Sgolds -

Yes [Opteron and A64 are 2 different die]. This is an area of confusion because they use the same core. (A lot of published garbage out there confuses 'core' and 'die' - much different things, of course!) Opteron and A64FX use the same die - and even the same package, temporarily. Socket 754 A64s have less cache, only one memory controller and only one HT. One core, two dice (or, some prefer, dies).

Now that we have established that you believe they are 2 different die, how about showing where AMD, or a reliable source says they are 2 different die. It is quite possible to disable portions of cache and aHT channels and repackage an Opteron die. You are very free at claiming facts but usually short on backing them up. I admit I have no data to prove my point but you've provided none either.

Sgolds -

Taiwanese company - you must be referring to UMC. That collaboration fell apart when UMC proved to be wholly unable to meet commitments. (They tested the deal with bulk Bartons, which UMC screwed up.

Ironic to hear someone from AMD complaining about another company failing to meet it's commitments. How about a reliable source to backup this accusation?

Sold Nov $13 Puts @ $0.50

Bought AMD shares with the premium.

Sgolds -

Elmer, Athlon64 and Opteron are two different product lines.

Ok, clear this up for me. Are A64 and Opteron 2 different die? I believe the answer is no and therefor AMD has been producing them since early in the year. A different package does not a large difference make. I don't understand why you're trying to draw a distinction here. Everyone seems to concede that A64 will not hit volume until 90nm. You seem to think 90nm will appear in Q1 and that's where the differences of opinion lie.

John -

SOur grapes, huh? AMD only had 3 weeks of pilot production for Athlon 64 and got "10s of thousands,"

Only 3 weeks of pilot production? What are all those Opterons? They've been making them since the beginning of the year.

Doug -

Are you sensitive tonight because your AMD covered call strategy doesn't fare as well when the underlying gets called away, or what?

You're right that my CCs don't do well in an up market but I can roll them up and out and still make a little money. But that's why I write the puts as well. They can't both fail to make money. Trying to sell some Nov $12s or $13s as I type.

I'm happy for those who have made money, despite my sobering views.

wbmw -

my impression from Microprocessor Forum was that networking processors is not the way to an Intel sized business (see my other response on this topic).

Actually I was referring to the entire Comms market, it was just the Network Processor subject that prompted my response.

ChrisC_R -

And what do you think might happen when AMD's 64-bit products go from 10Ks units/quarter to 100Ks units/quarter? Never mind, who needs one of your rude, sarcastic, obnoxious responses.

You remind me of someone else here who says he doesn't want to get in a pissing contest meaning he wants to be the only one who gets to piss...

To answer your question, if AMD can actually produce A64 in volume then obviously they will do quite well. So far they haven't and so far they are still losing money although the bleeding has slowed. The real question is did Intel reach capacity constraint? They shipped a record number of processors and chipsets while a lot of capacity is offline. Did AMD benefit from Intel's shortcoming or was there a real demand for AMD products in their own right? If the AMDroid prayer of Intel having 90nm problems turns out to be just another line of theinquirer nonsense then a huge amount of capacity is soon to come online and AMD may not enjoy the role of backup supplier when Intel runs out. Speculation and musing of course. We'll see.

Doug -

And Intel was unable to ship any P4 EEs or Prescotts. What does that tell you?

It should be obvious. Intel can make a ton of money without even shipping their highend product. That tells me something. What do you think will happen when they do? No hints now. You have to figure this one out for yourself.

morrowinder -

Intel gains ground on AMCC in net processor rankings

There is enormous potential here. Intel has been absorbing quarter after quarter of losses in their ICG group while they establish their commitment and credibility. I listened to a speech a couple of weeks ago by Sean Malony when he said Intel has the products, Intel has the design wins and all that is left is to execute. The customers really want Intel as their supplier. Now Intel has to deliver. I'm betting they will.

NaS -

maybe it tells me that Hector is more cautious than Jerry? I have no reason to believe it is a manufacturing problem, yet.

I'm sure you're right... They just didn't want to make money this quarter so they didn't start all the wafers they could have. Just being cautious. Don't want to make money unless you absolutely have too...

NaS -

it's a shame we have no way of knowing the yeild on SOI at dresden...

AMD was unable to ship more than a trickle into a demand driven market. That should tell you something...

Sun Microsystems Loss Widens; Sales Slide
22 minutes ago Add Business - Reuters to My Yahoo!



SAN FRANCISCO (Reuters) - Network computer maker Sun Microsystems Inc. on Thursday reported a wider quarterly loss than a year ago as weak demand for its servers caused sales to fall for the 10th consecutive quarter.

http://story.news.yahoo.com/news?tmpl=story&ncid=1208&e=1&u=/nm/20031016/bs_nm/tech_sunm...

