I just bought QQQ at $20.25
Wish me luck.
wbmw



You're a braver man then I am Gunga Din.......

Semi

This whole discussion has been a whole lot of fun (sarcasm)...

Actually I was going to say...... boring (no sarcasm). JMO

And by the way, this is just my opinion on how we might get back to discussing Intel and the things that may make the stock go up or down or just stay where it is for a while.
Nitt

Thank you, that's the first smart thing I've heard somebody say in the last page of posts or so.

I didn't come over here to bring the B.S. Waaaaaaaaaaa I don't like your moderation, crybaby discussion, that I left behind on SI. If anybody doesn't like the way this thread is run, then fine, get out. Just do me a favor and shut the hell up about it already and do it.

Somebody once said that people take these boards too seriously, maybe some people should revisit that advice.

I'm outta here until people grow up.

Semi

If "constrained" is an arcane euphemism for "problem", then so be it. I am not an insider in the industry and as such I am not privy to its jargon. I would tend to consider, though, that the webcast was intended for public consumption and as such, it would contain language that follows the more publicly accepted connotations.

--- That's one point of view. Of course, that point of view would be based on the assumption that AMD was trying to be honest, and informative on issues relating to thier 0.13u production, and not trying to sidestep or mislead the public. It appears that some do believe the honesty theory. I don't.

The evidence of lack of AMD parts >2200, 2 months after launch, and the unavailability of their announced top of the line parts until November, plus the general opinion on most web hardware sites, seems to suggest that most believe the misleading theory instead. It's a free country though. Everyone can believe what they want.

Semi

"You can observe a lot just by watchin'."
Yogi Berra

For the record, the AMD spokesman used the words "constrained by normal yields". Never was the word "problems" used. Your misattribution, if intentional, was pure FUD.

In Semiconductor Manufacturing talk, a Constraint, is a term generally used to describe a problem, and Being "Constrained" does mean that whatever you are trying to do is being limited by problems. Using the words "Constrained" and "Normal" when you're talking about Yields, should be an oxymoron, unless..... "Normal" is very very bad. JMO.

Semi

NEW AMD Logo For XP2700+

Paper Me......

http://www.hardcoreware.net/reviews/processors/axp_2700/1.php
It seems like just yesterday when AMD released their latest CPUs, the 2400+ and 2600+. Actually that statement is only half true right now - The 2400+'s are just starting to show up, and 2600+'s have yet to be seen in stores at all.
So it is interesting to me that AMD is launching another CPU before their last launch is even complete. I guess they're feeling a little extra from Intel lately; it seems as though every time AMD launches their fastest CPU, Intel responds just days later with a faster one.
And to make matters worse, Intel is actually able to ship their CPU's at launch date! In fact, before the 2.80 GHz was even officially announced, vendors were listing them on Pricewatch.
How these paper launches affect AMD in the long run should be interesting to find out. I tend to think it will hurt overall, and the fact that the paper launches are getting worse and worse every time concerns me.

---Oh man, Web sites making fun of AMD's marketing 2 days in a row. Those guys sure know how to make up funny stuff, eh??? Reminds me of an old joke.....

Fish Buyer: Hey, how come your prices are so high, the fish seller across the street sells fish that he says are better than yours, and he'll sell me 2 for the same price as every one of yours!

Fish Seller: Then why don't you go buy your fish from him then??

Fish Buyer: Well, he's out of stock on fish right now.

Fish Seller: Well, when I'm out of stock, my "Better fish" are 2 for one also!

Semi

The sad thing is I predict that many of those who decried Intel for the paper launches will offer up a multitude of rationalizations as to why its "okay" for AMD to do the same... to an even greater degree...

---You already gave an example of this happening. THG ripped intel apart during the Cu Mine Fiasco, but now.... all he has to say is........

Hopefully, this won't lead ambitious AMD fans to despair, because the Athlon XP 2200+ with the Thoroughbred A core has already been available for three months now. And, we've checked out the availability at various computer specialty stores: they're currently taking waiting lists for the Athlon XP 2400+ and 2600+."

---Paper launches and waiting lists, but a free "hopefully" pass for AMD from Tom.

