1.18.2008

The comeback kid? AMD's Q4/2007

SUNNYVALE, Calif. — Jan. 17, 2008 — AMD (NYSE: AMD) today reported fourth quarter 2007 revenue of $1.770 billion, an 8 percent increase compared to the third quarter of 2007 and flat compared to the fourth quarter of 2006 2 . In the fourth quarter of 2007, AMD reported a net loss of $1.772 billion, or $3.06 per share, and an operating loss of $1.678 billion. Fourth quarter net loss included charges of $1.675 billion, or $2.89 per share, of which $1.669 billion were operating charges. The non-cash portion of the fourth quarter charges was $1.606 billion.

This is quite a comeback considering so much negative news surrounding the company. The substantial increase in revenue while keeping OpEx flat helped drive the company to almost break even. In fact the computing group reported an operating profit of $21M. Overall AMD beats the overly negative market consensus of $-0.36 per share reporting in a loss of only $0.17 per share. Margins were up 3% to 44% which would lead anyone to wonder what could have happened if Barcelona was executed as planned. While the $1.606B goodwill charge significantly affects the valuation of the company, one can argue that the stock seem to have already taken this into account.

From the company’s business standpoint it appears that AMD is doing all the right things. Key to the improvement is the change in priority. Like Hector said, their number one goal is profitability while the second is serving the customers. Essentially that means AMD is optimising product mix to generate the most revenue as opposed to running both their Fabs at peak volume just to grab market share. It seems like Hector has indeed humbly learned his lesson and breaking the monopoly is off the agenda.

From a product execution standpoint AMD says that B3 is ready. The conference call painted a better picture of how the Barcelona fix is progressing. Engineering samples will be out in a couple of weeks while production samples will be shipped later in Q2. This sounds like Barcelona will only reach the general market right around end of Q2, probably Q3. AMD shipped close to 400,000 quad cores with the ratio of 2:1 (desktop to servers) in Q4. It is difficult to gauge the significance of that quantity in terms of margins but clearly it is just a large number thrown out there to impress people. It appears like it worked.

While it is evident that AMD is trending in the right direction, their strategy remains untested. Revenue may be trending upwards but so was demand in the 2nd half of 2007. The company grew revenue using existing products but their viability is relative to the competition who seems to be executing at a faster pace. AMD’s plan to return to profitability relies heavily on new products which appears to be limited to the mid to low-end segment (tri-core, 65W products). While this was proven to be marginally profitable in a healthy market, AMD needs to prove they can do the same feat in the midst of a price war. One can’t help but wonder if the actions of another entity plays a bigger influence on AMD’s profitability.

48 comments:

Roborat, Ph.D said...

(another continuation)

Since Q1’08 is seasonally down, AMD isn’t expecting to be profitable. The manner by which they reached their financial performance was quite desperate. They admitted that the savings they acquired through forced vacations won’t be replicated in the current quarter. They expect OpEx to rise in Q1 by 5%. I’m suspecting K10 production cost is starting to kick in as well. At the current margins AMD needs to be north of $1.8B to break even which is why AMD now aims profitability in the 2nd half.

AMD also says 45nm silicon is ‘out’. They expect to start ramp in the 1st half with revenue shipment in the 2nd half. Why AMD chose the word ‘revenue shipment in 2nd half’ instead of ‘out the market’ or ‘OEM shipment’ clearly means we wont see 45nm parts from AMD until 2009. That is more than a year from Intel’s 45nm availability since Nov 2007.

Ho Ho said...

"They expect to start ramp in the 1st half with revenue shipment in the 2nd half"

http://www.tgdaily.com/content/view/35672/135/

"Ruiz said that AMD that 45 nm Opterons have been produced and he is “pleased” with the initial result. However, he said that the company is targeting a H2 ramp of the 45 nm generation, which indicates that AMD will miss the originally promised mid-2008 launch of the processors. This statement also contradicts Meyer’s statement from the Q3 2007 earnings call that AMD would begin ramping 45 nm CPUs in H1 2008."

Anonymous said...

I think folks are misinterperting 'ramp'... clearly when AMD said ramp in H1'08 they meant equipment and starting production, and happily let folks think this meant product shipping...

I think AMD will ship some 45nm products in 2008, these may be 'samples' and/or products not quite available for purchase until 2009 (but 'shipped' ala 65nm), but AMD spinmeisters will be in full force saying - see execution to plan, we're closing the gap...if you have questions I will answer them when returning from forced vacation!

Also you folks need to use the AMD decoder ring 'mid-2008' means Sept or later (see K10 for details). Also 'launch' doesn't have to mean having actual prodcut launches (unless it is a 'hard launch across the board', in wich case means a handful of products will be available, and only certain products)

Why can't a SINGLE press person just ask AMD when customer will be able to BUY A 45nm PRODUCT! How hard is that? Why must the press be so lazy and not even ask what the heck AMD is parsing words on?

