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.
sgolds
Dothan would seem to be a great design for making dual core - it is designed for much lower power consumption
Yes. Next year.
For servers... it would require a new 32-bit platform to be introduced next year. I dont think the market will adopt that.
Intel's market share is large enough to do both.
Sure. Its just....If they decided to make a X86-64 bit Tejas and a dual-core dothan successor, how could they communicate where they are heading to?
Whatever they say there, it would make Athlon64 look like a panacea.
K.
elmer
Now who is acting on those predictions? If you really believe your prediction then how have you positioned yourself to profit?
My money is where my mouth is. Uploaded stocks, partially on margin. No options involved at this point, as I believe to have a grip on how things will develop - but not on when it will materialize in the stockprice. (The herd is walking too often for too long in the wrong direction.)
K.
elmer
50. For the record.
Happy New Year everybody.
K.
sgolds
Couple of things are to be considered: Changing horses to Dothan and successors for Desktop and Server would mean giving up Hyperthreading and PNI. Both features are not really important imo, but you know it is already in all these benchmark-applications, so it will hurt.
Second, Pentium-M-architecture is not exactly new from scratch, I dont know how far Intel could still scale it.
But most important, they would close the backdoor to migrate to X86-64, which many (including myself) believe is in preparation in Prescott.
Insofar I dont think this path is leading out of the land of misery.
I would rather expect they push LGA 775 as much as possible to get sufficient power into Prescott to scale it and throw BTX and Copper on the heat.
NP. People will like 4 Gigahertz without having any clue as they like 2,5 Gigaherz.
doug
Motherboard or socket, whatever the bottleneck is to get enough power into the CPU.
K.
elmer
You know what was meant is packaged, not repackaged, dont you?
K.
yb
The current Prescott socket requirements called FMB 1.5 support chips up to 3.6 Ghz. So, we won't see 3.8 Ghz P4 until Tejas comes.
As you are talking about socket requirements, ... not before LGA 775, would you agree?
As for the rest, your posting sounds like you have a high level of conviction of its content. Is it a synthesis of various sources or from a source you cannot quote?
K.
paul
Why would they want the extra packaging costs and the problems with interdie signals, when they can do it all on one die with lithography?
Because lithography is only a small fraction of the costs associated with a new process.
MCM is clearly an option if AMD will be forced to offer it before its second generation 90nm node (05/06), when dual-core dies will be creating enough volume to rectify a process. (Iaw, in case Intel would have a dual core Dothan-based Xeon offering ready next year, I expect to see that.)
I do not see a dual-core Prescott to be possible thermally, so I guess 05 will be the earliest date for dual-core CPUs. Rather based on Dothan-successor than on Tejas. And more likely on 90nm design rules than on 65nm.
K.
yb
Regardless what they said or what Mike has been told, looking at street-prices and at number of distributors and their prices for A64-systems (which at least give you some indication of what OEM pay for the CPU) is a way better base to guesstimate volumes by means of price elasticity.
If the latter did not change significantly during the recent months, Elmers dimension for AMD-64 CPUs sold for revenue is definitely the ballpark the game is played in currently. I dont remember AMD stating expected numbers. I remember them stating K8 would not be expected to contribute significantly this year already. In my understanding, that means not exceeding single digits percentagewise for Q4.
Which could be a slight understatement, i would concede. As well as expectations given to mobo-manufacturers tend to be overstated for obvious reasons - you want to have platform-competition out there.
K.
elmer
What else could it possibly be???
Inventories. Takes a couple of months to realign its structure while migrating to a new architecture. Requires a tough grip on prices, which requires a tough control of channel-inventories as well.
All that is happening right now. The latter things are obvious, the inventory structure alignments invisible- although there are hints how things are going there; introduction of Model64-3000+ before Christmas carries a message thisrespectively. Next message to be expected with the pricing-round early January.
K.
deleted - obsolete
andi g.