Semi

Intel X-Scale Processor PDA Win

http://biz.yahoo.com/rf/020930/tech_microsoft_pocketpc_1.html

ViewSonic, a closely held maker of computer screens and flat-panel displays, said the new PDA would sport Intel Corp.'s (NasdaqNM:INTC - News) 300MHz X-Scale processor, have 32 megabytes of flash memory, a 64,000-color display and a Secure Digital (SD) slot.

Semi

Bwaaaaa Haaaa Haaaaa Haaaaa Haaaaa

http://www.hardocp.com//image.html?image=MTAzMzM0MTY2MzY4aVM2UVdjQ2pfMV8xX2wuanBn

Hey, don't blame me, I didn't put it up on the Web. Blame Kyle at HardOCP. OK OK, so I'm guilty of not being able to resist posting the link here. So sue me.

Semi

I think it's a good read...
http://www.nytimes.com/2002/09/29/technology/circuits/29CHIP.html?ex=1033963200&en=3b60ec62ca6d0....
signed,
Bernard

I agree, it's a good read. It condenses everything that has been said for awhile now. So if someone hasn't been following Itanium/Opteron, it's a good compilation of the information already available.

Upon closer examination though, it really doesn't tell any of us who have been following the situation, anything that most of us didn't already know. It seems that the writers are saying that Itanium2 might be a success, or it might not. Opteron might be a success, or it might not. HP might have success with Itanium, or it might not. Intel might have Yamhill, or it might not. See what I mean?

Good article as far as gathering of information though..... Thanks.

Semi

Pressure is growing on AMD
SLUMPING EARNINGS, INTEL, DELAYED CHIP ALL HURT
By Therese Poletti
Mercury News
Advanced Micro Devices is under the gun.
The Sunnyvale computer chip maker has posted four consecutive quarters of losses. It's losing market share to archrival Intel.
And two weeks ago, AMD was forced to push back the launch of its newest chip for desktop personal computers, code-named Clawhammer, raising the specter of past manufacturing woes that once troubled the 33-year-old company. Worried investors have sent the company's stock down 30 percent since then.
``There have been a bunch of people who have been skeptical that there is a new AMD,'' said Nathan Brookwood, an analyst at Insight 64 in Saratoga. ``It's the first time in a long time that AMD has had a visible schedule slip. People who remember are wondering if this is a start of a new trend.''
http://www.bayarea.com/mld/mercurynews/business/4169842.htm

Semi

He may be very smart, but the lawsuit that would follow him to intel would not be worth it.

I'm assuming you're talking about some sort of "Intel stole our people" type lawsuit. I don't think that would happen. Historically, in order to prove something like that, a company would need multiple defections, and most of those people would need to then be hired by intel. Proving some sort of collusion based on one person, no matter how important they may have been to the losing company, would be extreamly difficult, and would drain a companies finances further.

Time, and Money. 2 things in short supply right now over at AMD. I'm not saying intel should recruit him, but if HE happened to make first contact, I would interview him. JMO.

Semi

Not Good IMO..... For AMD:

http://www.infoworld.com/articles/hn/xml/02/09/27/020927hnmeretsky.xml?s=IDGNS

AMD's Meretsky leaves

By Tom Krazit
September 27, 2002 1:06 pm PT
ADVANCED MICRO DEVICES (AMD) Fellow Wayne Meretsky has left the company, an AMD spokesman confirmed Friday. Meretsky was a driving force behind AMD's efforts to promote its forthcoming Opteron server chip.
Friday is Meretsky's "effective" last day with the company, said Morris Denton, an AMD spokesman.
"He announced a couple of weeks ago that he was leaving on his own accord, for personal reasons," he said. Meretsky may provide consulting services to AMD in the future, he said.

--- Yeah right, Consult In The Future. The same thing AMD said when Atiq Raza left, and it didn't happen. "Leaving on his own accord, for personal reasons". Translation, The ship is sinking, so I better get the heck out of here before it does. "A couple of weeks ago", shows no transition planning, and most likely a surprise to AMD. It doesn't seem to indicate he's leaving because of a better offer, more like "I'm fed up with this place, I Quit. Here's my 2 week notice".

