8.29.2007

AMD’s Last Options

What can be more useless than a stock downgrade after the significant drop? The rating of "Speculative Risk” from Citigroup’s Glen Young is the last of a series of downgrade given to AMD. And the reasons this time around are:

“… reflects recent product delays, the continued challenging landscape for microprocessor and graphics processor pricing and the increased volatility of AMD’s shares versus its peers [Intel, NVIDIA]… we see little prospect for [market] share momentum in coming quarters, and expect poor financial performance as a result,” he wrote in a research note this morning. “We look for management to provide a viable restructuring solution as a means to revisit our rating on the shares...”

We’ve discussed AMD’s need to restructure several times since the beginning of the year and the pressure only builds up as AMD continues to delay any announcements. But to be fair, restructuring is easier said than done which is probably why AMD hasn’t fully come forward with any plans. Invariably this is due to the lengthy discussions with legal and creditors about which course of action to take. While some companies restructure voluntarily like IBM or HP others go for bankruptcy which allows protection from creditors and gives the new administration more power to make drastic changes. For AMD and just about any semiconductor company with short product lifecycle, bankruptcy is not a favourable option.

Normally I would say that AMD doesn’t have a lot of options but if I was forced to give a few examples, this is what they would be:

  1. Keep the status quo, ride out the bad weather and hope for a miracle that the business picks up at least until they break even. Microprocessor volume demand grows ~15% YoY so AMD can sit around with two Fabs and wait for demand to outstrip supply. The only problem with this strategy is that it may take forever all the while being at the mercy of Intel's ability to manipulate pricing.
  2. Shut down Fab30 at the end of 90nm. If AMD wishes to return to profitability, this is a quick and effective move. Running 2 Fabs was made on an assumption that margins would be at least >45%. Clearly the environment has changed and AMD needs to face this reality. While this also helps AMD avoid the cost of upgrading the Fab to become cost competitive, this reduces AMD’s fixed cost, limiting the day to day cash haemorrhage. The challenge with this option and any other plan to sell or reduce headcount is the covenant they have with the German/EU government.
  3. Sell ATI. They already have the technology they need. They don’t need the additional quarterly loses. ATI will only continue to wither away if it continues to be under AMD. Not being able to deal and cooperate with Intel with equal footing as NVIDIA is a major disadvantage ATI doesn’t need right now. Separating the two companies can only be good for both.
  4. Go 100% Fabless. AMD’s margins are now close to ATI’s that this may only be the only viable solution. They can only blame themselves for rapidly commoditizing the industry. Of course we’re not talking about something immediate but rather in steps similar to #2 – shutting down the Fab once the process node becomes obsolete. We’re talking at the end of 65nm.
  5. Aggressive outsourcing. This is the mostly likely option AMD could take with a massive production offload to Chartered or TSMC, mostly for low ASP volume runners. Capital avoidance at 45nm makes this extremely attractive next year.
  6. Focus on niche. This requires a massive u-turn in its philosophy of “breaking the monopoly”. AMD definitely has the technology and roadmap to exit commoditized segments while delivering only on key markets. This will require a massive reduction in production capacity and headcount but if executed perfectly will protect AMD from Intel’s pricing pressure.
  7. Any combination with varying degrees of items 1-6 above.

AMD can still find a way out of bankruptcy. It’s just a matter of leadership adopting a different mantra and understanding that they don’t need to continue to knock heads with Intel in order to survive. In fact it is when they started deviating from Intel’s design approach and went ahead with their own ideas that they created their best product. Maybe it’s time for AMD to do the same with their business strategy.

48 comments:

Axel said...

Good analysis. My guess is that AMD will try to implement the following as the new asset light strategy:

- Shut down Fab 30.

- Completely outsource K8 / Griffin production in 2008 to cover the low end. This would bring the already low margins on K8 even lower due to foundry fees, but at least it keeps AMD in the game for low end desktop and mobile.

- Focus Fab 36 on K10 and the 45nm shrink. This will cover high end desktop & mobile, workstation, and server offerings.

The above, or something along those lines, is probably the only way they can make it.

Anonymous said...

Interesting options:

1) Shut down Fab30 at the end of 90nm. If AMD wishes to return to profitability, this is a quick and effective move.

F30 is now FULLY DEPRECIATED! As capital cost is ~50% of the overall wafer production cost, the chips coming out of F30 (even considering they are 90nm and 200mm) are pretty decent margins. Also consider these chips continue to be AMD's TOP BIN dekstop and server parts.

Now converting this fab is a whole different story - but this fab right now is pretty much gravy for AMD...they should keep operating it for as long as they can.