Keith -

Much better results than I expected. Good luck to you AMD longs.

confused -

can this be true?

It probably is true but not because of who reported it. The inquirer is the least reliable source of information out there, but that isn't any guarantee that the info is wrong :)

sgolds -

No time for pissing contests, so I will pass on further responses.

Translation: You want to get in the last piss..

sgolds -

Elmer, your revisionist history really contributes little to a rational discussion. Really! Austin was converted to flash well after Microsoft jumped ship on the original X-box design.

Well no kidding!!! Really??? So if they had to keep F25 running an old logic process then they'd need to build a new fab for flash expansion, right? No problem right because they're rolling in cash!!

Microsoft changed for cost reasons: They figured out that they were going to be shipping a C-note with every unit. Thus when Intel cut a price deal for Celerons, Microsoft bit. AMD refused to match price.

They aren't Celerons, they're CuMines and AMD didn't match Intel's price for the very reasons we've been discussing.

Try and follow along will you? You're missing a good conversation.

sgolds -

What other packet switching busses are there? Hmmm...

PCI-Express..

NaS

hmm, doesn't that equal the "INTC took it on cost" idea I floated to begin with?

In a round about sort of way it does. The claim has always been that Intel did it to screw AMD but I think if they really wanted to hurt AMD they would have let them have the contract.

Chipguy -

Other than their ubiqitous 1164/1165 high end FP chips in
the late 1980s Weitek was most famous for selling a clock
doubler SPARC chip to upgrade old Sun workstations.

That's the one. PC boards sometimes came with a socket to accept the Weitek.

NaS

since durons could and were being produced at the other fab (not dresden) AMD had another fab that was underutilized ready to pump out whatever Durons would be needed for xbox for a long time to come.

Maybe you didn't notice but the Austin Fab has been converted to Flash. The point is that AMD had no commitment to keep a process alive and active for years to come. I'm sure MS required that. Intel is still making CuMines as we type and will for years. In my view, Intel should have pulled out and let AMD have the design. AMD would have been stuck producing low margin processors on an old process. They could have contracted it out to a foundry but that takes lots of resources to convert it to a new process and you have to share the profits. AMD must have felt they had better use for their Engineering resources. Meanwhile Intel always keeps a process alive so they could just keep cranking out CuMines like jelly beans with essentially no resources consumed. No sharing the profits with a foundry either. Looks like it was a no brainer.

ChipGuy -

Wasn't it Weitek that did the same thing years ago?

NaS -

what's more reliable than millions of chips sitting around unsold?

A factory to produce more when the supply runs out.

Jerry R -

Please note that most AMD fans will never seem to care whether or not AMD makes money. Their Intel hatred allows them to rationalize the guiding principle to hurt Intel at any cost. AMD's management encourages this bunker mentality among the AMD faithful and feeds wildly optimistic projections to keep them happy and energized. Ignoring history, AMD fans can hope beyond hope that AMD's management will hold to their predictions this time.

You are absolutely right. I've tried to point out that you can't understand the AMD faithful without considering the psychology involved. But consider that many are AMD employees and they have a vested interest in believing. Battered wife syndrome. You can't really blame them. Had AMD had a business plan other than "hurt Intel at any cost" they might actually be a successful company by now.

A few people in this world will always cheer for AMD, regardless of their financial performance. It's simply blind faith driven by hatred for big bad Intel.

Why do you suppose they hate a winner? We've seen it many times. It used to be everyone hated IBM because they were on top. The fact is whoever is on top will be hated. It's a natural law.

Even if Intel has pushed their 90nm production to Q1, what material effect does it have? Intel still predicted a strong Q4, potentially their largest revenue quarter in its history. If AMD slips 90nm production to Q3/Q4 next year, what effect do you think this will have on their finances?

The point is that if Intel slips their 90nm production to Q1, and still makes a ton of $$$, they are a failure if they don't hold the interest of the Geek gammer websites. Remember success means you blast the most mutant aliens, not that you make money. Failure is losing to very low volume parts running selected gaming benchmarks.

What if AMD slips 90nm to Q3/Q4? Who cares? You think this is an investment forum?

NaS

INTC stole that one on price not on availability.

Sour grapes. MS went with Intel because they needed a reliable supplier.

blauboad -

I can quit anytime I want.

Me too. I've quite hundreds of times...

UnD

Still pondering that extra day. How about this?

What about it? Nice design win for Opteron.

UnD -

I like a man who's willing to make predictions! It's all about making money.

Keith -

That is remarkable growth. And Intel says the US market has yet to take off!

Sgolds -

AMD plans 90nm to be a two-quarter transition, just like their previous generations. Intel takes well over a year (maybe 6 quarters) to get all their x86 product on to a new geometry. So AMD will beat Intel to the 90nm volume crossover point, even if Intel ships first.