All said AMD had a reasonably good quarter, it's not clear to me it is sustainable, especially if the economy slows and Intel starts up with pricing pressure again. Also cutting costs works short term but cutting cost is not a sustainable 'growth' plan (unless AMD thinks they can continuosly grow the cost they cut each quarter!)

I'm also curious about Asset light as the INQ is saying AMD is mulling the NY fab - not sure if this is typical INQ B.S. Especially as AMD appears like it will struggle to convert F30 and I don't see how they need a 3rd fab with 'asset light'

Ho Ho said...

"I think folks are misinterperting 'ramp'... clearly when AMD said ramp in H1'08 they meant equipment and starting production, and happily let folks think this meant product shipping..."

In that case Intel is "ramping" Larrabee right now, starts product shipping in the end of this year and will launch early next year.

Of course that really means that they get the first silicon late this year (but they are preparing for it, it's the same as ramping!), they start sending out engineering samples next year and hope to launch H2 '09/H110.


"Why can't a SINGLE press person just ask AMD when customer will be able to BUY A 45nm PRODUCT! How hard is that?"

My guess would be that they know they won't get anything better than "This is not decided yet, it depends on what customers want/need. Of course all dates are subject to change" so they don't even try.

Tonus said...

Even when they know they'll get a canned response, the press normally is willing to ask those questions. It gives them a springboard from which to comment, and there's always the chance that they'll get a response that is different than expected.

I do find it aggravating that analysts and reporters aren't willing to ask such point-blank questions. Ramping and testing and all of these other terms are all well and good, but I am more interested in when I can actually buy an item. And the way AMD fudges their terms, we really can't tell what they mean.

So ask them already. Force them to either commit, or weasel out, or even just honestly admit that while they have all of these targets and are "happy" or "satisfied" with how things are going, they haven't a clue when they'll actually be able to put a product in your hands.

SPARKS said...

"AMD’s plan to return to profitability relies heavily on new products which is appears to be limited to the mid to low-end segment (tri-core, 65W products)."

This is all well and good in a perfect status quo market. However, there is a little company out there by the name of Intel Corporation, and they are quite a competitive company, I hear.

I wonder what will happen to this AMD best case scenario when, and if, INTC drops prices on ANY part of its lineup during the seasonally low 1Q to boost sales in the face of the current economic downturn. What pressure would this put on AMD'S ASP's, margins, and hypothetical profits?

A sub $200 Q6600 and E6850 would certainly put a damper on this rosy, perfect world, 1Q AMD market picture. Would it not?

I know what I would do.

SPARKS

Anonymous said...

"A sub $200 Q6600 and E6850 would certainly put a damper on this rosy, perfect world, 1Q AMD market picture. Would it not?"

Not just that, but the real battle will be in mobile... Intel will be pushing down with Penryn based mobile this year and will soon be pushing up in the low end space in 2009 (via Silverstone). This will really squeeze AMD as there gains have been in the low cost notebook market - if Intel can produce a lower cost product via silverstone in a market where cost (not performance) is the greatest concern, the Athlon/Puma will have no place to go as Intel's 45nm mobile will prevent those products from climibing the ladder.

Nehalem will soon be squeezing the most lucrative AMD market (4P+ servers) and Barcelona will become the new Athlon - price slashed to compete with a superior product. Only Barcelona 'black edition' just won't cut it in server world.

My guess is Intel will cede the low end desktop market until they have the capacity to push on this as well. If AMD loses significant ground in mobile and 4+ space, they will never become profitable.

AMD's best hope is the market slowdown is overblown, and they can start paying down debt as the overall market grows before Intel puts the capacity on line to really smoke AMD on pricing.

SPARKS said...

” My guess is Intel will cede the low end desktop market until they have the capacity to push on this as well. If AMD loses significant ground in mobile and 4+ space, they will never become profitable.”

“AMD's best hope is the market slowdown is overblown, and they can start paying down debt as the overall market grows before Intel puts the capacity on line to really smoke AMD on pricing.”

You’ve read my mind, my sentiments, exactly.

SPARKS

Anonymous said...

As the financials are in for 2007, perhaps it is time to look at:

"2007: A Year of Promise"

'Perhaps because of the lack of gains by Intel there has been a flurry of negative rumors about AMD lately. This includes ideas that AMD is behind in 65nm, is having problems with 65nm, and won't deliver K8L on the desktop until Q4 07'

What were those crazy naysayers thinking!

