For AMD Success Means Problems

An anonymous reader writes "AMD's success with its dual-core Opteron and Athlon processors has created something of a happy problem for the company. It can't make its products fast enough to meet demand. Just the same, with the Intel price war heating up and new 65-nanometer manufacturing technology being implemented in its factories, AMD has a lot of balls in the air right now." From the News.com article: "AMD's current pickle is the result of its success, which makes it a little easier to swallow for company executives. Demand is high, but the company's dual-core processors still use its 90-nanometer manufacturing technology. Intel's chips, on the other hand, are built using the smaller transistors provided by its 65-nanometer manufacturing technology. Not only is AMD using larger transistors, but its dual-core Opteron and Athlon 64 processors contain two processing cores integrated onto a single piece of silicon, or a die. This design has given AMD great performance during the past few years, but resulted in processors that were almost twice the size of its single-core chips."

Comments

I hate it when...

[MightyYar]

I hate it when my balls are in the air.

Re:I hate it when...

[smittyoneeach]

What good are balls in the air, when they've insufficient balls to release the specs and/or the driver source?

Re:I hate it when...

[Anonymous Coward]

I was hoping the first comment would be on that line.

Re:I hate it when...

[nschubach]

Both quotes are laced with falic symbology:
"AMD has a lot of balls in the air right now."

"AMD's current pickle[...]"

"[...]makes it a little easier to swallow[...]"

Re:I hate it when...

[TheLinuxSRC]

Obligatory Beavis and Butthead quote:

"hehe... he said balls"

Re:I hate it when...

[Anonymous Coward]

It's a fairly common expression in Sweden.

Re:I hate it when...

[drinkypoo]

"Balls in the air" was chosen because everyone has seen juggling; it supersedes "irons in the fire" because most people think that hamburgers grow on trees.

Re:I hate it when...

[Hoi Polloi]

Yes, it is a reference to juggling. It means having more things going on then you can handle.

Re:I hate it when...

[MightyYar]

Hålla många bollar i luften.

Re:I hate it when...

[Anonymous Coward]

"I hate it when my balls are in the air."

Better there than a rock and a hard-place.

Re:I hate it when...

[Firehed]

[insert phrase of choice regarding porn and only having two hands here]

Re:I hate it when...

[livewire98801]

Does that make me a vegitarian?

Re:I hate it when...

[Afrosheen]

Ugh. I'll spare a very long and insightful post for the hundredth time on this subject by saying simply,

  "STFU. Licensing issues. Buy ATI."

Re:I hate it when...

[smittyoneeach]

Aye, and after we've hit a 1,000 times or so, we can effectively convince these vendors that a market exists, and we do care.
Understood: licensing issues.
Understand: market demand.
If ATI produces more enlightened products, your market goes by way of Soviet Russia.

Re:I hate it when...

[Schemat1c]

...AMD has a lot of balls in the air right now.

That reminds me...

It's like...

My....

Nah, too easy.

Re:I hate it when...

[Anonymous Coward]

There is no such thing as an obligatory quote, ass picker.

Re:I hate it when...

[zacronos]

"Balls in the air" was chosen because everyone has seen juggling; it supersedes "irons in the fire" because most people think that hamburgers grow on trees.

Am I the only one who has no idea what that hamburger reference means?

"Too many irons in the fire" refers to blacksmithing, and has nothing to do with hamburgers (unless I'm really missing something). The idea was that a blacksmith would pipeline his work by putting more than one iron in the fire at once, putting another in the fire as he pulled one out to work. Having a lot of irons in the fire would mean he had a lot of tasks prepped and ready to be worked on -- but without the implication that they needed his immediate attention.

In contrast, having lots of balls in the air would mean you have lots of things going on which all require some degree of attention; having too many balls in the air can easily put you in a situation where you've taken on more than you can handle, while having too many irons in the fire really isn't a problem because they're just sitting there waiting for when you're ready. Thus, "balls in the air" is a more appropriate expression.

Re:I hate it when...

[Anonymous Coward]

I got too many balls in the fire. Hell, even one is too many.

heh

[Anonymous Coward]

balls in the air.

Does size matter?

[Riley Holmes]

As long as the processor fits inside of the PC case, I don't see why the bulkier size matters. If the performance is superior, it just doesn't make a difference.

Re:Does size matter?

[Anonymous Coward]

Size is heat, my friend, size is heat.

Re:Does size matter?

[OwnedByTwoCats]

For AMD, a bigger die per chip means fewer chips per wafer.

Which is a problem, when you can't manufacture enough to meet demand.

Re:Does size matter?

[man_of_mr_e]

You don't really understand how they make chips, do you?

It costs $X to make a wafer of CPU's. The more CPU's you get from that wafer, the cheaper each CPU costs. Large CPU Dies means fewer CPU's per wafer, thus high cost per wafer. Thus, each CPU die has a higher cost to manufacture than smaller dies.

Re:Does size matter?

