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.
Re: AMD mobile went from 12.2% to 15.1%. The KoolAid drinkers may be under the impression that AMD share is zero and dropping...
What these guys keep missing is that Northwood on 130nm was a 35w chip at desktop performance levels.
It's pretty easy to make a low power, 32bit X86 chip. It's harder to make a low power 64bit X86 chip. Look what happened when Intel went from Northwood to Prescott (and AMD had trouble going from Athlon XP to Athlon 64, too, until they added SOI).
AMD, thanks to its more advanced SOI FAB technology was able to produce 64bit mobile parts. Maybe Intel's 65nm SiGe tech will let them do it too (with Merom) towards the end of the year.
Maybe not.
Meanwhile, AMD's 64bit parts are gaining market share and Intel is looking somewhat silly with its "new" 32-bit mobile parts.
Intel extends the P4 line with new parts!
Intel ships low-spec P4s for low-cost PCs
By Tony Smith
24th January 2006 11:10 GMT
Intel has begun shipping a pair of low-cost Pentium 4 5xx class CPUs that have yet to be added to its public price list. The chip giant is also preparing to offer a low-price dual-core Pentium D 8xx processor in two months' time.
http://www.reghardware.co.uk/2006/01/24/intel_p4_517_524/
Well, when your binsplits are no good, and demand for your parts is dropping, you do what you gotta do.
Re: The move to 65nm is not a simple shrink
Yep, it may be the opposite. Rather than a shrink it's more of a relabeling of the 90nm process while increasing strain:
Maybe Intel should call it their new 80nm process....
Strain not scale marks Intel process, says consultant
Peter Clarke
EE Times
(01/24/2006 8:24 AM EST)
LONDON — Straining rather scaling transistors is the key to the latest 65-nanometer logic process from Intel Corp., as exemplified in the Pentium D 920 dual-core processor, according to Canadian technology and patent analysis consultancy Semiconductor Insights Inc.
In a statement issued Tuesday (Jan. 24) Semiconductor Insights (Ottawa, Ontario) did not discuss what transistor dimensions it found Intel using in its 65-nanometer logic process, but indicated that Intel has concentrated on straining its 65-nm transistors rather than scaling them.
<snip>
Semiconductor Insights prompted discussion in the industry when it claimed in January 2004 that the Emotion Engine + Graphics Synthesizer otherwise known as the EE+GS@90nm used in a Sony games console was made using a 130-nanometer manufacturing process technology, despite plans to implement the combination chip in a 90-nm manufacturing process and a naming convention that implies the use of a 90-nm process.
Sony denied the charge at the time, arguing that the definition of a manufacturing process was complex and that its own CMOS4 process was in-line with other 90-nm process technologies.
http://www.eetimes.com/news/semi/showArticle.jhtml?articleID=177103163
They aren't.
You'd better review your notes.
Re: You know, droidland... lah lah... such sad and creepy people
You forgot to add wealthy - especially compared to Intel holders who've seen the value of their investments collapse.
Maybe you should be doing a little less posting and a little more reading.
AMD has just opened up a major new initiative in the server market, to, at minimum, double their share.
And what happened to Dothan "shutting down" Turion? AMD has gone from nearly invisible in the mobile market, pre-Dothan to having a solid substantial share of that market.
I know, now you're all going to say wait for Yonah, Yonah will shut down Turion.
Sure it will...
Re: WHERE IS AMD WITH 65nm
Nominal feature width reductions has been delivering less and less of a performance improvement (speed or power use) starting at the 250nm node.
From 130nm, moving to the next node has introduced as many performance gating factors as it has improvement.
What has been improving performance?
AMD added copper at 180nm
Intel added copper at 130nm
AMD added SOI at 130nm
Intel added SiGe (point compressive) strain at 90nm
AMD added multi-plane strain at 90nm
Intel appears to be doing a simple shrink at 65nm
AMD is adding SiGe (point compressive) strain at 65nm
Of the process changes that have let designers improve performance since 1999, AMD presently has SOI (that Intel does not) and Intel has SiGe compressive strain (that AMD does not). AMD is adding SiGe compressive strain but Intel is not adding SOI.