'Desktop versions of K8L will come out in Q3. Apparently, the fact that the intitial supply of K8L is tight because of preordering has led some to believe that production won't start until Q3.'

Yeah more crazy talk....

'This would then have 45nm being released from Intel before AMD has any desktop K8L's.'

Hmmm...

'With normal expansion, Intel is likely to continue to slowly lose share to AMD.'

Ummm...not so much...

'This may seem unfair however AMD has cultivated better relationships with customers than Intel has in the past.'

Yeah the channel was VERY happy with AMD this year. Cray was ecstatic and I'm sure the OEM's were pretty pysched up!

'Overall, however, 2007 should see Intel's and AMD's product lines and graphics support move much closer.'

Yeah...about that...

'By end of 2007 Intel should have something to compete in 4-way but it will have increased competition in 2-way and the desktop.'

Umm...Intel is clearly struggling competing in desktop and 2-way server.

'Intel will have to be on its toes in 2008 to stay competitive and 45nm by itself won't be enough.'

Gotta hand it to Dementia on this one... 45nm by itself will not be enough, it looks like BOTH 65nm and 45nm will be enough! (oops...)

"Looking further ahead Intel proponents are quick to point to CSI but Intel will then be facing an AMD with two functional 300mm"

And that 2nd functional fab is... Chartered?

It's kind of funny - Scientia's forecast was that Intel would need the market to grow to do reasonably well as AMD would be capacity constrained and Intel would have a tough time competing... well 2008 is exactly the same - except reverse AMD and Intel. AMD better hope the market grows in 2008.

I'll be looking forward to his "2008: Another Year of (Unfulfilled) Promises"

enumae said...

Sparks
"...if, INTC drops prices..."


Projected price cuts.

SPARKS said...

Oooh, Enumae, you beat me to that one, buddy, as I scourer the sites daily. I missed that little tidbit. Thanks!


SPARKS

Anonymous said...

anonymous said...

"As the financials are in for 2007, perhaps it is time to look at:

"2007: A Year of Promise"


Obviously it's a good thing Scientia doesn't depend on his "scientific" acumen or prognostication prowess to make a living. Otherwise his corpus would have shriveled as much as his blog readership and credibility have.

SPARKS said...

Comeback kid my ass.

WARNING TO AMD:

BE AFRAID, VERY AFRAID



http://www.google.com/products?q=
Intel+e8400&btnG=Search+Products&hl=en&show=dd

http://www.google.com/products?hl=en&q=
Intel+Q9450&um=1&ie=UTF-8


SPARKS

Chuckula said...

Alright, this post is in response to censorman Scientia and his stupid AMD cult, posted here for freedom of speech purposes. As a recap, his latest post tries to put as much lipstick on the AMD pig as possible in light of their Q4 earnings, and reiterates the common fanboy belief that Intel's 45nm process is somehow completely broken:


So Scientia, you bring out the new AMD talking point that because Intel is delaying some mainstream quad core parts by about 1 month that this means Intel's 45nm process is a disaster and that AMD will fully catch up by next Tuesday. (Say, isn't 1 month about 1/8th the delay that AMD is taking just to ship a 2.4Ghz K10???) As usual this requires AMD fanboys to believe 2 conflicting ideas at once and drink the kool-aid:

1. Intel doesn't ship ANY quad cores! Those are just glued-together chips and nobody has ever bought any of them! Only AMD ships "native" quad cores!! Performance and power dissipation don't matter!

2. Oh look, Intel is delaying some quad cores by a month! Intel relies on Quad Cores to stay in business!! They are the highest volume part, and if there is ANY delay then Intel's entire 45nm process is lost for all of eternity!! It doesn't matter that Intel executed perfectly on the dual core mobile and desktop parts, nobody ever buys those anyway! And the 65nm chips will disappear in a poof of smoke too! Intel is doomed!



The fact that I ordered my e8400 THIS WEEKEND FROM A SUPPLIER THAT HAS THEM IN STOCK must mean... I dunno that Intel is doomed or something according to the AMD psychopaths.

I actually think that AMD fanboys are so conditioned to believing that delays and underperformance are so good that they think Intel's problem is not that it delayed introducing some parts, but that the delay is so short that the parts will actually be available just a month later. The real path to success is to have a propaganda rally to worship your inept CEO and then dump defective parts into the channel 6 months later!

Anonymous said...

"Intel 45nm is not yet generally available well after the supposed launch. A quick look at NewEgg shows only 3 items currently available: E5440, E5450, and QX9650. No mobile."

Yeah it is shocking that 45nm mobile parts are not available yet...especially as they WERE'T PART OF THE NOVEMBER LAUNCH! As opposed to the K10 parts that were 'launched' last Sept and the Phenom parts in Q4? How many Phenom SKU's are available on Newegg?

