AMD Q4 results

[NTMBK]

No, I'm not. I just said I'm not. Are you calling me a liar?

No, I'm calling you wrong.

 

[ShintaiDK]

Why would GF need to beat Intel? TSMC doesnt need to. There is a perfect business for both. Just look at TSMC, 1/3rd of everything they make is still 180nm+.

The arabs just want a foundry business. they dont want to lose 25-50B$ just to try compete with Intel on Intels terms.

AMD is the only company not fitting in anywhere.

 

[Homeles]

No, I'm calling you wrong.

You're entitled to your opinion. Too bad it's meaningless unless you can defend it.

Go ahead, show me how I'm wrong.

Why would GF need to beat Intel? TSMC doesnt need to. There is a perfect business for both. Just look at TSMC, 1/3rd of everything they make is still 180nm+.

The arabs just want a foundry business. they dont want to lose 25-50B$ just to try compete with Intel on Intels terms.

AMD is the only company not fitting in anywhere.

Exactly. GloFo has the higher ground here.

 

[SiliconWars]

So do you.

Yes, that includes Bulldozer. Don't tell me that I need to read when you are seemingly incapable of doing so yourself.

Bulldozer was like a pile of saint's bones — a holy relic for worship, but fundamentally worthless. Bulldozer was met with cult-like blind acceptance, and asinine disregard for third party reviews. Bulldozer is the epitome of the statement he made.

What about Larrabee? At least AMD actually delivered BD. All we heard for years was how Larrabee was going to be the bees-knees and the intel fans held on to it even when it became crystal-clear that it was a total washout. Now it's moved on to intels HD graphics which every year fail to beat AMD's even with a process lead and ever more resources being thrown at them.

It also includes their original Phenom. Another tragic disaster from AMD.

Yes clearly it's only AMD who can release bad products. Intel cancels them instead. Well unless you include Atom, which finally starts to look not completely terrible 5 years later. I'm willing to bet BD's successor in the 5th year will also look quite good. Come to think of it, AMD made it looks reasonable within 1 year with PD.

And how could I forget about Itanic? You really believe that intel does no wrong? MOST of intels products are pretty awful, it's only their high end x86 that is any good.

I attack anyone and everyone if they're presenting inane, unscientific, and intellectually dishonest garbage, like the sewage you're spreading. AMD fans just tend to post the majority of it.

It's also filled with putrid leakage from AMDZone. For every member that attacks AMD or their fans, there is another member that has a shrine in their bedroom dedicated to Jerry Sanders.

In fact you have a severe problem that you should probably get checked. I could almost hear the seething in your voice as you spewed that hate-filled garbage out.

 

[krumme]

Why would GF need to beat Intel? TSMC doesnt need to. There is a perfect business for both. Just look at TSMC, 1/3rd of everything they make is still 180nm+.

The arabs just want a foundry business. they dont want to lose 25-50B$ just to try compete with Intel on Intels terms.

AMD is the only company not fitting in anywhere.

Well hello, welcome to a thread already overloaded with noise, so you seem to fit perfectly then.

When the marginal production cost for a dualcore jaguar is perhaps around max. 2 usd per piece, and you can still sell it for something, i would say your 25-50B is a little over the top for an attack on Intels high cost.

You ask why GF need to beat Intel?

Well the answer is, to start earning some money in the future. Because weak competitors is a way to do it.

 

[Homeles]

What about Larrabee? At least AMD actually delivered BD. All we heard for years was how Larrabee was going to be the bees-knees and the intel fans held on to it even when it became crystal-clear that it was a total washout. Now it's moved on to intels HD graphics which every year fail to beat AMD's even with a process lead and ever more resources being thrown at them.

There are certainly numerous failed projects from both companies. Larrabee just happens to have been publicized.

Show me where there were expectations that Intel was magically going to beat AMD and Nvidia at their own game. I doubt you could produce such evidence -- you're just making this up.

Unlike Bulldozer, Larrabee didn't have a bunch of people idol-worshipping it. The failed project was just that -- a failure. Bulldozer on the other hand was released to market, despite all the reasoning for not doing so. And despite the massive failure that it was, people effectively called all of the third party a bunch of liars for coming to that conclusion.

Yes clearly it's only AMD who can release bad products.

Would you kindly go shove that irrelevant straw man from whence it came?

Since you seem to have completely missed the subject at hand, Intel's shortcomings are irrelevant. This thread is about AMD's failures.

Intel can release bad products, and certainly has. Even Intel would admit that Pentium 4 didn't turn out to be the greatest of ideas.