[Aadain2001]

1. Larger size = fewer chips per wafer

2. Larger size = higher average chance of defect per die

3. Larger size = more expensive to manufacture per processor

So you see, size DOES matter :)

Re:Does size matter?

[Anonymous Coward]

It is also worse than that.

The larger the die the lower the yield.

Taken to the extremes, assume a huge die. A single defect and you lose the entire wafer.

On a super tiny 100 dies/wafer a single defect only takes out 1%.

Of course on large dies, you might have small pieces of redundancy built in (all memory has this) so you can recover from some number of defects before you have to scrap it.

But still, the larger the die, the fewer you can make at a time and the lower the yield on the ones you do churn out.

Re:Does size matter?

[mnmn]

So right now building a new fab is at its most profitable?

How can we run out of silicon? Its in the sand. How come China isnt coming out with mountains of pure silicon for sale if the demand has been this good for this many years?

Re:Does size matter?

[Bender0x7D1]

It means more than that...

First, you are correct. If you get more chips per wafer, you can make more chips. Since the time to process 1 wafer remains consant. However, there is also more going on.

The second thing to worry about is the cost. If it costs (making up numbers) $100 to process a wafer and you get 10 chips, it is $10 to manufacture 1 chip. If you get 20 chips from the wafer, then it only costs $5 to manufacture 1 chip.

The third item is quality control. If there are any flaws in the wafer, the chip that is created over that flaw can not be used. So that chip gets thrown out. If we can get more chips from a single wafer, our percentage yield increases as well. Imagine that there is 1 flaw per wafer. If we only get 1 chip per wafer our actual yield is 0%. This would be very bad. Now imagine we get 2 chips per wafer. While it is possible that the flaw would affect both chips, most likely it will only affect 1, giving a yield of 50%. If you get 3 chips, your yield is 66%. This yield really hits the bottom line.

If you are losing 2-3 chips per wafer from flaws, then any increase in the number of chips is going to increase your yield percentage. If you call it a 25% increase in chips on a wafer, due to the 65nm instead of 90nm process, the percentage of chips lost to flaws will also go down and you just made more money.

Re:Does size matter?

[Belial6]

Actually, it's not a problem at all. It's a good thing. If AMD were already producing 45nm chips, and they were twice the size and slower than Intel's solution, THAT would be a problem. When you are doing well enough that you are outselling your ability to produce, and you still have not yet implemented your already developed technology, you are in a very good position.

Re:Does size matter?

[HermMunster]

But you know you are a sound company and you buy up all that cheap stock and make huge profits once the manufacturing issues are worked out.

Re:Does size matter?

[SillyNickName4me]

Yes, a larger chip definitely has disadvantages, but the doubling of size is unavoidable when doubling what is in there. WHat did they expect, 2 cores taking as little space as a single core?

Re:Does size matter?

[Anonymous Coward]

"Does size matter?"

Must...resist...obvious...jokes...

Motherboard makers may end up having to reduce the footprint of big cpus by turning sockets on their sides or lifting them above the board using mushroom shaped pillars. You'd have to be careful about putting heavy coolers on them.

Re:Does size matter?

[Anonymous Coward]

This really begs the question. I'm sure there's a reason somewhere but...
Why DO they use round wafers to create the chips from. Certainly there would be less waste with square wafers (though less and less so as the chips get smaller and smaller, of course). Are there other savings in the manufacturing process that make round wafers more cost efficient in the grand scheme of things than square ones?

Re:Does size matter?

[JimTheCactus]

Actually, for manufacturers, size matters in a big way.

Semiconductor manufacturing is, like with most manufacturing, an imperfect process. In general, for a given die (a die is one "chip" before it's placed in/on any kind of case) and a given manufacturing method there is going to be a manufacturing error rate that is measured in terms of errors per unit area.

These defects can stem from everything from a speck of dust getting into the system, all the way to a gas depositing process making a trace too thin (so the trace, and thus the chip, burns up during use.)

Therefore the bigger the die is the higher the likelihood that any one die will contain one of these manufacturing defects. Since a defect can occur at any point on the wafer, the less of the wafer that can be associated with any particular common point of failure (i.e. the smaller the die,) the more valuable any particular wafer can be and thus increase yield and profit of that product.

Re:Does size matter?

[elrond1999]

Although moving to 65nm probably would do AMD a lot of good, there is no point in going there if AMD hasnt worked out all the flaws in their 65nm process. 65nm is probably means more cost per wafer, longer design time and worse yield.

Re:Does size matter?

[onx]

Except that in this case your so-called "already developed technology" isn't really devloped yet (at least not by AMD, 45nm & 65nm tech remains to be exclusively Intel secret sauce).

Ed over at overclockers.com often talks about this subject http://www.overclockers.com/tips01051/

Re:Does size matter?

[morgan_greywolf]

Actually, it's not a problem at all. It's a good thing. If AMD were already producing 45nm chips, and they were twice the size and slower than Intel's solution, THAT would be a problem. When you are doing well enough that you are outselling your ability to produce, and you still have not yet implemented your already developed technology, you are in a very good position.

Oooh! Can you say that again?