Intel looks like it's headed for a dead end, since AMD will not only be going from 1 to 3 CPU FABS next year, but one or two of those FABs will be able to use more advanced fabrication technology than any of Intel's FABs.
What's the matter bo bo,
followed your own advice and now your net worth's half of what it was?
Re: They have themselves so convinced that Intel
This from the guy with a screaming sell on AMD in the low 20's and a screming buy on Intel when it was in the 30's.
Talk about someone who's "convinced himself."
Re: Go back and check the figures, numbnuts. AMD had an ASP for K7 in excess of $100 in 2001.
Yes, and Intel's ASP was about twice that for the Intel Architecture group.
What's you point?
Re: They never had 20% revenue share. They claimed 22% unit share in 2001, if that's what you're thinking of.
Exactly. And since AMD's ASPs were about half of what Intel's were back then, their revenue market share was about 10%.
AMD is now well into uncharted territory when it comes to market share from a revenue standpoint. And their market share will almost certainly continue to climb for a number of quarters (maybe a number of years).
Intel is into uncharted territory in terms of how small their revenue market share is - and it's dropping.
Re: while performing 2-3x slower using Rosetta. I can see the Rosetta part being a show-stopper for some end-users
So the Power PC based Macs are 100% to 200% faster than the Intel based Macs, unless you replace all your software, in which case the Intel based Mac is ~20% faster.
Well, if you're buying an additional system, and need to buy additional licenses anyway, then that won't be an issue (as long as an Intel version of every one of your Mac software programs is available).
OTOH, replacing an existing PowerPc Mac with an Intel based one won't make sense for a year or two (and I wonder if Apple will still be in business by then).
Re: My question still stands, however. Name one incompatibility
For us the issue is Microsoft Access (but that's neither better now worse in the "new" Macs.
Re: since they've released (apparently) Conroe 3.3GHz parts to various places they'd better get some real benchmarks published to show the lead is going to change.
That strategy certainly did the trick for Adam Osborne.
As far as publishing extensive details on how lame their current stuff is compared to what will be available at the same price in a couple of months, AMD seems to prefer waiting until the new stuff is close to being available for sale, but what do they know...
Re: Intel Chief Financial Officer Andy Bryant also noted that Intel's yearlong supply problems with desktop chipsets extended into the fourth quarter
Aw 'cmon.
Do you really expect us to believe Andy Bryant over WBMW?
Get real!
Re: Either way you cut it, (88.23-84.67 = 3.36) or (15.13-11.76 = 3.36) AMD only gained 3.36% revenue market share
You're right!
AMD is only gaining revenue market share at a rate of 28% per quarter.
As you've pointed out, at that rate AMD's revenue market share will be the following:
Q3 2005 11.76% Actual
Q4 2005 15.13% Actual
Q1 2006 19.36% SnowRider2 Projetion rate
Q2 2006 24.78% SnowRider2 Projetion rate
Q3 2006 31.72% SnowRider2 Projetion rate
Q4 2006 40.60% SnowRider2 Projetion rate
So I guess Intel has nothing to worry about.
Re: Riiiight, because for the 3rd or 4th consecutive quarter, Intel is having chipset shortages...
Then I'd expect that Intel is in frantic talks with TSMC and UMC to buy licenses to use TSMC/UMC FAB technology and then have TSMC/UMC engineers come out to teach Intel how to FAB a chipset.
Unless something other than chipsets is the real issue?
Re: I figured the patronizing tone would start settling in when the AMD crowd ran out of good reasons to cheer for the company
Yeah, it looks like AMD's biggest risk right now is that Intel might license technology from UMC or TSMC and learn how to FAB a chipset.
Re; I just picked up a bunch of AMD Feb. 30 Puts.
If you were a real believer, you'd have sold naked calls...
Re: And I wonder who's going to be the moron with the "sell" on Intel and "buy" on AMD for all of 2006? Might that be you? <VBG>
LOL !!!
Open mouth, insert....
Re: Intel claims Itanium eating RISC competitors
LOL!!
In other news, donkey wagons continue to take share from ox carts.