Smart guy, maybe INTC can use him????

Semi

Interview: Sun's Scott McNealy on Intel, Dell and more

"Ten years from now, you'll know whether I was on drugs or not."

http://www.silicon.com/bin/bladerunner?30REQEVENT=&REQAUTH=21046&14001REQSUB=REQINT1=55708


From the direction SUN has been going lately, my guess would be Yes..... He's on drugs.....



Semi

I been pretty busy lately ... But, I have had a chance to go through both of these presentations. One disclaimer: "I do NOT have any additional knowledge about this process that what is presented in the two papers". Having said this, here are my comments:

Shoooooot, that was way beyond my knowledge.... Sounds like you've had experience as a Process Development Engineer, or possibly Device Engineer..... Close?

Semi

Slightly OT: Still, though. Those people are more than just a statistic. My heart goes out to each and every one of them. Probably the scariest thing in anyone's life is to be without the security that they've always taken for granted.
wbmw

It's rough, I've been laid off in my career, twice. Both times, just before the holidays. Last time, it was the Medical Instrument Company in Palo Alto, Ca. They said that they needed to balance the books, so after working for 2 years to develop the new product, and 4 years improving it and trying to make it profitable, just when it started to make money, the company transferred the product to the main Los Angeles Facility, and laid off all 14 of us in the department.... Not Real Fun....

Semi

Without any clarification, this statement is kind of meaningless. Could you give some examples of problems caused by running bulk Si and SOI on the same production line?
There probably aren't any examples because nobody's pulled it off yet.
BTW, nice to see you've changes your alias to something more reasonable.
EP


Elmer,

It threw me for a second why I couldn't see the post you were responding to, or who the heck you were talking to. I was going to ask you what the heck was going on, Then I remembered... Right click, properties.... Ahhhh now I see.... I Guess somebody must have forgotten what I said....

http://www.investorshub.com/boards/read_msg.asp?message_id=508632

I wonder if you can set filters both ways (posts from and too) like on SI? Would be less confusing in the future....

Semi

Semi,
The first was credible, just lacking in information.
Intel working on a new DRAM standard with selected vendors? Is that industry politically correct, especially after Rambus?
John

Could be.... As was nicely (Thank You) pointed out to me in PM, I'm not always aware of everything that's going on......



Semi
"Occasional Mushroom"

Now This Sounds More Credible.....
http://www.theinquirer.net/?article=5585

The Economic Daily reckons Intel is seeking to increase its interest in Powerchip and collaborate with Elpida on memory production. Intel would be expected to define DRAM standards for the alliance, the paper says. Elipda and Powerchip would produce the goods.

Intel may well be involved in these apparent four or five-way discussions as it sees worldwide DRAM production concentrating into the hands of fewer producers. A multinational agreement involving Taiwanese and Japanese production facilities might allow the American chipmaker dabble in DRAM as a sort of glorified message boy.

intel trying to TAKE business away from the DRAM (Hey 8-/ did I get it right that time? well, as long as you're pleased eh?) makers. After all, intel seems to have the capacity and the muscle to do it.

Semi

08:05 ET Intel talking to Japanese chipmakers about DRAM production: Bloomberg (INTC) 15.20: Bloomberg reporting that the Economic Daily News out of Taiwan says today Intel has been in talks with Japanese chipmakers Mitsubishi and Elpida about making memory chips - the story doesn't cite any source for this speculation and an Intel spokesman would not comment. Separately, Korean chipmaker Hynix sold its flat panel division for $380 mln, which could provide liquidity to stay in the memory business longer. Both news items are being seen as negative for Micron (MU).

I find that hard to believe, based on the amount of 0.18u and 0.13u capacity intel has on-line. Why would they be talking to a competitor about making chips for them? That would mean that intel decided to idle the Ireland Fabs over Christmas, due to poor market conditions, and at the same time contract out manufacturing. It doesn't make any sense.

Semi

Yousef: You clearly know a lot more about the process than I do. I would only add that it is likely further complicated by the fact that the same Fab is being used for two other product lines.