Not all of them.
http://finance.yahoo.com/q/it?s=INTC
K.
yb
the next big stuff will be Transmeta-like on-the-fly emulator, where hardware instruction set is not part of the standard at all and the manufacturer is free not only to add new instructions, but also to drop the old ones without any compatibility problems. Both AMD and Intel will have their own instruction sets and emulator will do the job.
This is an interesting vision, as it is just widening the "Hardware follows Software" - approach currently communicated from Sunnyvale.
It could materialize if hardware-capabilities will stay far ahead of software-needs as they are today for most segments of the market allowing to sacrifice ressources for all sorts of emulations.
Personally I do not expect this to prevail, although I do not see killer-applications on the horizon to change this in the near future.
As for "without any compatibility problems", this would be very, very surprizing to me. Within the last two decades, I never bumped into any emulation close to be without any compatibility issues.
On "good article", reading Ed Stroligo's pieces since a couple of years now, I dont know if I should be worried about him or if he decided to make some Faustian pact some time ago. (I tend to be worried, his last piece was completely unnecessary and hardly coordinated with Satan Claras editor.)
K.
paul
If Intel were to introduce an extended AMD64, with some extra instructions, would the licensing agreement give AMD full access to that?
Sure. Instruction-set issues between Intel and AMD are not a matter of licences, just a matter of timing: Those "extra" Intel instructions would have to be implemented into AMDs architecture - this can only be done in the next generation design - or even the overnext.
MMX, SSE, SSE-2, PNI, all the same pattern. AMD-64 as well btw, the other way around.
In your context, if Intel would introduce a X86-64 CPU in 2004 with additional instructions, these could probably be implemented in AMDs second-generation 90nm-design. If Intels X86-64 comes only 2005, that would be already too late for AMDs next generation design - assuming Intel would not let AMD know earlier what exactly (and completely) they are doing instructionwise (which I would not count on, as they did not for previous instructions).
K.
Semi
so I don't understand what makes the Inquirer say those things
Mike was explicitely quoting the source of his piece - even with a link to it.
Then he did ask Intel for confirmation and posted the denial he got on it.
I dont see anything to complain about. K.
elmer
Remember what happened in the late eighties with X86.
See the pattern?
Merry Xmas
Klaus
Merry Christmas to everybody - eom
yb
Yes. Dual cores will be the next thing. About clockspeed, dont forget thermals, it would need to operate at mobile-like frequencies to stay within any reasonable envelope.
90%, well, yes, on extremely parallel apps maybe. On average expect 40% or so, minus 30% for frequency-penalty. There will be no catapult for performance here, just linear continuation of the performance-curve.
As for calling it DualK8 or K9, i do not care about monikers.
K.
elmer
Well, looking at the earning surprise history over the last three quarters it is pretty clear that there is a widening gap between AMDs development and the perception of it.
http://finance.yahoo.com/q/ae?s=amd
Looking at the current quarter estimates, its going to widen again in January.
Obviously the learning curve of the Consensus seems to be a not really steep - for the careful version.
As for a plain-text version: These guys are falling back more and more behind the curve.
K.
yb
A second generation 90nm process is on AMDs roadmap for about six weeks already for H2/05 which somewhat excludes 65nm to be very close to that, right?
But well, it still could be a hybrid process. 90nm logic and 65nm SRAM. If so, its a question of semantics which moniker you use for it. Or a matter of taste, if you will.
Currently AMD needs to show a roadmap people can be confident that AMD can execute it. You know, to convince people that AMD can leverage 65nm and 300mm simultaneously in a brandnew fab is not really consistant with the perception of "AMD", which in this context stands for Always Major Disappointments.
Not that I would like that. What I like is AMDs Management faces the fact it has a credibility problem and takes action to change that. Which Bob R. explicitely stated recently.
K.