Some of your other options just aren't practical - AMD couldn't go 100% fabless- this would kill any hope of competitiveness with Intel. There is no foundry capable of keeping up with Intel's scaling timelines. And AMD will lose a lot of performance benefit by not being able to tune the chip design and Si process with each other. If they are forced to simply put in a design that meets a foundires lot file (process specs), it will likely be suboptimal in terms of the max performance theoretically achievable. And you already touch upon additional margin pressure as there is now the added cost of the foundry margin.

They also can not sell ATI at this point - who would buy it? AMD would get pennies on the dollar. ATI is now crippled into an AMD only chipset solution and still has to compete with Nvidia in this space. Intel looks serious about entering the discrete card market and Nvidia seems to be firing on most of it's cylinders so discrete cards is not going to be a cakewalk either. Who in their right mind would take ATI at this point in time?

I think AMD's only chance (short term) is to go back to niche market and they HAVE to raise ASP's to get profitable. They are just going to have to swallow their pride, cede back market share to Intel and hope Intel is more interested in raising their own margins/ASP's and ending the price war.

Once they get profitable (or are on a clear road back to profitability) they can then convert F30 and /or build the NY fab. This would allow them to move chipset/graphics production in house - which would allow better margins and help smooth out capacity excesses - and then start to lower prices again in an attempt to gain share back at a helathy and sustainable rate.

There just isn't much to restructure other than cutting out some of the ATI business units.

The key for AMD to profitability is raising revenues, not cutting costs.

Ruiz and the rest of the Sr management should not take any bonuses this year and should have much of their future bonuses tied to either stock price or profitability, with a reasonable base salary. Options and RSU's should be stopped until the team can demonstrate they have a clue. They should not be able to ride on a bonus gravy train while they cut employees and their stock price tanks.

SPARKS said...

If this is true, INTC is NOT going to give AMD ANY latitude.

I said Otellini would lower the boom, and here it is. They are out for BLOOD.

http://www.theinquirer.net/default.aspx?article=41990

SPARKS said...

AND THIS:

http://www.dailytech.com/article.aspx?newsid=8656

SPARKS

Roborat, Ph.D said...

Anonymous said: F30 is now FULLY DEPRECIATED! As capital cost is ~50% of the overall wafer production cost, the chips coming out of F30 (even considering they are 90nm and 200mm) are pretty decent margins.

i said "end of 90nm". As in no more 90nm demand and stop the Fab from upgrading to anything else. The savings from this action are direct and immediate - no capex, lower fixed cost, cash for 200mm tools.

Some of your other options just aren't practical - AMD couldn't go 100% fabless- this would kill any hope of competitiveness with Intel

how is that different from today? Is AMD still competetive in its execution? In fact it looks like TSMC will be well ahead in 45nm than AMD and knows a lot more in 65nm high volume manufacturing. K10's design has been available for quite some time. AMD just can't make them properly. So don't brush off going 100% fabless just yet. At AMD's current state, they can't possible be worst off as it is.

Roborat, Ph.D said...

Cont'd...

As for ATI. pennies for a dollar is better than millions in quarterly loses with a dwindling revenue stream faster than they could lay off ATI employees. I'm sure there will be some interested in ATI - for the right price.

In any case you don't stop yourself from selling something that's costing you money just because you paid a lot for it.

SPARKS said...

All good analysis, indeed, Doc. But, it looks as if the situation has been rendered academic. Intel refuses to relent from the pressure, then there is this.

http://www.theinquirer.net/default.aspx?article=41985

SPARKS

Axel said...

Anonymous

Also consider these chips [from Fab 30] continue to be AMD's TOP BIN dekstop and server parts.

That is only true until K10 ramps. Once faster K10 parts begin to be produced in volume from Fab 36, Fab 30 becomes a dinosaur because Fab 36 produces K8 more cost effectively and AMD can switch the higher speed K8s to 65-nm. This will probably occur roughly six months from now. AMD have three options at that point: Convert Fab 30 to 300-mm/45-nm, hang onto it until they can afford the conversion, or sell it off.

I believe the latter is the most likely to occur, based on AMD's desperate cash situation. In fact, they may even sell it off sooner. There's a rumor on The Inq of a Russian buyer having purchased much of the 200-mm tooling. If true, then AMD are already giving up Fab 30 capacity today to secure the cash.

Giant said...



If this is true, INTC is NOT going to give AMD ANY latitude.

I said Otellini would lower the boom, and here it is. They are out for BLOOD.


That report is true SPARKS. If you look on Intel's reseller website under pricing they confirm the September 2nd launch for Tigerton. Launch speeds at up to 2.93Ghz.