Intel cancels them instead.

Because they have competent management.

Well unless you include Atom, which finally starts to look not completely terrible 5 years later. I'm willing to bet BD's successor in the 5th year will also look quite good. Come to think of it, AMD made it looks reasonable within 1 year with PD.

Reasonable, but still leaving everyone with an empty stomach.

I like the Bulldozer concept -- it's brilliant, but the initial batch of products were a tremendous flop. Had you actually read any of my postings on Steamroller, you'd notice that I have very high expectations of it.

And how could I forget about Itanic? You really believe that intel does no wrong? MOST of intels products are pretty awful, it's only their high end x86 that is any good.

Once again, you really need to learn how to argue. Making up positions that your opponent obviously doesn't hold just makes you look like a fool. In addition, the argument you are making is not even tangentially relevant.

For the second time, this is about AMD's shortcomings; Intel's shortcomings have no place here, seeing that despite their occasional failures, they are doing very well financially.

In fact you have a severe problem that you should probably get checked. I could almost hear the seething in your voice as you spewed that hate-filled garbage out.

I'd much rather be angry than be wrong -- like you.

 

[Phynaz]

Thanks for the derails guys. I've asked for this thread to be locked.

 

[SiliconWars]

Any thread with Homeles in it would need locking if derailing was a reason, because that is his #1 reason for being here.

 

[Homeles]

Any thread with Homeles in it would need locking if derailing was a reason, because that is his #1 reason for being here.

My #1 reason for being here is to challenge people's arguments. That is one of the points of a forum.

I've stated this multiple times.

 

[mrmt]

The moment Mubadala uses AMD - with production on TSMC - because GF can not do it now - to dump temash/kabinit/kaveri whatever on the market, Intels margin is easily under attack.

I think first you are overestimating Kabini market bracket, and overestimating AMD structure to take on Intel like that.

Kabini competes against Celeron/Pentium, and given current AMD margins there isn't too much to shave on gross margin. Shave 3 or 4 points in AMD margins and AMD will be in dire trouble, both cash flow negative and losses. Intel, as bad as it would be in a year of high CAPEX, they can withstand a price competition on this market bracket.

As for GLF be able to withstand this competition, I think there is a limit on how much Mubadala wants to put on this project. If they were really interested in AMD, they could simply not charge AMD everytime the WSA clauses were broken, but instead they charge AMD every penny they can. This is not the behavior of a company who has a war chest big enough to take on Intel in a multi-year multi-billion battle.

 

[SiliconWars]

My #1 reason for being here is to challenge people's arguments. That is one of the points of a forum.

I've stated this multiple times.

Declaring your own rules ie "This is a bash AMD thread not a bash Intel thread", then declaring victory based on that is not "challenging people's arguments".

 

[ShintaiDK]

Well hello, welcome to a thread already overloaded with noise, so you seem to fit perfectly then.

When the marginal production cost for a dualcore jaguar is perhaps around max. 2 usd per piece, and you can still sell it for something, i would say your 25-50B is a little over the top for an attack on Intels high cost.

You ask why GF need to beat Intel?

Well the answer is, to start earning some money in the future. Because weak competitors is a way to do it.

Great start there from you

I dont see any rationale in what you say. Jaguar seems to be something of a misfit between 2 segments. Jack of trades, master of none. And AMD is quickly becoming non existant in the segment. -37% YoY with no sign of recovery.

AMD is simply "another foundry customer". And the arabs did minimal investment into AMD to make sure they use GloFo and keep on paying. They already got their investment back several times.

It sounds like a rerun of all the nonsense when the arabs forst got the fabs. AMD wuld leapfrog Intel in processnode and there was no end to all the great things. What happend? AMD got its rearend smacked by the arabs that still mug them every single time they get the chance.

GloFo dont need to beat Intel to earn money, as already shown. Try lookup TSMC and see how much their 28nm for example account for.

If a Jaguar dualcore cost 2$, then a 3570K can be made for 10$ or less.

 

[Homeles]

Declaring your own rules ie "This is a bash AMD thread not a bash Intel thread", then declaring victory based on that is not "challenging people's arguments".

Dude, read the title. Read the context of the argument I made against amdisstaying.

It's not that hard. Everyone else in this thread seems to have figured it out.

I am victorious. If you have a problem with that, try disputing my argument, point-by-point, instead of the grasping at straws that you are doing now.

 

[SiliconWars]

Great start there from you

I dont see any rationale in what you say. Jaguar seems to be something of a misfit between 2 segments. Jack of trades, master of none.