HailMary, Reseller Mike
If they don't distinguish this in some way, the clueless consumers will assume they are of comparable performance
Hmm. The clueless consumer will look at the same number and the same price. Nothing else. (Talking about Athlon numbers, they usually believe they are talking Megahertz, and barely ever mention the little plus behind the number.)
Now - assuming Mageek has it right and the (list)prices are indeed identical - (the other) Mike has a fundament to speculate on Performance-Rating-monikers. If the (numerous) indications pointing to higher thermals per clock for Prescotts will be validated as well, this fundament looks rocksolid to me.
If Prescott will indeed have rated numbers, I mean, why not? As long as its founded on relative performance, it is completely allright with me.
AMD did it for a while, and the market ate it.
I am personally tending to follow Reseller Mikes idea of putting things together here - unless somebody offers a more plausible approach to explain the pricing strategy as rumoured.
K.
Petz
Could be. Whatever the reason is or the reasons are, looking at it from a reasonable distance it does not matter much.
As I brought it up, I should probably adjust my perspective looking at it.
K.
Keith - WAD 2000
Not excited at all about it. Check w:o for first impressions.
K.
sgolds
IIrc you dont have it right in some details, but that is not important as you have the basic design of the program correct which enables you to see the background of my recent "Crow"-posting.
Sorry I dont intend to go into details of this program again, as i consider it as an unpleasant episode which I hope I can just sit out within some six weeks or so - I just knock on wood it is gonna luckily end without any hazzle. After that, I sure hope AMDs management is fully aware they owe their shareholders compensation for the pain they had with their investment for a while - which is pretty easy to deliver: Any price above 30 until the upcoming stockholders meeting would erase any bad memories.
K.
sgolds
Here we are:
http://amd.edgarpro.com/redirect_frames.asp?
filename=0001012870%2D03%2D001178%2Etxt&filepath=%5C2003%5C03%5C14%5C&cols=7%2C0%2C4&SortBy=receivedate&AD=D&startrec=76&res=25&pdf=0
By the way I tracked back to one of my postings about AMDs ESO-exchange-program
http://www.investorshub.com/boards/read_msg.asp?message_id=1152243
Interesting enough to read your reply on this one as well after couple of months
K.
sgolds
do you have a handy reference so I can read that ESO exchange?
Sure.
http://amd.edgarpro.com/redirect_frames.asp?filename=0001012870%2D03%2D001178%2Etxt&filepath=%5C....
Exchange ratios have been modified slightly before the program was set in force in summer, but the terms in general have not been changed.
And I have seen nothing similar before as well. If you track back to my thisrespective postings on this board you will find quite upset comments about it.
K.
sgolds
I hope the AMD program has a guarentee that doing the exchange will really result in a lowering of the option price
No.
K.
sgolds
and that date is early enough in Q1 so that there is insufficient information about Q1 (and FY2004) performance to impact individual employee decisions
May I suggest that you read the terms of the eso-exchange program currently under way?
There is no individual decision to make at all in January. These decisions (to participate or not) were made long ago.
K.
sgolds
Thanks lot for your valuable recommendations.
K.
Keith
Defining your security level is the important item that you need to think about, if you have questions I can help with that.
TIA.
Apparently, the notebook I ordered has WLAN capability already (although it is not mentioned in the technical specs), based on Intersil's Prism (dont know yet if thats integrated or by means of PCMCIA-card. If its b only I would hope it is card based so that I can easily upgrade it later, if it is b/g i dont really care what it is as long as it works.
I will let you know as soon I have my hands on it and will bother you with setup-support then.
K.
Keith, twimc
when you talk about "crow", you must have some ideas in mind?
Nothing particular at this point. But there is many many possible crows to be served in various cooking styles I could think of.
For the example you adressed: Fasl-Integration-Charges: No. Thats completed. But they could always close down JV1s 0,35µm production, take charges for it, tool JV3s extension-building as planned but not executed yet and issue another Convertible Senior Note for it. And yes, everything already in 10-Q. Possible discontinuation of manufacturing facilities and financing transactions covered in save haven statements, JV3 expansion explicitely noted. Just one of dozens of possibilities.