Anonymous said...

"i said "end of 90nm". As in no more 90nm demand and stop the Fab from upgrading to anything else."

My bad - didn't catch the end of 90nm part.

"how is that different from today? Is AMD still competetive in its execution? In fact it looks like TSMC will be well ahead in 45nm than AMD and knows a lot more in 65nm high volume manufacturing"

Be careful you don't fall into the same trap that Dementia often does, not all 65nm processes (or 45nm) are the same. Just because one is competitive on schedule doesn't make the process competitive - I don't know how good TSMC's 45nm (or for that matter 65nm) process is. The interaction between CPU and process is very tricky and on a whole different scale then chipset (and probably graphic chips). Having no control over 1 of these two components would put AMD at an even greater disadvantage to Intel.

Giant said...

Barcelona performance results are in:

http://www.xtremesystems.org/forums/showthread.php?t=157136

Slower than Core in terms of IPC. Core also goes to 3Ghz, a full 1Ghz faster than Barcelona.

Anonymous said...

1. How about this miracle: griffin, Barcelona, Phenom, RD790, R6xx line, …
And Intel manipulating pricing, ho please if they lower it too much they get hurt too, hurt to the point some one will see problem on Intel and the stock goes to the toilet and Intel BK Q4/2008.

2. How shutting down a fab will help? You do know that both fabs are near each other so the employees for both are the same? Tell me how do they lose money with two fabs? Because of the electricity account of two is bigger then one? So they build the fab and now they don’t need it so let it hang in there, hoo please…

3. Sell ATI?!?! Who is ATI for you, people? buildings? “They already have the technology they need.”, And what you do with the technology if you don’t have people to work with the technology, hoo please… “Not being able to deal and cooperate with Intel with equal footing as NVIDIA is a major disadvantage”, but Ati will withdraw from Intel product line, why do they need Intel… In fact Intel needs more Nvidia and Ati then the other way around…

4. “Go 100% Fabless.”, The biggest crap I have ever heard, you do know that AMD primary business in the early days was to manufacture CPUs (better manufactured than Intel) not to design them, and now that they had made a major feet of having two fabs they go fabless? Come on, my brain is dying in here…

5. Tell me do you know what CPUs have been manufactured in Chartered? Which models and clock speed? When you have that answer, put it here.

6. “Focus on niche.”, That would kill then for sure, niche = no products no volume, the end for sure. Better focus on volume don’t you think. “This requires a massive u-turn in its philosophy of breaking the monopoly”, so that’s what you want a Monopoly, nice…, by the way some Hitler dictator would be nice too, no? ”will protect AMD from Intel’s pricing pressure.”, Like VIA right, I see where you want AMD.

7. Any combination with varying degrees of items 1-6 above would kill AMD and that’s the all point isn’t it?

Giant said...

1. How about this miracle: griffin, Barcelona, Phenom, RD790, R6xx line, …

What miracle? Griffin is just a rehash of K8 Turion with larger cache and some power saving enchancements.

Barcelona and Phenom results are in (see my post above), prefragged by Kentsfield and Clovertown.

RD790 is just a chipset. There's nothing revolutionary about it. About the only major change is PCIE 2.0 support, but Intel and Nvidia chipsets that are upcoming have this featureas well.

R6xx is quite underwhelming. Ati did not raise the bar, not one bit, with these new products. That's what I call a flop, the likes of which haven't been seen since the Geforce FX.

Roborat, Ph.D said...

1. How about this miracle: griffin, Barcelona, Phenom, RD790, R6xx line, …
And Intel manipulating pricing, ho please if they lower it too much they get hurt too, hurt to the point some one will see problem on Intel and the stock goes to the toilet and Intel BK Q4/2008.


the question is, how much is too much in intel's case? It seems to me that a few % points of its margins drives AMD to the ground.

2. How shutting down a fab will help? You do know that both fabs are near each other so the employees for both are the same?
Tell me how do they lose money with two fabs? Because of the electricity account of two is bigger then one?


if you do not understand the cost to run ultra-cleanroom environments then i see no point trying to explain it to you.
So they build the fab and now they don’t need it so let it hang in there, hoo please…

as opposed to continue running it at a loss?

4. “Go 100% Fabless.”, The biggest crap I have ever heard, you do know that AMD primary business in the early days was to manufacture CPUs (better manufactured than Intel) not to design them, and now that they had made a major feet of having two fabs they go fabless? Come on, my brain is dying in here…
who cares about what AMD's history is or its primary business. As a business, its supposed to make money. AMD's "feat" of building 2 fabs is nothing now but a gigantic fixed cost that they can't afford.