It may appear to be that in some ways, however it will be sold in two clear markets. Temash is for high end tablets, and will be the only high-end volume tablet on the market. How much volume that is remains to be seen. Kabini is AMD's notebook chip for 2013. This is where all the volume is coming from. Believe me if it wasn't for the WSA AMD would probably ditch Richland altogether this year.

AMD is simply "another foundry customer". And the arabs did minimal investment into AMD to make sure they use GloFo and keep on paying. They already got their investment back several times.

This isn't really true. AMD carries industry experience the likes of nobody else does. Who else fabbed their own chips and also deals with the largest pure-play? AMD, and only AMD. GF has a lot to learn from AMD and you're forgetting that by and large, GF *is* old AMD.

If a Jaguar dualcore cost 2$, then a 3570K can be made for 10$ or less.

The problem Intel has is, the 3570K can't be sold for less than what it is else there would be a ripple effect on the rest of their CPU's. So they save a few bucks every process node? They are still stuck with the same market segmentation that will be their downfall.

Jaguar probably costs more like $5 each, and this is actually relevent to AMD, who can make $10 or $15 more on the chip assuming it's good enough. AMD can raise prices with better products, intel can't lower theirs.

 

[SiliconWars]

Dude, read the title. Read the context of the argument I made against amdisstaying.

It's not that hard. Everyone else in this thread seems to have figured it out.

I am victorious. If you have a problem with that, try disputing my argument, point-by-point, instead of the grasping at straws that you are doing now.

Yes Les I'm sure we call all figure out what the original Phenom and Jerry Sanders has got to do with AMD's Q4. :hmm:

 

[ShintaiDK]

What do you think AMDs revenue would be if they only sold Jaguars? A 2 issue wide core thats mainly an Atom competitor. Remember a new Atom uarch comes this year, first one since the start.

AMD is already down to only shipping ~13.5 million CPUs per quarter. And some 35-40% of those are already Brazos.

AMD is being the new VIA.

 

[Homeles]

Yes Les I'm sure we call all figure out what the original Phenom and Jerry Sanders has got to do with AMD's Q4. :hmm:

It's rather unfortunate that you don't understand the concept of "context."

Why aren't you calling him out for turning this into some AMD persecution sob story?

 

[SiliconWars]

What do you think AMDs revenue would be if they only sold Jaguars? A 2 issue wide core thats mainly an Atom competitor. Remember a new Atom uarch comes this year, first one since the start.

Ok I exaggerated a bit but I don't think it would be much worse than what it is now. It's plain as day that AMD is scaling down production on everything over 25W, which is the way the industry is going. I expect Jaguar to be 60%+ of AMD's sales in the 1-year period after launch.

AMD is already down to only ship 13.5 million CPUs a quarter. And some 35-40% of those are already Brazos.

Well there's a couple of things that nobody has picked up on yet. GF's 32 is very mature now, so dual cores like the A4's and A6 Trinity's will basically be non-existent in terms of Richland. We already know that AMD plans to use Jaguar as A4 and A6 this time around.

This is a big help, as the chips coming out of GF will mostly all be A8's and A10's, which is what is needed for such a big die. We can see from the current quarter that Trinity is holding it's value and shipments doubled.

So the current low margin A4's and A6's are being replaced by higher margin Jaguars, and the chips coming out of GF can easily pay for themselves.

 

[SiliconWars]

It's rather unfortunate that you don't understand the concept of "context."

Why aren't you calling him out for turning this into some AMD persecution sob story?

I would have done, and meant to, however it was easier to get sidetracked by your nonsense. I don't agree with the new guy's posting methods in a lot of ways, however he has a point on the direct attacks against him which are just ridiculous.

 

[Homeles]

I would have done, and meant to, however it was easier to get sidetracked by your nonsense.

Nonsense? Where? Is that the only rebuttal to my argument you can come up with?

I don't agree with the new guy's posting methods in a lot of ways, however he has a point on the direct attacks against him which are just ridiculous.

Well then perhaps he shouldn't have made a user account with a provoking name, and dedicated his first post towards attacking this forum.

He knew very well what he was getting himself into. He has no one to blame but himself.

 

[SiliconWars]

Nonsense? Where? Is that the only rebuttal to my argument you can come up with?

Well then perhaps he shouldn't have made a user account with a provoking name, and dedicated his first post towards attacking this forum.

He knew very well what he was getting himself into. He has no one to blame but himself.

A chosen user name - unless obviously offensive in a way that is generally regarded offensive to the majority of the population - should in no way open up a user to direct attacks the kind that this guy has been getting. Accusing him of being a future serial killer is extremely bad taste, especially due to recent events.