One dead platform eating the decaying flesh of another dead platform.
Single core G4 runs applications 30% to 60% faster than Core Duo dual core - OUCH!
http://www.aceshardware.com/forums/read_post.jsp?id=115151807&forumid=1
The biggest problem that may face both Intel and AMD is this:
This is a real shocker (to me, anyway).
Keep in mind that between realtor fees, legal and tax costs, it's hard to get more than about 94% of the selling price of a home sale.
Surge in no-money down loans could bite
Report: 43% of new home buyers chose no down payment last year, which could hurt as market cools.
January 18, 2006: 6:57 AM EST
NEW YORK (CNNMoney.com) - More than four out of ten first-time home buyers opted for no-money-down loans last year, a move that could prove disastrous for those buyers if the housing market cools, according to a report published Wednesday.
Citing a report by the National Association of Realtors, USA Today said that 43 percent of first-time buyers put no money down last year.
The median new buyer only put 2 percent down on a $150,000 home in 2005, the report said. Half of all new buyers put down more than the median and half less.
Lots more at: http://money.cnn.com/2006/01/18/real_estate/downpayments/index.htm
Re: If you look at the results without the estimates and expectations, Intel is still making a ton of money.
And while he's at it, he should note that (absent forward expectations) it's a good time to add to Ford and GM holdings, too.
Re: nothing is easier than arguing his lame points.
Of course it is. As long as you're willing to make stuff up.
Why don't you tell us again that Intel makes money on flash.
Magellan Fund dumping Intel
Worried about a price war and their compeitor's rapidly increasing capacity, no doubt.
http://www.marketwatch.com/news/story.asp?guid=%7B38049962%2D1083%2D45EA%2D88A5%2D9FBCDFB399E3%7D&am...
Re: Cayenne
One of the car magazines (I forget which one - maybe Automobile or Road and Track) tested the new Jeep Grand Cherokee SRT8 and the Jeep was faster through the slalom, faster on the skid pad, and faster in acceleration.
Curiously, the Jeep's (by the factory, for the SRT8 model) lowered and modified suspension, made the Jeep less desirable than the Cayenne for serious off road work.
Just an FYI, but the Jeep appears to run rings around the Cayenne.
Regards,
Dan
Re: what is the latest about the Yonah release? When will it happen?
They're out there, but the early shipping parts don't seen to spec out too well.
What will be interesting to see is what speeds they ship at, once (if) Intel gets their 65nm process sorted out.
Intel had trouble yielding Dothan at more than 2ghz, and Yonah has more heat to dissipate, so it may be tough to yield parts that do well on single threaded code.
There is a Yonah based notebook available at CompUSA.
It's a 15.4" widescreen, weighs under 7lbs and has a 3 hour battery life. Runs at 1.66ghz but that battery life would have been one core running at a lower clock speed (at least, that's what their power saving design strategy calls for) Sells for $1,199. Yonah might not be the savior Intel was hoping for.
http://www.compusa.com/adproducts/product_info.asp?product_code=336977&pfp=ADPRODUCTS
Remember when Intc was considered high end and AMD was sold as low end?
Here's a listing of the systems that are being advertised by Best Buy and Circuit City this week. Seems like Intel is losing the "hearts and minds" battle. Note the progress AMD is making in the mobile space.
Best Buy Notebooks:
AMD $1,500
INTC $1,350
AMD $1.200
INTC $1,200
AMD $850
AMD $750
INTC $550 Featured
Best Buy Desktops:
AMD $1,060 Featured
AMD $930
INTC $850
AMD $520
INTC $460
AMD $320
Circuit City Notebooks
AMD $1,550
INTC $1,100
INTC $900
AMD $850 Featured
AMD $750
INTC $650 Featured
Circuit City Desktops
AMD $940
INTC $890
AMD $580 Featured
AMD $430
Whats next for COVAD
Tomorrow.
Re: It's not fraudulent, and in fact it's how the finances are computed throughout the entire industry.
Sure it is. Most companies have divisions that lose $400+ million on $47 million in sales.
Can you link to a couple more examples?
Just think of me as being from the "show me" state.
