Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
keith
Fujitsu is committed to Sparc, Sun is as well, IBM is committed to Power, HPQ is committed to Itanium.
All of them are global companies, but still it is two cultures dealing with committments in a different way.
K.
cg
...narrow price, performance, and application range band of x86 PCs and servers.
U r kiddin, aren't you? ;)
K.
"It is amazing the bad press spin that Intel faces"
Haven't seen this kind of complaints for just too long. ;)
K.
grimes
It is neither superior to Floating Gate nor to MLC-technology, but cheaper to make.
The effects are not yet really visible, FASL really only has begun migrating to Mirrorbit last year. Expect increasing revenues and linearly increasing profits every quarter for the the next two years, while they are ramping 130nm mirrorbit in JV3 and migrating to smaller geometries and Mirrorbit in Fab 25, even if competition remains fierce.
K.
mas
Absolutely d'accord with your last statement. And I might add it is a good (Intel-IR) idea to shift the focus of perception of Intel from µP to global growth opportunities leveraged by Intels core-competences (marketing, manufacturing, technology, in this order) to keep Intels investors confident in the Common even when earnings will slide or drop - for a while at least.
Flash counterattack does not worry me for FASL at all - neither from a price - point of view (NAND pricing is more what FASLs pricing will be oriented at competitively) nor from a technological point of view. Whatever node it is, Floating Gate is moribund. If Intc had Strataflash already on 90nm, I would consider it a threat, FG @ 90nm is not.
Aggressive pricing by Intel will only accelerate the exit of other NOR-manufacturers - except for some niches.
K.
Replying from AMD-64 notebook (WAD 2000)
Which model is it?
Which OS?
Fan always on (and annoyingly loud) like on mine?
Resolution and quality of display?
K.
keith
yes, interesting. A very nice escape from the nanosecond thing: "Intel is making it as well". Semantically implying "....already, now, as we speak". Smart.
By saying "Dell will offer X86-64 bit servers even before software will be available for it", Mikey again invented another Dell-language-term of time. This one is a lot harder to argue against than the nanosecond one.
Now, coming back to our recent exchange: Do you understand these kind of (lukewarm) statements as product-announcements?
K.
Keith
But the statements from the OEMs suggest to me that they have reason to believe INTEL won´t delay Nocona another 7+ months
It won't. They will do the very same as they did with EE.
In Serverspace, it is even easier to play this game. Barely anybody would be crazy enough to use brandnew server hardware without having evaluated it before, so every Server-CPU is extremely low volume for couple of months. Call it a beta - release.
Then, if it is good enough to run 32bit OSes and apps flawlessly, bout 90% of users wont complain and trust Intel that it can handle i32e as well if they would want to use it.
It probably wont be able to do so, but when it comes to that (in two years time?) Intel will certainly have something that can.
We had these things before, i just remember the Pentium Calculation error, the 1,13 GHz part. No big deal to replace couple of thousand CPU for Intel.
Basically, it is business as usual. (yawn)
K.
Keith
As I said, short memory, I dont remember which OEMs did announce it publically. That is why I asked you.
However, what I remember clearly is Intel announced Prescott for H2 last year.
As for Nocona in Q2, i would think I already posted very clearly what I expect and what not. (The fact only Nocona has been announced as i32e (and not Prescott at the very same time) tells a lot about where Intel is to anybody familiar with manufacturing models of the industry: Exactly as you posted it, the IDF-message I received clearly is another emergency edition-announcement)
K.
Keith
All the big OEMs have said they will offer products based on it this year
Sorry, I have a rather short memory. Could you remind me what all the big OEMs said last year about Prescott-Systems for Holiday Season? And what did they say about Dothan-based Notebooks? ;)
K.
wmbw
My view is simply that investors need to be pragmatic about the risks involved, and not assume that they can ramp up K8 as fast as their roadmaps indicate
Sure. However, I would be neither surprised nor disappointed if 90nm would slip another quarter (or even two) for the following reasons:
1. I would expect 130nm beyond the scope of this year - no performance penalty.
2. It could even be favorable to run a mature 130nm a couple of months longer if yields are fine with it and if 90nm-process is not mature enough (yield- and/or designwise) to leverage economic effects from smaller dies.
But then, considering AMD had AO Silicon in November last year, volume ramps in the third quarter seem to be feasible if no major setbacks occur. Actually, for AMD which was usually forced to ramp immature processes for competitive reasons, this is unusually comfortable timeframe to be at volume. In essence, AMD will be able to choose when it is favorable to migrate within some six months-window just from its own processes and not for competitive pressure from Santa Clara. As I said, unusually. Would think this will benefit Gross Margins couple of points.
...Intel's problem being with the 90nm process, that means they have the potential to really turn things around in the short term if they fix their issues Intel's problem being with the 90nm process, that means they have the potential to really turn things around in the short term if they fix their issues
That would depend on the issues. Talking about leakage: If it is material-related, it is a matter of luck if they find some other material and still a matter of time to implement it processwise. If it is transistor-geometry-related, oh boy, no short term fix then. So, what you say is possible, but completely unpredictable for me.