As long as the process is the same for the different products, it really shouldn't have that big an impact. For example, intel runs P3, P4 and Itanium on the same production lines. It's simply a matter of swapping the reticles in the Litho Steppers to lay down the different product patterns. IF on the other hand, you're trying to run Bulk SI and SOI (2 different processes), on the same production lines..... that could be a problem.

Semi

The best "knob" to turn is poly linewidth ... Thus it is both a litho and etch issue. Both of these need to be closely monitored. I have seen the poly etch bias be as hard to control as the litho bias. I have also seen etch bias from isolated to dense features and PFET versus NFET biases at etch.
At least in Litho, some of these can be corrected on the reticle. Pushing the process just reduces the margin for devices to be out of spec (Ioff, BVdss, Vt, ...).
Make It So,
Yousef

Yeah, what he said, but it's also critical for other layers. For example, if your Contact Layer pads are too small, Post Etch, the VIA interconect might miss the Pad alltogether, resulting in only a partial connection between layers, and resulting in dead die. But Yousef is right about speed. Poly is frequently referred to by P.E.'s as "The Speed Layer", for reasons Yousef described. Litho creates the patterns, but depending on the Etch Bias, the Post Etch Critical Dimensions are the "Data Feedback", to the Litho folks to tell them if their Focus/Exposures are set correctly.

Thanks Yousef!!!

Semi

So AMD may be able to babysit their fab lines and make a couple of highspeed samples but they clearly can't manufacture them.

Agreed, and guess who's doing the baby sitting. Like I said, there's probably some pretty miserable P.E.'s over there.... Been There, Done That, Hated It. Thankfully I don't need to do that anymore......



Semi

But seriously folks....I have always believed that the SI Intel thread before TC, PE and El and others started being picked on, was much better than any other sources, combined.

Sure, hindsight is always 20-20, but it's pretty irrelevant at this point, isn't it. I received a PM that the AMDroids on SI are happy that I'm gone, and I'm happy here..... so it's time to move on IMO.

Semi

AMD doesn't seem to be able to keep up. Why don't they just cherry pick a few chips and introduce 3000+? Paper yes, but that's about all they ship at the high end these days.
WHy dilly dally around?
Well, no one ever accused AMD of marketing.

That assumes you're getting any 3000+ off the line. Process Parameters are a balance between Speed and Yield. If you are optimizing your line for speed, instead of hitting within the process window, then there's a certain point that the Yields fall off drastically. If you are already pusing your Process way past it's designed Process Parameters, as I suspect AMD has been doing for awhile, then you will eventually end up with no yield at all. I bet there are some pretty miserable P.E.'s at AMD.

Semi

I´m pretty dissapointed with the way this thread has evolved, looks like I have to find a new home. At least I tried, and got good anwers from wbmw, semi and yousef (thx guys) before the bashment started again. I´m not interested in reading that rubbish (or even posting), as real contributions would simply get lost in the noise. It´s a shame.
KD2000

Keith,

Are you sure that's what you want? I have a feeling that things will die down after awhile. The reason I say that, is because my feeling is that most of the bashing is coming from people who are currently posting on Silicon Investors AMD Moderated Thread. I haven't been there for awhile, but what I'm hearing is that many of them are P.O.'d that the people they enjoyed bashing... left. I hear that iHUB is a hot topic of conversation over there.

You'd think they'de be happy to have things to themselves, but it doesn't seem that way, since apparently several of them have decided to acquire accounts here (under different names of course), so that they could continue their intel bashing, while claiming innocence. Apparently bashing isn't any fun over there for them anymore, with all the other posters agreeing with them.

The same thing happened several years ago, when intel.com had a message board. It was great fun to bash, when they were bashing directly at intel, but when the board was closed down, several of the AMD fans stated they were going to another particular place. Almost all of them stopped posting at that new place after a short time. I think it's the same Psych that makes them insist the intel fans who left were all paid intel people. No fun for them otherwise....