If anyone is offended by the name "amdisstaying" then I think that points more to a problem on their side as well tbh.

 

[Abwx]

Well then perhaps he shouldn't have made a user account with a provoking name.

LOLZ

Typical of a narrow minded AMD hater , the most ridiculous
and bad faithed argument i ve seen so far in this forum.

There are people with the acronym Intel in their name ,
yet no one did see them as being provocative.

 

[amdisstaying]

Eyefinity, I will put "Homeles" on my ignore list.

Back on-topic:

Intel is a tough competitor. That is a true statement.

IF I were running AMD, then I would pull JHH on Intel.

JHH wanted a court date with Intel. That date was December 6, 2010. Intel is also a paper Tiger. Intel doesn't want to face a Jury on a Judge. Nvidia and Intel entered into a serious negotiations. They also asked the Judge to postpone the trial date for 2 weeks.

JHH & Intel agreed for a settlement. Intel will pay NVidia $300 millions per year for the next 5 years. Intel and NVidia also signed some cross-licensing agreements. JHH will not give Intel NVidia's Graphics designs, but Intel can use NVidia's patents. So JHH got $1.5 billions from Intel. NVidia will not make any chipset for Intel. Actually NVidia decided to abandon chipset business completely. NVidia's main focus became the Tegra family of products.

AMD and Intel will negotiate their new cross-licensing agreement during 2014 or 2015.

AMD should not license AMD64 to Intel. AMD should ask for $1 billion in royalties for the next 5 years after 2015. AMD can take Intel to court and let the Judge or the Jury decide the case.

If Intel refuses to pay AMD $1 billion per year in royalties, then AMD should not license AMD64 to Intel.

There is no way Intel could survive without the AMD64.

AMD will be a graphics & ARM64 company. AMD can stop all X86 business. AMD will have only 4000 employees. So if AMD do what JHH did to Intel, then Intel will give AMD $1 billion in royalties every year.

A fixed amount of royalty can be set for each CPU. So x*5 is Intel's payment to AMD, where X is the number of the CPUs made by Intel. Y*5 is AMD's payment to Intel, where y is the number of the CPUs made by AMD.

Let us just say X = 300 millions, then 300*$5 = $1500 millions, that is, Intel's payment to AMD.

Let us just say y = 100 millions, then 100*$5 = $500 millions, that is, AMD's payment to Intel.

So AMD gets $1500 millions - $500 millions = $1000 millions per year from Intel.

 

[Nemesis 1]

It may appear to be that in some ways, however it will be sold in two clear markets. Temash is for high end tablets, and will be the only high-end volume tablet on the market. How much volume that is remains to be seen. Kabini is AMD's notebook chip for 2013. This is where all the volume is coming from. Believe me if it wasn't for the WSA AMD would probably ditch Richland altogether this year.

Temash is for middle tablets . Haswell will be the highend tablet / with better graphics . Some are confused about haswell. INtel designed for the mobile market not the desktop market not the server market . It was built ground up to be a mobile chip . Kinda like Dothan . except way more efficient. ULV is the normal for haswell the desktop chips will be ported to higher speed less efficient process. As for Larrabbee being a failure lol. will see how many knights corner sells this year. likely about 5 billion worth.
Its been hammered home now for 2 years by intel . haswell is designed from ground up as mobile. Temash will have to fight it out with tegra4 in the middle tablets . the highend is all intels. Intel has to price within the market it aims at . even if its always the highend stuff. But intel has shown a willingness to enter low margin makets with the lexington phone . so saying intel will not price to win the game going against both AMD and arm . Intel hands are completely untied. Intel will price to win . The only company that can go toe to toe now on pricing is samsung. I wonder what it cost for a foundry to produce a chip. Than I wonder what that foundries margines SHOULD be . Are fabs in business to lose money for AMD /NV/ Apple ect.ect.ect. Intel can and samsung can out price all others legally. This is no longer an x86 war.

 

[Ajay]

What do you think AMDs revenue would be if they only sold Jaguars? A 2 issue wide core thats mainly an Atom competitor. Remember a new Atom uarch comes this year, first one since the start.

AMD is already down to only shipping ~13.5 million CPUs per quarter. And some 35-40% of those are already Brazos.

AMD is being the new VIA.

Well, things have certainly been going this way since I first read your prediction (AMD xforms into the new VIA). 2014 might change that trajectory a bit (new big core server CPUs), but that's only if they are still solvent.