Low power Nocona parts are a definite possibility.
I cannot argue against it, as you neither post what low power would be not specify a timeframe when that would be.
From the comments made above it is still possible, but not exactly highly probable we see Low Power Noconas anytime soon.
K.
Darbes
If the agenda Goto published recently is not completely bogus (and many things what we saw and see would fit into his picture), it would appear to me that iT has way to go for a launchable product.
Looking at the Nocoma announcement for Q2 from that viewpoint, it would look like nothing much more than a FUD shot.
I am inclined to this view of it - consequently I would be very surprised to see much more than samples of i32e this year.
Say, couple of thousand CPUs.
In case I will be proven wrong on it, I would be nothing less than astonished if the product would not be plagued by incompatibilies of any kinds, bugs, worms, louses, flee and the kind of problems you expect from launching a product which is simply not ready for the market. Let alone competitive performance - at least for the AMD64 part of it.
In case I will be proven wrong on that as well, I will gently lift my hat in Intels direction - before eating it. (I promise to post a picture of it here by year end, if it will happen.
Hugh
K.
wbmw
Fast forward to today, and 64-bits is still irrelevant on the desktop, not to mention mobile.
Assuming you are referring to current applications, I agree.
However, from an economic viewpoint 64bits are relevant at the time they play a role in buying decisions.
This will be the case until at least Longhorn, which will service >95% of desktop OS market share, supports it and the required drivers. Longhorn is what? 2006 now?
I agree on this navigational waypoint. So, second half of 2006 is about after half of the useful life for a System purchased this year. Which makes a perfect leverage to switch over to a 64-bit operating system by then.
Thats why I believe 64-bit is highly relevant already today for educated users - I would hope this rationale will kick in for corporate purchases already this year. For the consumer side I am confident that as word spreads that 64 bit promises benefits for gaming and video stuff you bet 32bit will become a no-no very soon for gaming kids. As soon as that, publishers will have little choice to tout this horn.
WRT to elmer, i sure hope he will be continued to be proven wrong with his hypothesis that AMD will fail to deliver on the manufacturing side (again) with K8. They did not in the current node - although the progress is far less impressive as with their 180nm copper process in 2000 when they doubled clockrates within a year or so. Very much more unimpressive on the yield side, but that is no miracle with more layers and larger die on SoI.
Actually, i guess elmer would have been pretty right on the manufacturability of K8 (at least in the current node) if AMD had not designed lots of redundancy mechanism into K8 - what elmer did not know, at least he never mentioned this turf even when asked about.
K.
OT Tenchu
Thanks. About the age I was when I had the same conviction. ;)
Only when Intel succeeded I understood what one of my academic teachers told me long before: "Always keep in mind, in economy only change is sustainable."
K.
Techu
...the only one that matters is the bottom line.
If you don't believe me, try opening up a small retail shop right next to your local Wal-mart and see if you have any success.
That is what I was convinced of before Intel took the PC-CPU market from Big Blue.
K.
p.s.: May I ask how old you are? (PM appreciated, just curious).
keith
7,5M Server-CPUs in 2003 sound plausible. Would imply that machines counted as servers use Server-CPU versus Non-Server-CPU in a 2:1 ratio.
K.
Quote of the day:
We believe Itanium is an endangered species. Ashok Kumar
http://www.mercurynews.com/mld/mercurynews/business/8035675.htm
(Sorry if that has been posted before, the source says it was in a report from yesterday.)
K.
wbmw
I can well imagine the scientific community (better said, some part of it) is delighted to port their apps to IPF. These guys do neither have to care much of the time needed to do it nor do they get the bill for hardware.
As for professional apps, that is different. These guys have to think about the relevant market for their ports. Given it would require same effort to port to IPF and to AMD-64, what would they do first?
K.
wbmw
If these programs are used now,
and are available in EPIC format,
and these programs behave like Spec-fp relations performancewise,
and the last 10% of performance do really matter,
and costs for Hard- and Software do not matter much,
then this would be my decision as well.
K.
wbmw
If I were buying a CAD/CAE/CAM workstation for performance, Opteron would not be inside.
Which CAD/CAE/CAM software do you use currently?
K.
sgolds
"(but it might make some dandy flash)."
As long it is floating gate products, there is nothing to worry for FASL, which plans to be at 110nm Mirrorbit few months later than Intel plans to be ramping 90nm floating gate (same wafer size). K.
ChrisC_R
Lol - Yeah. K.
Re: The Street Knows what's going on
If the street had any clue of what's going on, wouldn't we see money flow from Intc to AMD? K.
Thanks <eom>
Keith
Thanks for your take.
"It´s up to AMD to change that view, if a change is justified."
In a way, i would agree. Well, see what Bob has in his pocket for next week.
K.