Bashing is no fun, unless you have someone to bash. Give it time, they'll get tired of it....... eventually. JMO

Semi

Michael Dell Story

http://www.fortune.com/sitelets/innovators/dell.html

Semi

BMW,
Re: "Intel plans two-week shutdown of Ireland fab (Not a good sign, IMO.)"
Agreed ... Is this happening at any other INTC Fabs ??
Make It So,
Yousef

Yousef,

I'm not aware of intel announcing any other Fab warm downs...

Semi

Intel Brings Low-Voltage Processor to High-Performance Embedded Boards
Monday September 23, 11:03 am ET

SANTA CLARA, Calif.--(BUSINESS WIRE)--Sept. 23, 2002--Intel Corporation today unveiled the latest addition to its embedded processor roadmap -- the Low Voltage Intel® Pentium® III processor at 933 MHz -- for use by communications equipment manufacturers in high-density blade systems.

http://biz.yahoo.com/bw/020923/230689_1.html

Semi

Intel Is Kicking Silicon at AMD
The upstart's soon-to-be launched Clawhammer chip, while a step forward, may not be enough to keep pace with the giant

---Did you see that new intel inside commercial during Monday Night Football? The commercial zero'd into a small circle in the middle of the screen, then overlayed a statement like "intel inside... my homework, then the circle spreads out, and several scenes are played showing young people using intel technology in innovative ways for completing homework tasks. I only saw part of it, but in addition to Homework, I saw segments on Digital Pictures, and Adults at work. I like the way intel showed examples of people using the technology, instead of just standing in a crowded corner with "Deer In Headlight" looks, saying "Do Me".



Semi

You seem upset to think someone may think you're associated with Intel. I'd be proud to be associated with such an upright forthright organization.

--Not upset at all. Perhaps you just think I should be upset, because you seem to be. So sorry to disappoint you....

If Andy'd referred to 'our boys in blue' would you have taken it to refer to Intel? I'd have taken it to refer to the US Navy(or maybe IBM, but certainly not the blue boys of the Intel commercials).

---IF IF IF.... The point is, he didn't. I think I've shown that the so called "Andy Grove" was pretending to be the Chairman of the Board Of intel. And I'm pretty sure everyone can see through your B.S. too.

Goes to show how easily one can mis-interpret meanings.

---I think everyone can pretty much see who's trying to mis-interpret things here, and who caught them doing it. I'll leave it to the individuals reading the posts. I have a feeling they're smarter than you obviously give them credit for.

P.S. - Do you really miss me that much on SI that you had to come on over here "Win"? I would guess it's pretty boring over there with only AMD Cheerleaders left, eh?

Semi

Phudd You are claiming to be the Chairman of the Board of a Company that is relevant to this industry. You are acting as an imposter and that is no doubt a violation of the rules.

**************************
Are you sure he's not one of these?
1 ANDY GROVE CATHEDRAL CITY CA
2 ANDY GROVE PANORAMA CITY CA
3 ANDY GROVE LA QUINTACA
4 ANDY M GROVE NORTH TONAWANDA NY
Which Andy Grove do you think he's impersonating?
Just curious.
**************************

---Since you're "Just Curious", I'm happy to help you out. It's pretty clear that he was suggesting that I'm an intel Process Engineer, when he referred to me as "Our Process Engineers".


So, perhaps you could prove me incorrect by posting on SI. I'd appreciate it. Also, that usuck fellow knows enough to at least call it SRAM rather than DRAM. Surprisingly, Semi doesn't know the difference. Gee... I thought our process engineers weren't so parochial.
Andy Grove


---How many of those other people are associated with Process Engineers.....? Just Curious.

Semi

So, perhaps you could prove me incorrect by posting on SI. I'd appreciate it. Also, that usuck fellow knows enough to at least call it SRAM rather than DRAM. Surprisingly, Semi doesn't know the difference. Gee... I thought our process engineers weren't so parochial.
Andy Grove

SIGH....., SRAM... DRAM....., RAM is RAM. From a Process standpoint, It's all the same to me. Maybe you have me confused with a Device Engineer.