This is not true. Intel has never had all it's x86 products on a single process generation. There is still demand for older generations of processors in many embedded designs. This is another area where AMD loses design wins because companies know they can't rely on a continuous supply for year after year. At Intel the older process generations are fine tuned, the equipment is depreciated so they serve to produce chipsets and other products that don't quite need cutting edge technology. In assembly and test the equipment enjoys a longer lifetime too. This is a tremendous advantage for a company to not have to keep scrapping used equipment and facilities because they aren't useful for bleeding edge products.

Sgolds -

Every time someone points out that Intel seems to be following AMD in technology, you change the topic to financials.

What was your point, that AMD is ahead in technology? I bet you base that on something from theinquirer right? I see Intel producing leading edge products in high volume with high margins. I see AMD producing mostly lagging technology (XP) while managing to ship small quantities of leading edge products which suffer from serious yield problems. Your idea of who is ahead of who is based on who can introduce the best low volume product regardless of whether or not it can be produced commercially at a profit. If profits are your goal, as in the case of Intel, then you only push the technology as far as you can reasonably produce. You seem to think that's all that's possible for Intel but you fail to realize the difference between the two company's philosophys. Is Intel delaying 90nm and if so is it because of a problem? You naturally conclude it's yes to both because you embrace AMD's philosophy of push everything to the ragged edge and promise the moon regardless of what you can deliver. You assume Intel must think the same way but you fail to consider that they are under no pressure whatsoever to hurry when they are making tons of $$ with their current technology. AMD is not challenging them with their technology in the market place. Maximizing profits is the guiding philosophy at Intel and you aparrently can't conceptualize that. Whatever Intel is doing >>it's working<< and whatever AMD is doing >>isn't<<. You always want to avoid the financial failures and try to make the measure of success something that would best be discussed in academia where business issues play no part. It's the beauty of the idea there, not the profitability of the company.

Sgold -

That was the second proposed Mustang design. First, dual core. Then that was dropped in favor of a large cache design. Finally the whole thing was dropped as a solution looking for a problem.


Ahhh yes, that was the excuse after long delays. Others pointed out that AMD had never produced a large cache design and it was almost certainly a manufacturing problem.

UnD

Wouldn't you want to sell the puts just after such an announcement, if it does indeed arrive soon?

Sure. Are you making a prediction? Be very wary of AMD's product announcements when they are also announcing losses. Anyone who has followed them for long knows this pattern.

Alan -

I will believe in 90nm from the various companies when I see some products for sale. So far all I have seen is words and guesses. The last data point we have is 130nm, where Intel did indeed lead by almost a year.

Myself, I consider a companies track record just like I would consider an individual's. If I know someone has a history of delivering what they promise then I would be much quicker to believe them than someone who makes promises but rarely keeps them. Intel has a track record and so does AMD.

Sgolds -

Intel is not doing a very good job of leading.

LOL! You guys are too much! You've made up a mythology of your own, agreed amongst yourselves that it's official reality and draw conclusions from there. The reality is Intel posted a great quarter with record processor shipments, record chipset shipments and near record profits. Q4 looks even better. The reality tomorrow is AMD will post it's 9th straight quarterly loss on it's way to 10. They'll promise profits are right around the corner and you'll believe them again. It's the battered wife syndrome.

In my view AMD is doing a very poor job from a distant 2nd place.

UpNDown

Are you expecting AMD to announce an FX-53 tomorrow? With SPEC results? Big splash? Best results for any shipping desktop processor? Later in October, later in the year, or never?

Why ask me? How would I know?

Sgolds -

They once had a design dubbed Mustang, which was originally proposed to be a dual K7 core processor.

Mustank was a large cache Athlon.

AMD sets date for son of Opteron

http://zdnet.com.com/2100-1103_2-5091845.html

Chips based on the K9 architecture will likely be released--at least in sample quantities--by the second half of 2005, Weber said.

Anybody actually believe this?


AMD is also looking at adding threading to future chips. Simultaneous multithreading essentially allows a chip to run two applications, or two "threads" of the same application, at the same time--thereby reducing the time it requires to complete a task.

Once again AMD plans on copying Intel's ideas. Can't these guys do anything original? First they copy Timna with it's ondie memory controller and now this!

YB -

To be sure you'll have to go back and listen to the CC but they did say they gained market share in the value segment. I didn't listen to the entire call. It's possible they gained share across the board.

Buggi1000 -

For me it is given, that AMD should reach 10% unit-
shipment growth - sounds crazy, but that is my feeling!

Could be but their ASPs may have suffered to do so, so I don't thing you can necessarily infer a revenue increase.