5. Tell me do you know what CPUs have been manufactured in Chartered? Which models and clock speed? When you have that answer, put it here.

do you have a point or are you just trying to brag?

6. “Focus on niche.”, That would kill then for sure, niche = no products no volume, the end for sure. Better focus on volume don’t you think.

you mean like they're doing now? how is that going for AMD?

“This requires a massive u-turn in its philosophy of breaking the monopoly”, so that’s what you want a Monopoly, nice…, by the way some Hitler dictator would be nice too, no?
huh?

”will protect AMD from Intel’s pricing pressure.”, Like VIA right, I see where you want AMD.

huh? does VIA have products like Opteron MPs or 4x4?

7. Any combination with varying degrees of items 1-6 above would kill AMD and that’s the all point isn’t it?

as opposed to standing still and continue to lose $2B annually? it looks like you're the one who wants AMD to act like a startled deer in the middle of the road

SPARKS said...

"1. How about this miracle: griffin, Barcelona, Phenom, RD790, R6xx line, … "

I believe you have misunderstood the DOC's intentions. Given AMD's current finacial situation, AMD MUST do something radical to stop hemoraging expendatures. Yes, these ideas seam very radical, however, desperate times call for deperate measures.

Doc's analysis and subsiquent scenarios are ways for AMD to save the company, not ideas to cripple AMD any further. The bottom line here is, they need to generate extremely successful products and/or cut CAPEX, maybe both.

Therefore, rather than criticize the DOC's possiblities, perhaps YOU could offer alternative senarios conserning AMD's survival in realistic terms; as apposed to the conjecture and misinformation AMD offers on a daily basis.

The fact is Nvidia and Intel ARE on a serious roll and AMD has cut their margins to severe losses. There are serious delays in product yields and avalability, and product launches are underwhelming, performance wise. They are not making a profit because their products are not competitive.

These facts are undisputable, and AMD must do something. Do you have any ideas?

SPARKS

Anonymous said...

Roborat, Ph.D,

I have pity for you; you are not much different from Scientia or maybe Sharikou too, except you don’t delete the posts.

You didn’t understand that I was making an analogy to
company monopoly = dictator company.

The company who as a monopoly will do what it wants whatever it’s good or bad to the market (end users if you prefer), and that is technology/price/performance/features/… you name it!

I will not reply to the other non replies you have give me, just let you with this from Anandtech site, maybe this will bring some light to your head:

It bears mentioning that there exists some potential issues between now and then however. First and foremost is the 800lb gorilla: Intel. Under normal conditions Intel's support is critical for new extensions to take off, the only exception to that has been AMD64/x86-64, which came at a time when that specification was far better suited for the computing industry than Intel's own IA64 specification. Simply put, we believe AMD will be unsuccessful with SSE5 if Intel is unwilling to add it to their own processors; the performance improvements are important, but it's not an AMD64 situation where AMD has the influence and technical advantages to get what it wants.

But with those warnings out of the way, don't take this as a disapproval of SSE5. We believe that this will be the most important extension to the x86 instruction set since SSE2, and the new instructions like MADD in particular can offer the kind of performance improvements AMD wants to hit while avoiding the need to increase clock speeds significantly and the problems that result from such.

And Sparks, Maybe just waiting for what AMD would do until the end of the year?
Since AMD will release complete new products for all the market segments? Then maybe write this article if something is not too rosy?
Giant for example said RD790 is just a chipset nothing special, so what to say from all the Intel 3x line, complete junk? Complete waste? What’s interesting there too? DDR3? Like if it was the most need tech in Core 2 CPU, memory bandwidth. Bah.

Roborat, Ph.D said...

thanks for the link giant.
http://www.xtremesystems.org/forums/showthread.php?t=157136

interesting development. anyone interested in writing an article to summarise the chatter? i can't post it under their name.

email: roborat2000@yahoo.com

SPARKS said...

"And Sparks, Maybe just waiting for what AMD would do until the end of the year?"

I need not post all the previous posts going back 1 year this date. I will say, however, that I saw this entire chain of events happening to AMD at the time. This purchase of ATI, for the price they paid, was just plain wrong. Let me give you an example of what I posted on thr Inq.....

"If you didn’t see this coming with the ATI/AMD buyout you are as naïve and short sighted as Hector Ruiz. This unholy alliance would lead credence that an all out assault on DAAMIT has been brewing. Business wise, it costs both INTC and NVDA nothing to obtain the monopoly everyone talks about drop in their respective laps by default. Fact: Those stupid rumors your Charlie (correctly) wrote about last year about INTC buying NVDA were, rightly, preposterous. At this juncture, however, INTC doesn't need to spend billions, and NVDA can hold on to its nuts. They can crush their competition together; corner the market by doing what they do best, individually. Talk about being between a rock and a hard place. The bottom line? It cost AMD 5 billion to put themselves in this position.