Last Post on that Topic.... Filter Set.

http://www.investorshub.com/boards/read_msg.asp?message_id=504537

Semi

Hello everyone; does anyone know what the die size of the
new 2.0 Ghz Celeron will be, will Intel actually go to the
trouble of making them smaller than the Northwood core, in the
past only part of the L2 cache was active on the Celerons. I
guess there are pro's and con's to both these approaches. I see
some are already on sale at a local screwdriver shop up here
in Canada. I am surprised I haven't seen any reviews of it yet,
it should be a decent over clocker for the enthusiast crowd.

---It's usually not worth the trouble to do much of anything for a bottom of the line product. Plus, you have to be careful not to have so good a lower end product that it steals away sales from your higher end. My guess would be that the transition to the 0.13u Process should result in a smaller die size, and lower power requirements. The overclocking ability should scale as well as the 0.13u P4.

One thing, is that Celeron 0.13u is still on the 400MHz Front Side Bus, where P4 is moving to the 533MHz bus. There's been some reviewers complaining about that, but that's specifically done to differentiate it from the P4. Intel probably won't change that. From intel's point of view, you want the higher bus performance of P4, pay the higher price of P4.

P.S.- I didn't find too many reviews with a Google search. Sorry.

Semi

Chip Makers Try to Score With Gamers
http://www.latimes.com/technology/la-fi-chips22sep22(0,69134...
Computers: Intel and AMD boost marketing, encourage development of faster software in hopes more consumers will upgrade their PCs.
By ALEX PHAM, TIMES STAFF WRITER

When the going gets tough, the tough play games.

With the semiconductor industry in the doldrums, chip makers Intel Corp. and Advanced Micro Devices Inc. are coaxing game developers to churn out bigger, faster games to give consumers a reason to run out and buy new computers......

Semi

VIA Tries Again : The P4X400 Chipset for Pentium 4
http://www.gamepc.com/labs/view_content.asp?id=p4x400&page=10

The Final Word
If you've read through our article to this point, you most likely know that our impressions of the VIA P4X400 chipset and the Soyo Dragon Ultra Platinum Edition motherboard haven't been entirely favorable. While the P4X400 seems to be a step up from what we've seen with VIA's previous Pentium 4 chipset, the P4X266A, it seems as soon as the P4X400 finally hits the market, it's practically out of date already.
The performance of the chipset simply isn't up to par. Even with CAS 2/2/2 DDR-333 memory at the highest speeds certified for the chipset and motherboard, performance of the Soyo Dragon Ultra board couldn't outpace our SiS 648 DDR-333 or Intel 850-E RDRAM platforms. With Intel's upcoming DDR-333 and Dual Channel DDR-333 products on the near horizon as well, it looks like the P4X400 will be pushed further down the totem pole in no time at all. It's a pity. If it had actually hit the market in volume a month ago, it would have been touted as a true performance player. By today's standards, it's surprisingly mediocre.

---Not good for VIA, IMO. The CEO is making the same mistake that Jerry Sanders made, letting his personal feelings get in the way of business. When you try to rub Goliath's nose in it, you better be sure that your slingshot kills with the first blow. If not, it's unlikely you'll get another chance.

Semi

RE: Hmmm, it doesn´t come as a surprise to me either, but still, UMC was originally supposed to produce SOI Bartons even (the article states that as well)!?! It makes sense that, now that Barton is manufactured at FAB30, to go with SOI for them (which also seems to show a certain amount of confidence in it). But I still can´t understand why AMD would overestimate UMC´s capabilities as they apparently have done.
As for the rumours, it seems that quite a few were simply made up. I do think that Heye´s statements are closest to reality, but we´ll see about that.
KD2000

KD, not surprising to me either. I'm guessing that UMC is discovering what I've known for a long time. Making micrprocessor logic circuits, is a heck of allot harder than making DRAM. Many times, if a wafer yields poorly at end of line, we'll peel the chip back, layer by layer to determine root cause. Rarely, do we find something wrong in the cache areas, most of the time, the issue lies in the logic circuits. Logic circuits are much harder to make than DRAM Cache. Something that I believe UMC is beginning to realize also.

Semi

Re: "The postponed launch of the K8-core processors is indeed due to problems in the back-end verification of SOI technology and minor changes in processor design, Richard Heye,
vice president of platform engineering and infrastructure at Advanced Micro Devices (AMD), confirmed on September 20 during a visit to the company's partners in Taiwan."