[edit] Keith
"This is what analysts are seeing, and one reason why UBS reduced the target estimate to 17$ exactly one month after they moved the target up to $20."
You are serious with a claim Tom Thornhill would be unaware of waterfalling semiconductor pricing?
Assuming you are aware of it, let me put it this way:
Where do you see the XP sweetspot (sales) right now?
K.
(scratching head about continuing deliveries of XP1700+ CPUs in the 30 to 40 Dollar price range. That is really the background of my question).
maybe a new accounting scandal?
That is not what IRS is looking for. Just trying to milk out a lil more of the cow. K.
Keith
while Athlon XP pricing continues to drop.
Are you saying XP ASP drops or even "continues to drop"?
K.
elmer
Bye for now. Take care. K.
Dan
AMD's sales were up 100% year on year in that market.
That is the official numbers. Before that, India was a traditional place where many many grey market CPUS from Asian OEMs ended up, which were counted somewhere else.
Now, as AMD managed to clean this mess up, Indian resellers can hardly blame AMD for that. So they blame AMD and its distributors for something else.
Very understandable, but no action required from AMD here at all.
K.
smallpops
Thanks much for digging up and posting this link.
While everybody has the right to be wrong and consequently there is nothing wrong with changing minds, Gartners analysts are cheeky enough to start their new assessment:
"Gartner has long believed that a market opportunity exists in the space between IA-32 processors and Itanium server technology."
But well, from a biological point of view, opportunistic strategies are successful for parasites.
K.
joe - Great Post!
Also, it is not clear if the 2 of these fabs can supply all of the x86 microprocessor demand
Yes! Intel is not sure either; that's why they build this third (spare) fab in Leixlip. It was about time somebody raised this question here.
K.
Upndown
I would love to share your take on it. And I appreciate MSTs move to make it public now as a preview, _not_ public beta. ( taking half an hour to read through the Beta-NGs gives a good idea of why that is).
Dont get me wrong here, everything is fine - anybody who went through different OSs would not expect anything else.
K.
I_banker
The only thing holding me back from buying is the release of socket 939.
I see. And SG-Stepping, which is the other side of the very same medal.
If you are lucky, there will be Socket 939-boards available with a complete set of 64-bit drivers. Which is the bottleneck currently.
K.
Windows AMD-64 beta is public now
http://www.microsoft.com/windowsxp/64bit/downloads/upgrade.asp
K.
(yes i know it is called "preview" and "Windows XP 64-Bit Edition for 64-Bit Extended Systems" but I cannot spell it - let alone remember for longer than five seconds).
yb
Are u sure? Notebookmakers see demand for the worlds most powerful notebooks: http://www.digitimes.com/NewsShow/Article.asp?datePublish=2004/02/02&pages=A7&seq=41
K.
Keith
Sun hasn't yet built much of an x86 server business. As of the fourth quarter of 2003, the sixth quarter after Sun launched its first Xeon servers, the company still hadn't made it into the top 10 shippers of 32-bit x86 servers, according to preliminary data from market researcher Gartner. No. 1 HP sold 442,976 of them, while 10th-place Hitachi sold 7,340, Gartner said.
http://asia.cnet.com/newstech/systems/0,39001153,39166961,00.htm
K.
doug
Increased BIST capabilities make more sophisticated redundancy concepts possible to improve yields. Consequently, many "black" transistors most probably are to be explained as being part of redundancy-mechanisms as well.
Looking at 300mm, 90nm, LGA-package and maximising yields by means of enhanced redundancy mechanisms, Intel is on an uncompromised cost strategy. I would not be surprised if they can drive down costs by far more than 30% which you would expect from 300mm/90nm economies alone. Could even come closer to 50% after the migration is completed next year.
By end of next year, Intels ASP could well be a two digit number - without gross margins going down at all.
K.
[edit] Steve
I dont think there is any lack of talent at all at Intel.
It is just they have the philosophy of running a company according to an uncompromised money-machine-approach: Basically they are just researching for common intuitive misperceptions and capitalize on it by means of products delivering what people perceive as value, leveraging it by means of branding and all instruments of the marketing-orchestra. Playing all these instruments together to make a symphony for the public is the core excellence of this company.
Iaw they dont care a d..m s...t which benefits their products bring from an educated (better said merit) point of view. What they really care about is making their products as cheap as possible and get as much as possible for it.
You bet Intel engineers know pretty well P4 is a deception. But it makes tons of money and pays for salaries and stock-options.
K.
btw: Accordingly, Dirk Meyer is a great architect. But I somewhat suspect the background of Fred Weber (of physics and theater sciences) is a lot more important for AMD, looking at how K8 has been stage-managed for the public. Now, there is professional variety show offered right next to Intels old-fashioned operahouse.
yb
Is it the same as Newcastle? I hear talks from last year and no meat so far...
CG has nothing to do with Newcastle.
Newcastle is here since last year. It is just not what you expected - for the very same reason you thought Xeons are something else as P4s (or Celerons, btw).
K.