I think about what you're suggesting, the more I like it. BTW, you're not the only one that feels dropping the dual-core range and leveraging the tri-cores against Intel's dual-cores while letting the quad-cores battle it out, should be AMD's new strategy.

Last I saw AMD's roadmap, however, not only were dual-cores still alive for the foreseeable future, but uni-core processors were there for the near term, too. It appears AMD is thinking very different tactically, and probably perceives the need for much greater stratification in its product range. Hopefully, it will pay out down the line.
 
Actually, the more I think about what you're suggesting, the more I like it. BTW, you're not the only one that feels dropping the dual-core range and leveraging the tri-cores against Intel's dual-cores while letting the quad-cores battle it out, should be AMD's new strategy.

That would be one way to implode AMD faster. Sell big fat expensive dies downmarket for the low price of duals. Do you really think that is the way to reverse multi-hundred million dollar losses? It isn't.

AMD needs to sell smaller dies for higher prices, not larger dies for lower prices.
 
Back
Top