Apparently when parts don't show up that Intel HASN'T LAUNCHED YET, it should be treated as a delay?

The guy is really in denial now. If anyone posts anything there that he dosen't like he says it is wrong and unsubstantiated (or just censors it). However he can readily make claims like 'issues' with 45nm and have no supporting DIRECT data on this, but it should be taken as gospel! Even if you believe the 45nm parts are delayed ("IF"); could there not be several reasons for this... it vould be the stepping (shrink issue), it could be a process issue, it could be a demand issue...

As always he takes a singular observation, ignore all possible explanations and zeroes in on the one that feeds into his pre-determined conclusion. The guy is a joke!

He also claims that 45nm is not so great based on the fact that the very FIRST 45nm stepping is only marginally better than the FINAL mature 65nm stepping.... Gee I wonder if he will apply that same lgic to AMD's 45nm process when it first comes out? ...and what does it say about AMD's CURRENT 65nm process if you compare it to 90nm?

InTheKnow said...

Is anyone really surprised to see moderation turned back on at Scientia's? I really think the best thing to do is just avoid the site altogether.

I wish there were a rational AMD centric blog out there, because I think an opposing view to my admittedly Intel biased view is a good thing. But I've learned the hard way that Scientia's blog isn't that place. The heavy handed censorship prevents an open an honest discussion.

Of course my final post on his site to that effect was immediately removed.

Chuckula said...

Scientia: First of all I'm cross posting this at a real blog so your censorship won't work.
Second of all, there is NO delay on any of the dual core mobile or desktop chips in 45nm. For mobile, the reason you don't see them on NewEgg is because they all go to OEMs since people upgrading mobile CPUs in laptops is EXCEEDINGLY RARE. Please show me OEMs that are complaining about not having chips. As for the desktop chips, I personally got my e8400 from Tankguys. By later today it will be fully available on the bigger sites because they tend to have more respect for Intel's blackout dates.

Secondly, since you live in bizarro world where you (and not Intel) are the judge of when products are "supposed" to launch read this post from Anandtech WRITTEN IN AUGUST OF LAST YEAR:
Penryn's Launch Schedule

The launch will work like this: late this year (Q4), Intel will introduce a a new Extreme Edition processor based on Yorkfield (quad-core Penryn). Given that most Extreme Edition chips start out with a $999 price tag, that's where we'd expect the first quad-core Penryn to fall as well.

More affordable Penryn won't arrive until Q1 2008, which is great news for AMD. In the beginning of next year, Intel will introduce both mainstream Yorkfield and Wolfdale (dual-core Penryn) parts. We'd expect Wolfdale based chips to run at between 2.33GHz and 3.00GHz, while Yorkfield will probably top out at 3.33GHz.

In Q2 2008, Intel will push Penryn even further down the line and introduce a Wolfdale based E4xxx successor. This version of Wolfdale will have a 1066MHz FSB and drops support for Intel Virtualization Technology and TXT.


This is the link: http://www.anandtech.com/showdoc.aspx?i=3069&p=2

So it looks like Intel said it would be coming out with the mainstream 45nm parts in January... and guess what they did! The only thing wrong in the rumors from 6 months ago are a minor bit about clock speeds (but everyone can overclock to 4Ghz anyway) and the recent announcement of a minor delay in some quad core parts. Please show me one single solitary promise that AMD has made (other that a promise to lose money) that has been even close to being fulfilled like this was.

Roborat, Ph.D said...

First of all, by Q3 AMD should finally have at least 2.8Ghz (K10)chips.

considering that everything he predicted in 2007 turned out completely wrong, i'm not surprised. If AMD can't hit 2.4Ghz with a crippled tri-core. What makes him think AMD can do it with a quad.


A takeover is technically and financially possible except that I can't think of a single company that wants to get into the frontline processor business and slug it out with Intel.

Private equity investors? There's tons of them. Where to begin?
KKR, Merrill Lynch, etc.
AMD at $3B plus $5B debt is peanuts to these guys.

Axel said...

Scientia censorman

A quick look at NewEgg shows only 3 items currently available: E5440, E5450, and QX9650.

I see that he conveniently forgot about the 2.83 GHz and 3.0 GHz 45-nm Harpertown quadcore server chips that have been available on Newegg for weeks now. The slower ones are backordered, no doubt because 65-nm Clovertown currently has no competition so why should Intel introduce Harpertown en masse and cannibalize all those sales?

InTheKnow said...

I think Scientia's claim that all 45nm chips are from D1D is spurious as well.

Perhaps Orthoganal can confirm that F32 (just across the parking lot from F12) is indeed up and operational?