No mater how you cut this pie, it is indeed a dangerous scenario for the struggling AMD/ATI fighting a two front war against two very large and powerful adversaries with different agendas----presently. Truth is, after all, there is blood in the water, and both NVDA and INTC are on a serious roll. More to the point, as much as you can feel AMD’s pain, not only will this happen, it makes perfect business sense."

In fact, many waited, so was Cray, so was the channel, so was SUN, so were loyal AMD fans. I did't and I won't.

Wall Street, bankers and creditors, can only wait so long. Intel and Nvidia are not sitting on the side lines, sipping Latte's, waiting for AMD to catch up, either. Trust me. Does this by your estimation make them bad people, Evil Empires, Nvidiots? No they are smart companies who made the right choices at the right time. AMD, unfortunately didn't.

Look their are no "bad chipsets". Some people at AMD/ATI gave their proudest best to bring out a good product, and worked hard to do it. I feel for these people. This is REAL competition, however, the American way. AMD is getting the fair fight they asked for and PAID FOR, 5 BILLION TIMES.

I'm an American investor. I saw this coming and I made good money on both NVDA, and INTC. Does this make me bad person? No, I am a capitalist pig. Waiting till the end of the year? Trust me, if I saw anything of value coming from AMD/ATI, I'd sell INTC and NVDA at tomorrows opening bell in goddamned heartbeat. I wouldn't wait a minute more.

As for now, I just don't see it happening. NVDA and INTC are just to strong. There are some smart people on this site, industry insiders, they KNOW what they're are talking about and I'm making money. If I wait, I lose. I wouldn't sell my wife, kids, family, or friends. Tech companies are very, very negotiable. Word.

SPARKS

Anonymous said...

Barcelona amazing! Nope
DTX amazing! Nope
Scientia's newest hope of crushing Intel? "SSE5" LOL

Also, it seems that Scientia is taking a spanking over at the Zone. He's already owned up to his faulty crystal ball, next he'll hopefully be gone from the web :D

Anonymous said...

Barcelona amazing! Nope

Link ?

pointer said...

someone said ... 4. “Go 100% Fabless.”, The biggest crap I have ever heard, you do know that AMD primary business in the early days was to manufacture CPUs (better manufactured than Intel) not to design them,

since the bolded part goes unchallenged, I'll put up my statement here:
No, AMD didn't better manufactured CPU; and in facy, it used to have all sort of yield issue in the old days. My old data point (7-8 years ago) is that AMD had 60-70% yield at the test end.

What exactly AMD did last time when it did the Intel clone, was speeding up some MHz by resolving some Intel CPU speed path

Anonymous said...

I don't think that Intel will hurt itself in order to wipe out AMD. Intel can't simply push pricing or products with the sole intent of forcing AMD out of business, they have stockholders as well.

At present, Intel has a pretty balanced strategy where they continue to improve their product and release it on a schedule that allows them to maintain an acceptable level of profitability without giving AMD much of a break.

Besides, AMD's biggest enemy has traditionally been AMD, not Intel. Intel really just needs to maintain its own course and watch as AMD struggles because it's so badly managed. Granted, they've always had a smaller margin of error, but they are so badly run that it makes you flinch.

Anonymous said...

"Intel can't simply push pricing or products with the sole intent of forcing AMD out of business, they have stockholders as well"

Intel stock has gone up ~25% this year as they have put the pressure on AMD. Could they have had a greater return? Probably... but a 25% return is doing OK for your stockholders (especially after many years of near 0 returns with the exception of the dividends)

It is AMD who has forgotten (and punished) its stockholders. This market share at all cost is great for the Napolean running the company but what exactly is that strategy doing for AMD stockholders?

As for the ATI purchase - well hindsight shows it may not have been a great deal (at least at this point in time) and some may have predicted this. I do respect AMD taking a shot at this, granted it looks like it may have been the wrong call, but they took a shot. This was a far better strategic decision than the slash and burn market share strategy. At least with ATI, AMD rounds out their product line, has some minor hope of competing in corporate space with a platform and has a chance of cutting out some new markets with things like fusion.

So will any of these work? Maybe not, but I respect the fact that they took a chance. Clearly Intel was heading in a similar direction (Larabee development was/is the worst kept secret, and while AMD has hyped fusion it is not clear whether Intel had a program already in place to look at this or not - they were just less public about it).