Well SOI technology is a "front-end" issue, so it isn't that ... How about LowK/Copper integration ??!! That's a back-end issue !! Sounds like AMD is in a lot of trouble with the .13um Copper/LowK process. This was discussed and warned over on the SI thread ... Of course the AMDroids there
are very "blinkered". <ggg>
Make It So,
Yousef

--Suppose what he Meant by the statement of "Back End Verification", was "End Of Line Yield and Bin Splits are bad, due to the Front End SOI"? These Non-Process types get their statements from the Process Engineers, but somtimes it is difficult to get the correct picture through the heads of people who are not involved in the Process. That would make your statement true also..... Right? Still, not good, if there is an issue with SOI, it would affect most of the Front End Layers. Depending on what exactly is the problem, based on the need to delay, it doesn't seem like a small issue to me.

Maybe intel was telling the truth after all about the lack of benefit of Partially Depleated SOI. JMO

Semi

P.S. - What's an SI Thread, some New Process Defect??

Semi, can you please take a look at this post here, and comment on the presentations. I would really appreciate hearing from a process expert on this new technology. Thanks.

O.K., just so you know, I'm looking at it from the process side, ie: making the gates, so I would have to take intel's word about the actual performance of the gates themselves. That would be more a Device Engineer.

That being said, I would really LOVE to have lower Poly aspect ratios illustrated on page 10. The comparison of the Source/Drain Tsi Profile Height seems to suggest that the tri-gate would have much lower height requirements. Maintaining profile height, not to mention sidewall uniformity is one of the most difficult parts of the manufacture. One thing I also noticed, is that the tri-gate seems to eliminate the isolation layer present in the Dual Gate. Elimating a Layer reduces costs.

The second thing that caught my eye, was on page 20, where intel states that Multi leg tri-gates increases total current. More current good, less current bad. An oversimplification maybe, but explaining would take pages.

The third advantage I see is the layout width of 0.60 of current transistors. If I'm reading that correctly, this should reduce die size significantly. Could you imagine 2 or 3 meg of cache..... on a P4? With the same die size, and the same power requirements?

One thing I can't help noticing, is there's no discussion of..... Cost. That's an important thing to leave out, and I would have to assume that if it were cheaper, intel would say it. They don't. Overall, it looks like quite a breakthrough, but I have a sneeking suspicion that we'll see it in Itanium, and Xeon, way way before you see it in P4 or Celerons...

JMO

Semi

AMD's Dresden fab is reason for delays
From the Inquirer, so take it with a grain of salt:
AMD's Dresden fab is reason for delays
Trouble at the chip mill

I remember when I was working on a Process that was trying to sqeeze every last little bit of performance out of a design on an old process, because the New One wasn't ready.

We had to limit ourselves to the 1 Best performing Litho Stepper, and even then, Etch was meeting with Litho twice per day to review the Steppers Focus/Exposure data vs. the Post Etch "Critical Dimension Data", to retarget the Stepper. It was a Huge amount of labor and tool downtime, and even then, the Bin-1 (Highest Speed) Yields were..... less than desireable.

Pure speculation of course, but I wonder if the same thing is happening at AMD?

Semi

RE: This brings to mind another topic that I feel is necessary to discuss.....

Good points. I actually went back yesturday, after reading your notes, and re-read my first 3 posts, and edited accordingly. I think one of the biggest mistakes that I made over on SI, was to allow the obvious Flame Baiters (Well Know People) to get under my skin, instead of just setting "ignore". You know how it is, if you could go back and do it over, you might do it differently? Hindsight is 20-20. I agree with with the post that I believe Greg made on the iHUB AMD Thread, Water Under The Bridge. I believe I can make a contribution to the knowledge base of this board, relating to Semiconductor Manufacturing Process concepts and/or issues, and so I commit to ignore and set filters for similar baiting of myself here.

Semi

P.S. - By the way, I stink at spelling (Hooked On Phonics), so please excuse any future spelling errors, you'll usually get the meaning of what I'm saying anyway.