Though why we should believe anyone who can talk to the guys that work there when we have a pronouncement from Scientia is beyond me.

enumae said...

Scientia said
...and the power draw is just a little better than the G0 stepping...


XbitLabs

TechReport

A 25-30% decrease in power consumption under load is not "a little better".

------------------------------

I tried posting this on Scientia's Blog, but I don't think he will post it, so, can someone explain how he can conclude that a reduction of 25-30% considered "a little better"?

Tonus said...

My only question is in regards to statements like this one:

"AMD first produced Brisbane samples in Q2 2006 and delivered product in Q4. If AMD has samples in Q1 then product delivery should by Q3."

I am assuming that he is stating that AMD went from sample to shipping product on 65nm within ~6 months, and that therefore it is reasonable to expect them to go from sample to shipping product on 45nm within ~6 months.

If this is the case, is it a reasonable assumption? Or are there enough variables in processor design and process shrinks to make such an assumption highly unreliable?

ConScientia said...

can someone explain how he can conclude that a reduction of 25-30% considered "a little better"?

Easy.

Intel improvement of 25-30% = small improvement
AMD 1-3% improvement = breakthrough

Roborat, Ph.D said...

...therefore it is reasonable to expect them to go from sample to shipping product on 45nm within ~6 months... If this is the case, is it a reasonable assumption?

where are these 'samples' and what conditions are they in? It's fairly safe to make such statements considering how vague the term 'sample' silicon is. One can even argue that Intel at this point in time have 'sample' 22nm silicon. Sure they're not working silicon but hey, it's a 'sample'. The guy doesn't know what he's talking about by making a technical milestones off from a vague marketing reference point.

Orthogonal said...

InTheKnow said:
Perhaps Orthoganal can confirm that F32 (just across the parking lot from F12) is indeed up and operational?


I didn't realize people thought this was an issue, it was widely publicized that F32 received factory certification and was officially open back in October '07. Risk starts are generally well into the line when that happens too.

I honestly have no idea how much product has shipped from there, but it has been going for quite some time now. F32 hasn't reached 100% capacity yet since there are still ramping tool quals at this time, but they are running max wafers for the supported capacity...

Chuckula said...

you think that Intel is doing a paper launch? You think that newegg has no products?

CHECK IT OUT: http://www.newegg.com/Product/Product.aspx?Item=N82E16819115037&Tpk=e8400

Or maybe newegg isn't good enough all of the sudden?

TRY HERE: http://www.zipzoomfly.com/jsp/ProductDetail.jsp?ProductCode=10007603

Oh look... It says "In Stock" for an e8400. That's what we call a hard-launch Scientia, it means that you actually put out a product that people can actually buy on the actual dates that you anounced 6 months ago. I'm getting sick of this bullshit that Intel's 45nm process is somehow completely broken when they've shipped more fully operational 45nm parts in the last 24 hours than AMD will all fucking year.

SPARKS said...

http://www.zipzoomfly.com/jsp/
ProductDetail.jsp?ProductCode=
10007603


GET ONE TODAY!!!!


enough said.


SPARKS

Anonymous said...

"...therefore it is reasonable to expect them to go from sample to shipping product on 45nm within ~6 months... If this is the case, is it a reasonable assumption?"

Reasonable is tough to say... once again Scientia is using EMPIRICAL #'s and applying them to the future. This is much like saying a stock has gone up every year in the month of Feb, so I should bet the farm on that stock right at the end of January this year and not look at anything else. As always you need to understand the fundamentals behind things in order to make a prediction/assumption like this and quite frankly Scientia lacks the experience, technical education, and real life manufacturing experience to speak authoritatively on this. Simple having a blog doesn't qualify him to make these statements, though it certainly doesn't stop him!

So back to the reasonable question... is it possible... YES... is it reasonable? impossible to tell.

Here are some factors:

1) What is AMD's targets for initial 45nm products. If they do like they did with the 65nm transition (release products at 3-4 speed bins below the previous generation, that increases the probability of going from sampling to production rather smoothly.
2)If by "sampling" AMD means fully functional chips at their OEM's for them to develop BIOS's, then that also increases the probability. If sampling means they are internall debugging it, that is a whole different story. As usual in AMD' recent history, they choose to be very vague and non-specific with their claims so they can back out of things if they need to down the line.
3) What yields is AMD willing to live with on 45nm? It is easy to get a few center die yielding and get those die out for sampling. In this hypothetical case, are they willing to start ramping with less than ideal yields for PR purposes? If so this makes the likelihood of going to prodcution better. Keep in mind and functional and sampling chip can be made on a poorly preforming and poorly yielding process... addressing these types of issues (if they exist) can require a lot of effort.