As for the SSE - I think for the good of the industry, it needs to be done via consortia. This gold mine mentality (if I announce first I can set it up my way and stake my claim) is really not good for the industry and SW is already falling far enough behind where I really don't think we want them deciding whether to do development for two sets of instructions or have them try to decide which one will be better.

SSE5 may be good - but it looks more like AMD saying we got caught with our pants down on SSE4 (and were only able to partially implement on K10) and therefore we will get head of the curve.

SPARKS said...

Here we go. The heads are rolling or the rats are bailing, your choice:

http://www.hexus.net/content/item.php?item=9747

Anonymous said...

Roborat and Sparks here it is what I was talking about. You don’t need to castrate the company, just need to wait a while:

AMD Ships Quad-Core Opteron Processors for Revenue

-The $$$ money loss start shrinking today.
-The OEM are also waiting for the superb chipsets (RD790).
-The 2600XT is a success, almost all the manufactured got sold immediately.
-OEMs love the Turion + Nvidia/Ati chipsets (cheaper notebook/more profits) no need for Santa Rosa costly, useless and unneeded expensive tech.
Santa Rosa is one of latest examples that the monopoly is a bad think that forces stupid “standards” that hardly anyone needed or wanted (OEM/Users/People).

SPARKS said...

Nice picture. By the way I am extremely flattered the put me in the same sentence with the, Doc. That said, I will be willing to bet from a manufacturing, yield, performance and availability perspective the he will have something more to say about this, yet another, paper launch. So will others.

Financially, this thing, along with some benchmarks, by independents, better compare with the high end C2Q’s, at least. Pricing; where do you begin? Frankly, financially speaking, I don’t think it will be enough to stem the tide of tsunami losses for the 3rd and 4th quarters. Further, it’s about time they committed to something. Don’t you think?

Finally, third quarter results are coming in Sept with the real launch and ALL the numbers including benchmarks will be out. Still think it looks good? By all means, buy a couple of hundred shares.

Good Luck.


SPARKS

Anonymous said...

-The $$$ money loss start shrinking today.

CPUs running at or below 2Ghz in extremely limited quantities are not going to stop AMD losing a truckload of money each quarter. Over the last three quarters AMD has lot more than $1.7bn. After this quarter that loss figure will be far higher than $2bn.

-The OEM are also waiting for the superb chipsets (RD790).

No matter how good the chipset is they're still stuck with old K8 CPUs. The only reason AMD is moving any real quantity of these CPUs is due to the pathetically low prices they're selling them for.

-The 2600XT is a success, almost all the manufactured got sold immediately.

They're selling this product, but hardly making any profit on it. They're being forced to sell them at $100 because of the mediocre performance vs. Nvidia's parts.

OEMs love the Turion + Nvidia/Ati chipsets (cheaper notebook/more profits) no need for Santa Rosa costly, useless and unneeded expensive tech.

This is FUD. It's common knowledge that Core 2 Duo is much faster than Turion. 20% faster clock-for-clock and it has higher frequencies. There's no need for outdated, slow, Turion mobile CPUs.

Giant said...

Note: I posted this on Scientia's "non biased" blog. Since he has resorted to deleting comments once more I'm also posting it here.

gutterrat, Scientia also censored my posts as well.

Scientia claims to be open and non-biased, but has he ever deleted a post made by one of the pro-AMD people here? No. Never.

In fact, this post and your post above mine will probably soon be deleted.

It's pathetic how AMD fanboys like abinsein claim crap like "I know people who are testing K10 and Penryn and K10 is faster." without providing even the slightest modicum of proof.

Sharikou might be a bleating AMD fanboy, but at least he doesn't censor posts like Scientia does.

Giant said...

These scores come from XtremeSystems:

They are a comparison at 2Ghz between Barcelona and Harpertown (both unreleased products):

If you want all the pretty pictures etc. as before the link is: http://www.xtremesystems.org/forums/showthread.php?t=157136

Harpertown results on the left, Barcelona on the right.

23.3s vs 39.7s for Pi
20.48 vs 17.13 in "relative speed" for Fritz
2454 vs 1896 in single cpu cinebench
15334 vs 13295 in mutli cpu cinebench
6.25x vs 7.01 cinebench scaling (78 vs 88 percent)
10.4s vs 10.6s for wprime 32m
310.9s vs 327.4s for wprime 1024m

As it is plainly clear, Barceloa is behind Harpertown on IPC. The only test Barcelona is better in is the cinebench scaling.

But, what's also important to note, is that this scaling advantage advantage will be decreased with the Harpertown CPUs that feature a 1.6Ghz bus, up from 1.33Ghz today. It's not much, but should be good for a few %.