To prevent from any misunderstanding: I'm not saying whatever they could serve would not be a plausible or even necessary move. I'm only predicting how everything with any possibly negative impact they will bring on the table in January will be likely perceived by the street.
I can not comment on common practice. Manipulating stockprices is illegal at any time by whatever means in the U.S. (and elsewhere).
K.
Keith, and twimc: Patience
Yes. Could be it was the announced date for earnings reporting of Q4 results differing from usual pattern which caused recent negative sentiment of the street about AMD, as it probably nourishes fears of some crow to come on the table just before the strikes of exchanged ESOs will be fixed end of January.
Which I strongly hope wont come: If the board of directors and management indeed would elect to serve any crow in January I personally would consider this to be nothing short of just overtaxing shareholders. Somewhat understating communication while an exchange program is under way is - although stretching plagued "Chance of your lifetime"-Investors nerves - understandable to some extent. Any breaking negatives couple of days before fixing-date of strike-prices for new (exchanged) employee options to be granted would be definitely beyond any sane Investor-Relations-policy. It would conteract any management's efforts to transport the "New AMD" message to the street, as it would certainly be perceived as the continuation of the old cheeky "We're good at using other people's money" policy.
fwiw
K.
Keith
Well. Bought at least this one for a rock-bottom price. If I could say this for AMD-common, I would look a whole lot better though.
See my post on w:o for some more details on WAD 2000.
K.
I-banker Copied. Thanks lot & implications...
Checked google for a source and found: 32-bit card bus....
Which reminded me I gotta make sure there will be drivers available for Windows-64 before I buy.
Oh boy. Many many years ago I promised myself to stop pioneering anything in Soft- or Hardware. Which is what I (mostly) did consequently for the last ten, fifteen years. Now I strongly tend to fall back into the old habits.
Naah. I'm gonna use Windows2k on it for quite a while and let the others do the pioneering!
You dont believe ?
Sigh. Neither do I .... ;-(
K.
I_banker
Yes, I would have preferred that. And Thin and light as well.
And 5 hours battery runtime. Below 1500 USD.
Just could not find that.
Plus the book I found was an occasion which I just could not resist.
K.
sgolds
Sure its gonna be a multiband card, which is capable of g (b) and a. My question was, which manufacturers solution should I choose?
K.
Petz
Medion WAD 2000. 7lbs, 3000+, 15", 512KB, 60GB, no WLAN included.
K.
p.s: Recommendations anyone which WLan-card (a/b/g) i should buy for it?
deleted duplicate
For the record: Just ordered
naah. Not more shares, as I am skewed towards AMD to some unsane level already.
But an Athlon-64 notebook.
K.
yb
I think we are close to have Athlon shortages in early Q1.
Can I interprete that as ... Athlons under 50 Dollar streetprice... ?
that is my hope as well.
I can repeat that I think AMD will lose some marketshare in Q1.
Can I interprete that as ...in terms of Volume-MSS? I would not argue at all - and I would not bother at all.
However, in terms of Dollar-Marketshare, I hope AMD can gain a lil bit in Q1.
The sweet spot will be 2600, as I don't think sub-2600 models will reappear anymore.
Well, from a production viewpoint, yes. From a shipment viewpoint, that seems pretty optimistic to me. Dont forget, after selling out Thoroughbreds there will be a pile of Thortons - no idea of how big this pile is, but should be sufficient for couple of quarters to satisfy the 50-65 Dollar street-price range. (XP 2500+ will be around 75 Dollar then and will continue to pull value-buyers-demand on it).
Plus, more water into your wine: There is a still a huge pile of Appalbreds as well - and AVNET still has (plenty?) Fab25-Durons in stock.
Anyway, I agree: ASP up next quarter. And consecutive quarters as well. Which is the obvious strategy as long as AMD is capacity-constrained, isn't it?
K.