So the main risk is that an additional stepping or two will be required - this could completely blow the schedule out of the water and lead to some of the issues you see with K10 schedule. This could be for yield issues, speed path issues, power issues, design issue (with the shrink) to name a few reasons.

As 45nm for AMD is essentially a dumb shrink with minimal process improvements I think the risk of things being vastly worse (at least no worse than 65nm already is) is fairly low. Also it is unlikely they will massively screw up a simple, or to use an industry term, "dumb" shrink. And AMD seems perfectly happy spinning out new products under the performance of previous generation through claims of "OEMS and customers" demanding it, so it appears that won't stop them here either.

So bottom line it is likely a decent assumption... the flaw in his argument is that 45nm will not be THAT MUCH cheaper than 65nm and the ramp rate will not be that steep - so margin impact will be rather small in 2008, especially as the initial 45nm products will likely be low ASP parts as they will in almost all certainty be non-top bin parts for AMD.

Anonymous said...

"Some might also have noticed that AMD has scaled back its expected 2008 volume from 100 million units to 80-90 million units. "

A minor footnote lost in a 100,00 word blog... This is a 10-20% lowering of expoectations and is HUGE! I don't buy the simple, this is because F38 ramp is lower than expected, because if the demand was truly there, AMD has foundry capacity they could use at Chartered! I think this lowering represents a mix of both realization of a slower overall 2008 growth and a softening of demand due to the competitive situation with Intel. Clearly if AMD was back in the 'we're selling everything we can make' hay day of 2006 then they should be able to ramp up the Chartered outsourcing capacity to offset the slower F38 ramp.

As usual another bogus argument and more excuse mongering. AMD just had a >0.5Bil cash infusion, if they really needed the F38 capacity and the incredible demand for K10 is real, then they would ramp it more aggressively, or ramp the foundry capacity - the lowering (BY AMD!) of unit volume expectations is not indicative of supply/capacity issues, it is a tell tale sign of lower than previously expected demand! To deny this is simply burying your head in the sand.

To put a capper on the supply excuse...
- 10Mil chips is ~200K per week...
- let's assume AMD gets 200 die/wafer (nless we are talking all quadcore this is lowballing things)
- That's 1000 wafer starts per week! (if you factor in dual cores,and more die per wafer this # gets even smaller!)

You don't think if they needed to AMD could make this up by increasing starts in F36 or at Chartered! Come on! Yeah a supply/capacity issue!

InTheKnow said...

Scientia said
...and the power draw is just a little better than the G0 stepping...


To which enumae replied...
A 25-30% decrease in power consumption under load is not "a little better".

and provided links to back up his numbers.

Scientia (to give him credit) replied...

Yes, I'm going to have to agree with you. The 45nm E8200 at 2.66Ghz does draw 1/4 less power than the G0 stepping 65nm E6550 at 2.3Ghz. And, these seem to both be in the same 65Watt class. That is pretty impressive. So, yes, 45nm is going to be important for Intel for high density server applications.

So all well and good. He modified his blog to correct his error as well.

This just leaves one question in my mind. Why make such a bogus statement in the first place? It isn't like Penryn's power savings is a big secret. The links enumae provided are not new or hard to find.

If Scientia has done his homework that poorly on such an obvious, easily verified claim, it makes me question the validity of his other assertions as well.

InTheKnow said...

On a less happy note for some, it looks like Intel is pushing out Larabee.

The Ars Technica article doesn't give any reason for the apparent slip.

Anonymous said...

'This just leaves one question in my mind. Why make such a bogus statement in the first place?'

Because it is up to readers to disprove ANYTHING Scientia says, while anything anyone else says has to be supported and is assumed to be incorrect. Thus Scientia can simply make up 'facts' to support his already formed conclusion.

Conclusion - Intel 45nm is no good... let's see how can I support that? I know, I'll say it is only marginally better than 65nm. I'll also say they have no 45nm mobile parts yet (even though Intel said these wouldn't launch into 2008). Then I'll just say yields are bad because I know it'll be impossible to prove or disprove this statement. Then I'll just sprinkle in something about only D1d producing parts... heck it's not like I have to support any of this crap, I'll just keep throwing up barriers until people give up and support my version of the 'truth'.

The more bogus statements you throw out the more likely one is to be correct...even a stopped watch is right twice a day! The other reason is simple denial... he doesn't want Intel to be better at anything, so he will not carefully read any review which might contradict this perception.

His discussion on cash on hand (somehow as a metric of health for intel) and his admission of mistakes in that area is also laughable. In admitting his error on the calculation, he completely missed the forest while looking at a tree. Cash on hand for Intel is a useless metric - there is dividend, stock buyback and other cash expenditures which make this less of an ideal metric for evaluating health.