Anonymous said...

"Frankly, financially speaking, I don’t think it will be enough to stem the tide of tsunami losses for the 3rd and 4th quarters."

AMD's profitability in Q3 and Q4 will solely be dependent on how much the overall x86 market grows. If it grows and AMD can loosely retain market share at current level they will cut losses some.

If the market doesn't grow (which would be unusual for Q3 and Q4 which are generally seasonally stronger) or if Intel cuts back into market share then losses will continue as is. Either way K10 will have next to ZERO impact.

To put things in perspective servers account for ~5% of AMD's unit shipments. If they manage to go from 0-50% by end of year on K10 (which is pushing it), K10 would still only account for ~1-2% of AMD's unit shipment in H2'07. Throw in another 1-2% for desktop and you'll see why even Hector was claiming limited financial impact months ago before the delays even occurred.

It's frustrating listen to all these AMD fanboys spouting nonsense about finances and costs (like 65nm) when they don;t have a clue other than 65nm EQUAL 2X 90nm, GRUNT, GRUNT - SAVINGS MUST BE HUGE?

Oh you say aggregate die size will not actually decrease that much as AMD is introducing new products with more cache and 4 cores? Oh you're saying the extra metal layers and process technologies on 65nm makes the wafer cost more? Wait you're telling AMD is only getting ~30% smaller for an equivalent die and not a 50% shrink? Still the savings will BE HUGE I TELL YOU!

Anonymous said...

Nice picture. By the way I am extremely flattered the put me in the same sentence with the, Doc.

Since you reply to questions directed to him on the same subject, what would you expect? ;)



CPUs running at or below 2Ghz in extremely limited quantities are not going to stop AMD losing a truckload of money each quarter. Over the last three quarters AMD has lot more than $1.7bn. After this quarter that loss figure will be far higher than $2bn.

But I bet isn’t another 600 million.


No matter how good the chipset is they're still stuck with old K8 CPUs. The only reason AMD is moving any real quantity of these CPUs is due to the pathetically low prices they're selling them for.

So what? It’s better to sell some AMD chipset with some AMD CPU than only NVIDIA chipsets with AMD motherboards don’t you think?


They're selling this product, but hardly making any profit on it. They're being forced to sell them at $100 because of the mediocre performance vs. Nvidia's parts.

Don’t really think so. If that would be the case you would see larger prices differences between the 2400 and 2600 lines. The 2600XT is normally faster than 8600GT and the 8600GT cost more…


This is FUD. It's common knowledge that Core 2 Duo is much faster than Turion. 20% faster clock-for-clock and it has higher frequencies. There's no need for outdated, slow, Turion mobile CPUs.

Not really FUD… Seek for press releases and you will see HP, DELL, ACER are making new models announcements. The Turion could be 50% slower than a core 2 duo, most people who buys notebook don’t really care about its performance more the price and/or features. If you worked on a computer store you should know that. And the outdated Turion mobile CPU was the first to the market with 64 bits. US market is already at 44% AMD 56% Intel not bad for an outdated CPU no?


Besides read this and you would all see why we don’t need Core 2 Duo X6800 processors:

Current Analysis West said that the best-selling AMD desktop PC in July was HP's Pavilion a6110n with an Athlon 64 X2 4400+ processor, a 320 GB hard drive and 2 GB of main memory for $545; Intel's best-seller was the Pavilion a6120n with a Core 2 Duo E4400 CPU, a 320 hard drive and 2 GB system memory for $610.

Surprisingly, AMD-based notebooks also showed an unusually strong presence during the month of July. 44.8% of all U.S. retail notebooks integrated an AMD processor – a record high for the company. Unit share was up 11.8 points from 33.0% in June and up 20.5 points from 24.3% in January of this year. Intel Centrino notebooks, on the other hand, were estimated at a retail unit share of 55.2% in July - down from 67.0% in June and more than 75% in January.

Equally interesting is the fact that AMD-based notebooks are not much cheaper than Intel-based systems. According to Current Analysis West , AMD notebooks sold for an average of $748, while Intel systems were priced at $764.

SPARKS said...

"But I bet isn’t another 600 million."

No you won't. Not a dime.

You don't have the balls, or the money, to risk anything.

Your right about one thing. It will not be 600 million. AMD is selling tooling, parts of FABS, their own backyard, and god knows what the hell else, just to stay alive, and of course, to look good while their bleeding to death.

SPARK

SPARKS

SPARKS said...

Oh yeah, read this.

http://www.motherboards.org/reviews/hardware/1721_1.html


SPARKS

Anonymous said...