Should be easy since you claim it's "how finances are computed throughout the industry"...
Re: IMO, the deal most likely enables AMD to proceed with industry memory technologies in their memory controllers without fear of litigation. If they hadn't signed, it would have exposed them to more of Rambus' "crookedness".
I can't argue with you there (try not to fall over) but I still don't like this.
When Intel tried to work closely with Rambus, the end result was pretty bad for Intel.
Now AMD may be working more closely with Rambus and it worries me. Maybe they're just not taking chances, but even so it's a bad idea.
Paying Rambus anything is making your first payment to a blackmailer - it usually gets worse, they don't just go away and AMD only signed a 5 year agreement. Low priced heroin for 5 years can get you into trouble down the road - if not right away...
Re: it's a rather large bucket, which includes things from special employee compensation (bonuses, for example, which can't be placed under R&D), to income
Sounds fraudulent, shouldn't compensation for employees in the CPU segment show up as costs for CPUs? Did they pay their landscaping staff $300 million in bonuses?
Whatever.
The original issue was whether or not Intel has managed to make a significant share of its profits at anything other than CPUs and tie in sales. They've managed to lose money in many creative ways (as you've pointed out) and have certainly developed a large, money losing flash business, but outside of the one prouduct, no luck at making money.
Re: 90->65nm crossover ~ Q3 2006 as you well know.
No, I clearly didn't - got a link?
Will you be convinced that Intel is doomed if the majority of their shipped product in Q3 isn't from 65nm lines?
From the time of the .18 to .13 shrink, there have been few significant advantages from shrinks other than saving a few dollars on wafer costs. Improvements have come from things like Al to Cu, adding strain, adding SOI, adding SiGe, etc. After spending $2 Billion or so to upgrade a FAB to the next shrink, you save about $5 per CPU on silicon fabrication costs. Unless you need some aspect of the new process to stay performance competitive (and keep ASPs up), you lose money on the first 500 million parts produced at that fab.... There are a lot of cases where you're better off just running more wafers using existing equipment (if the performance of those parts lets you charge competitive prices).
Note that Intel's flagship product, Itanium, tends to be late to the shrink party.
Re: intel never convert's old fab's to a new process for CPU's, they just build new fabs for the new process from the get go
Ummm. Nope.
For example: http://www.intel.com/pressroom/archive/releases/20030218corp_a.htm
Re: You're just plain wrong. Wishful but wrong.
So tell us, excluding the generally profitless flash segment, what sales revenue do you see Intel receiving that isn't dependent on CPU sales?
I'll help you get started. Intel sells a fair number of embedded parts for things like blackberrys and disk controllers. They also sell quite a few network controllers (the network chips on a lot of motherboards, including AMD motherboards. For example, I've got Opteron servers that have Intel network chips and Intel based RAID controllers.). Now estimate the numbers and revenue per unit and see how much of Intel's profits come from there.
Re: The development of Core chips is the first in a series of bet-the-company moves that Mr Otellini is making
And if even one "bet the company" move doesn't work out?
Re: we won't have the majority of our production on 65nm until mid-2007
When will Intel have the majority of its production on 65nm?
Another thing to keep in mind is that AMD will be keeping its current FAB at 90nm through 2007, since AMD's 90nm SOI process continues to be competitive.
Intel's 90nm process is bulk so they may feel the need to convert all existing lines to 65nm to keep up with AMD's more advanced technically but larger feature size 90nm process.
Re: The statement was made that Intel derives x revenue from non CPU items
Duke, if it makes you happy to think that their tie in sales of chipsets and motherboards means that Intel isn't virtually 100% reliant on CPU sales, that's great.
But unless and until Intel starts supplying motherboards and chipsets for VIA and AMD CPUS, then a 25% rise or drop in Intel CPU sales will continue to mean a 24+% drop in gross profits at Intel.
Which was my point, and which was not some sort of "for argument's sake only" point.
Re: I guess this means that Rambus is a good and virtuous company
Rambus is a bunch of crooks. AMD shouldn't have signed.
Re: what the "other" group includes
I'm not the one who brought up other than PC CPUs and their support chips.