I must concede it did serve one very important point - as little as he understands about process and technology, he may know even less about financials.

Next thing you know he'll be saying AMD is closing the process technology gap!

InTheKnow said...

I must concede it did serve one very important point - as little as he understands about process and technology, he may know even less about financials.

Now that is a frightening statement, because he knows almost zilch about process. And to make matters worse, he has alienated everyone who does know something about the process side of things and driven them off of his site.

enumae said...

IDC Unit Share

"Processor vendor shares in 4Q07 did not change significantly from those in 3Q07. On an overall unit basis, Intel earned 76.7% market share, a gain of 0.4%. AMD earned 23.1%, a loss of 0.4%. These shares are identical to the shares of 2Q07.

By form factor, market share changes were very modest. In the mobile processor segment, Intel earned 81.9% share, a gain of 1.1% and AMD earned 17.8%, a loss of 1.1%. In the PC server processor segment, Intel earned 85.4%, a loss of 0.6% and AMD earned 14.6%, a gain of 0.6%. In the desktop PC processor segment, Intel earned 72.1% share and AMD earned 27.7%; share changes were negligible.
"

Giant said...

CUPERTINO, California—January 22, 2008—Apple® today announced financial results for its fiscal 2008 first quarter ended December 29, 2007. The Company posted revenue of $9.6 billion and net quarterly profit of $1.58 billion, or $1.76 per diluted share. These results compare to revenue of $7.1 billion and net quarterly profit of $1 billion, or $1.14 per diluted share, in the year-ago quarter. Gross margin was 34.7 percent, up from 31.2 percent in the year-ago quarter. International sales accounted for 45 percent of the quarter’s revenue.

Apple shipped 2,319,000 Macintosh® computers, representing 44 percent unit growth and 47 percent revenue growth over the year-ago quarter. The Company sold 22,121,000 iPods during the quarter, representing five percent unit growth and 17 percent revenue growth over the year-ago quarter. Quarterly iPhone™ sales were 2,315,000.

“We’re thrilled to report our best quarter ever, with the highest revenue and earnings in Apple’s history,” said Steve Jobs, Apple’s CEO. “We have an incredibly strong new product pipeline for 2008, starting with MacBook Air, Mac Pro and iTunes Movie Rentals in the first two weeks.”

“Apple’s revenue grew 35 percent year-over-year to $9.6 billion, an increase of almost $2.5 billion over the previous December quarter’s record-breaking results,” said Peter Oppenheimer, Apple’s CFO. “Our strong results produced cash flow from operations of over $2.7 billion during the quarter, yielding an ending cash balance of over $18.4 billion. Looking ahead to the second quarter of fiscal 2008, we expect revenue of about $6.8 billion and earnings per diluted share of about $.94.”


http://www.apple.com/pr/library/2008/01/22results.html

Very impressive, the Intel Mac sales are increasing at an incredible rate.

SPARKS said...

Boy, oh boy, ya gotta hand to Dementia for calling the INTC delays on 45nM! Man, the clocks, well, all hell must beak loose when ya get ‘em over 4 Gig! And, the price is so terrifyingly high!

I really don’t know why you guys even bother to click on the ASSHOLES website and read his bilge. This man is beyond ridicule.

As a side note:

AMD, be afraid, very afraid.

http://www.zipzoomfly.com/jsp/
ProductDetail.jsp?ProductCode=10007603


http://www.newegg.com/Product/
Product.aspx?Item=N82E16819115037



SPARKS

InTheKnow said...

I really don’t know why you guys even bother to click on the ASSHOLES website and read his bilge. This man is beyond ridicule.

To put it simply, the level of misinformation offends me on a professional level and I feel a need to correct the errors somewhere that open minded individuals might find the real information.

Anonymous said...

'Very impressive, the Intel Mac sales are increasing at an incredible rate."

It is impressive (though the stock is/will get pounded anyway), however it is easier to show large increases when you are starting at fairly low levels. This phenomena was also seen at AMD when they started gaining some market share, but it generally is not sustainable as the market share and volumes increase.

That said, the early evidence shows the switch to an Intel processor for Apple was a very good move and I could see Apple easily getting to 10% if they are wiling to cut prices a bit. And Intel locking this up for some time was significant - the exclusivity will not likely be permanent but clearly Intel having sole access to 6% of the market was/is a shrewd business move. At this point it is probably not worth Apple's time and effort to have a 2nd source in AMD, but if they grow, Apple will likely fall into the Dell trap.