Brent Rehmel aka biggest AMD apologist/propagandist aka Scientia used to say that AMD's quotes of 20% better for INT and 50% for FP were not K10, but quad K8 scores, obviously, only a conspiracy that a widowed nut could concoct. Recently, AMD has demonstrated SPEC scores in line with previous projections.

Moral of the story? Scientia, own up to your 5464069406th broken crystal ball!

Anonymous said...

Just wait for the Anandtech K10 3.0Ghz demo article and we will talk if Scientia was right or not...

Anonymous said...

What does the 3GHz Phenom have to do with another conspiracy by Scientia being busted???

Speaking of that 3GHz Phenom, damn! Look at that cooler.

Anonymous said...

opps wrongs link

here it is
gigantic cooler needed by 3ghz phenom

SPARKS said...

Here ya go.

http://www.xbitlabs.com/news/other/display
/20070902004215.html

SPARKS

Ho Ho said...

"It’s better to sell some AMD chipset with some AMD CPU than only NVIDIA chipsets with AMD motherboards don’t you think?"

If would have been even better if they could have sold the chipsets to AMD and Intel CPUs. You know, the other company that sells roughly 80% of all the CPUs.

Also, have you got any idea how much marketshare has ATI lost since it was aquired by AMD?

Anonymous said...

In addition, they lost ~$100m/quarter when Intel canned the chipset deal they had with Ati. They would still have this revenue stream if it weren't for selling out to AMD.

SPARKS said...

WTF?!?!

"Next week expect more from HEXUS on AMD ‘Barcelona’ and how its own customers are confused and frustrated over AMD’s extremely restricted media product sampling ‘strategy’ of this new product... which, it seems they believe, could cost them a packet at the expense of AMD endeavouring to tactically influence a, um, initial ‘proper understanding’ of the chip and its platform proposition by the hands-on sampling of the most appropriately friendly of ‘media partners’... "


http://www.hexus.net/content/item.php?item=9747

AMD is gonna TANK. GET OUT while you still have your nuts.

SPARK

Anonymous said...

scientia has a small brain

Anonymous said...

niarb llams a sah aitneics

Anonymous said...

Lighter side... found this at Yahoo board:

> Fanboi Charlie's (aka "Groo") only response on "K10 benchmarks start to trickle" thread....

> http://aceshardware.freeforums.org/viewtopic.php?t=96

> "Know."


> LOL!!!!!! By the way his somewhat rather extremely short (one word) response, he "knows" AMD is over.... precious!!!

Anonymous said...

AMD Asset Light - Job Wanted

Look at this!

Job Title: Director, CPG Product Cost Reduction
Company: AMD
Location: Austin, TX
Job Description

BSEE and 10+ years of experience in a related field.

Product and Test Engineering experience required. Knowledge of product cost drivers and proven track record reducing semiconductor production costs in volume manufacturing.

Experience working with all facets of semiconductor manufacturing—die production, assembly, and test. Ability to work in and drive cross functional teams in both direct
supervisory and matrixed models.

Experience with both internal manufacturing and foundry/OSAT.

Primary Purpose & Key Job Functions

Drive cost reduction initiative for high performance client microprocessors.

Cost reduction and form factor reduction will enable these microprocessors to be applied to emerging and non-traditional applications spaces.

Short term cost reduction opportunities will include existing die designs targeted toward smaller, less expensive packages and reduced test flows. Long term opportunities include die reduction through feature removal and die shrinks.

Annual cost reduction goals (unit cost) will be planned and budgeted. Execution on cost reductions may utilize internal wafer fabrication, assembly and test sites as well as foundry plus outside assembly and test.

At AMD, we are committed to equal employment opportunity. AMD does not accept resumes from headhunters, placement agencies or other suppliers that have not signed a formal agreement with us. Our supplier base is restricted to specified hiring needs. Therefore, any resume received from an unapproved supplier will be considered unsolicited, and AMD will not be obligated to pay a referral or placement fee.

http://messages.finance.yahoo.com/Stocks_%28A_to_Z%29/Stocks_A/threadview?m=ts&bn=893&tid=1599731&mid=1599731&tof=1&frt=2

SPARKS said...

It begins, the performance per watt scenario/mantra. As I said before, damage control.

SPARKS

SPARKS said...

OOPs forgot this part.

http://www.vnunet.com/tv/?channel=vnunet.com%20TV&clipid=1386_vnunet_0096

SPARKS said...

Giant,

Very tasty!

http://www.vr-zone.com/articles/Asus_P5E3_Deluxe_X38_Mainboard_Preview%21/5217.html

SPARKS