'I really don’t know why you guys even bother to click on the ASSHOLES website and read his bilge. This man is beyond ridicule'

For me it is comical for someone so ignorant (in the true sense of the word) in certain areas comment so authoritatively on them and not even realize his mistakes even after folks try to correct him - it is simply amusing and a good study in fanboy psychology. It is also humorous to watch his 'following' (for lack of a better word), simply accept whatever he says without evidence and dismiss anything that contradicts him. I love the empty 'excellent analysis as always' comments.

BTW - where did Abinstein go? I miss his 'insights' into Intel's poor yields and his Si process 'knowledge'.

Giant said...


BTW - where did Abinstein go? I miss his 'insights' into Intel's poor yields and his Si process 'knowledge'.


Shortly after all the Phenom reviews were up I grabbed some of Abinstein's old posts including one where he claimed Phenom IPC would be higher than Yorkfield. He even had one post where he told me to "remember my words" when I claimed Phenom IPC would be inferior to Kentsfield.

I reminded him of these posts, and provided links to about a dozen reviews of the Q6600 'fragging' Phenom by an embarrassing margin.

I haven't heard from him since!

pointer said...

Giant said...

BTW - where did Abinstein go? I miss his 'insights' into Intel's poor yields and his Si process 'knowledge'.

...

I reminded him of these posts, and provided links to about a dozen reviews of the Q6600 'fragging' Phenom by an embarrassing margin.

I haven't heard from him since!


So ... it is all your faults? :)

or he might be busy fixing the TLB bug, asssuming he is professionally involved with that CPU ... after being too busy accusing others as FUDer inclusive of a true AMD supporter in AMDZone.

InTheKnow said...

I want to eliminate all the disinformation and confusion in this entry right now.

You and Enumae have both asked about 45nm production at Intel. Here is the official Intel FAB 32 announcement from October 25, 2007.

I have had some people here argue with me and try to claim that D1D is a high volume FAB. It is not. D1D is capable of real production but only at a fraction of Intel's regular FABs. This is clearly stated by Intel in referring to FAB 32.


I don't know what D1D's 45nm output is. But I do have reliable information that they were around 2.5K Wafer starts per week on 65nm. This is 10K wafer starts per month or between 30-50% of AMDs total capacity. It is true that this is small by Intel's standards, but D1D does put out significant volumes.

"Intel Opens First High-Volume 45nm Microprocessor Manufacturing Factory"

"Intel first produced 45nm processors in its Oregon development facility, called D1D, in January and is now moving into high-volume production with the opening of Fab 32."

The fastest that Intel could get Copy Exact running on FAB 32 would be 30 days later. This would be November 24. Then it takes 90 days for these chips to be manufactured and distributed. This would be February 22 at the earliest.


The author clearly pulled this piece of information out of his butt. He fails to take install and qual time into account where tool matching would take place. And 30 days to get the first lot from the starts to e-test is probably optimistic.

He is also under the delusion that when F32 opened they were just starting to work on getting the line into production.

However, a more likely scenario is 60 days from opening for one test batch then an additional 90 days for actual production and distribuation. This would be March 24. This is why I've said that all production so far is from D1D.

As orthogonal already pointed out, the line was already qualified with risk starts in line when the fab "opened". F32 would be pushing out product by the end of November with an October opening. Sadly, the world doesn't work in the nice leisurely way this author would have you believe.

enumae said...

"The first 45-nm production wafer rolled off the line several weeks ago." John Pemberton, the Fab 32 plant manager during a presentation officially opening the new facility.

Extreme Tech

I tried to post this on October 28, 2007, but I got no response Scientia.

Tonus said...

Well, the great thing about all of this is that at some point, these companies have to actually produce and sell a product, and then we can better determine who is right or wrong.

A lot of the technical details are over my head, but basic hardware reviews and product availability are not. However, I am grateful for the replies I got to my question. Thanks!

Anonymous said...

How many vacations can ruzi force his employees to take? I doubt theres enough days in the year for enough vacations for amd to make a profit. I'll bet they got that 45nm silicon just like they had those 3Ghz chips too. Theyll be lucky if they get that silicon into anyones hands by Jan 2009. If amd does try to push the clocks up to anything good like 3Ghz theyll fry. amd is finished.

pointer said...

Well, you can call me the cum back kid!

pointer said...

Anonymous pointer said...

Well, you can call me the cum back kid!


when did I post this??? Anyone know what was the loophole that someone can pretend as me? I know someone did that to Sharikou ... (shit ... now i am linking myself to Sharikou ... :) )

pointer said...

hahahaha, although he can pretend himself/herself to be me, but by just looking at the icon, then you can know whether it is a real me or not. My icon is a 'Blogger icon', while the pretender is with the 'Anonymous icon'

by simply copy the whole text, including the icon, then paste, it will reveal it too:
Anonymous pointer said...

pointer said...

need to suck some dick