This is what "Intel already has 45nm technology" whiners are missing entirely. AMD's sales are outstripping production and they've already got the 65 nm technology waiting in the wings in order to keep up with demand.

Die size doesn't matter a whole lot to the end user of the product unless we're talking laptops. Then it matters. But when we're talking desktops or servers (as is mostly the case with dual core 64-bit chips today -- few laptops are using 64-bit chips), then the end-user doesn't care so long as the case it's sitting in can keep it cool.

But it does matter to the manufacturer...and so what if Intel has 45 nm technology? As long as AMD can produce a better, faster chip and keep up with demand, they're sitting pretty. It's that simple.

Re:Does size matter?

[opti6600]

It's from the way the silicon ingots are drawn and cooled. Slicing it is a lot simpler than wasting a ton of possible product space making it square for a particularly OCD manager.

Re:Does size matter?

[smbarbour]

With myself playing the devil's advocate as I am frequently wont to do...

While the larger size does increase the chances of a defect for a die, the chance of a defect for a wafer remains the same.

However, with the smaller size, that same defect is also more likely to be a catastrophic defect. (The difference between having working chip and a non-working chip)

Re:Does size matter?

[TheRaven64]

I was talking to a former Intel senior engineer a while back, and he made the point that a hexagonal die would be a better choice. It would have a larger area for external connections and you could fit more on a circular wafer than with a square die.

Unfortunately, a hexagonal die is a lot harder to design, and harder to cut out of the final wafer, than a square one. The second problem is probably solvable relatively cheaply. The first is not.

Re:Does size matter?

[TheRaven64]

On a super tiny 100 dies/wafer a single defect only takes out 1%.

Assuming the defect is in the middle of a die. Assuming random placement of the defect and a square die you lose 1-4%. Depending on the ratio of die to defect sizes, you may lose 2-3% on average.

Re:Does size matter?

[TheRaven64]

This is true. AMD actually do quite well out of defects. A fully-functional die is an 800-series Opteron. It has loads of cache and four working HyperTransport links. If some of the cache doesn't work (most likely, since most of the die of any modern processor is cache) then it becomes a cheaper model. If some of the HT links aren't working, then it is downgraded to a 400, 200 or 100 series. In the worst case, it becomes an Athlon 64, or even a Semperon. If an entire core is too defective to use then it becomes a single-core part.

You can work out AMD's yields by how they price their parts. The ones near the pricing sweet spots are likely to be the ones they get the biggest yields of.

Re:Does size matter?

[Aadain2001]

This is true of all uP manufacturers. Most build in redundancy into their designs to help with this. For example, a processor with 4MB of L2 cache has physically more cache than 4MB. There may be an extra couple of rows of transistors for a block of memory so if a row or two is corrupted the block can still be functional.

Re:Does size matter?

[Bing Tsher E]

As long as AMD can produce a better, faster chip

I think you meant "a hotter chip that costs more to produce."

And as far as volume, not being able to ship to potential customers is a disaster. It means your competitor's market share grows. If AMD and Intel were producing identical commodity products, it wouldn't matter as much.

Re:Does size matter?

[Bing Tsher E]

A fully-functional die is an 800-series Opteron. It has loads of cache and four working HyperTransport links. If some of the cache doesn't work (most likely, since most of the die of any modern processor is cache) then it becomes a cheaper model. If some of the HT links aren't working, then it is downgraded to a 400, 200 or 100 series. In the worst case, it becomes an Athlon 64, or even a Semperon. If an entire core is too defective to use then it becomes a single-core part.

[putting on clownish AMD fanboy hat just for fun]

"At what point does the die become defective enough that they ship it to Intel to be a Pentium IV?"

[rim-shot]

Re:Does size matter?

[servognome]

When you are doing well enough that you are outselling your ability to produce, and you still have not yet implemented your already developed technology, you are in a very good position.

If you can't produce enough that means you can't capture additional marketshare, you alienate customers, you give the competition time to create a competing product to retain marketshare, and your manufacturing costs remain higher so your chips aren't as profitable.

Re:Does size matter?

[afidel]

Actually I very much care how much power my CPU's draw. I only have 50KVA available in my current UPS. I'm already at ~66% capacity. For me every watt counts because wasting 100 watts per dual CPU server means I can get that many fewer servers in before I hit the wall with my current UPS. A new larger UPS and generator big enough to power it is on the order of a quarter million dollars, so spending more per server to get cooler, equally capable processors isn't a big deal. We've been using Opterons because they have been giving us the best MIPS/Watt but that has changed with the new Core based Xeon's, only unfamiliarity with the models and lack of standard images is really keeping us from moving over right now (and standardization).

Re:Does size matter?

[Xabraxas]

Die size doesn't matter a whole lot to the end user of the product unless we're talking laptops. Then it matters. But when we're talking desktops or servers (as is mostly the case with dual core 64-bit chips today -- few laptops are using 64-bit chips), then the end-user doesn't care so long as the case it's sitting in can keep it cool.

Not true. Most laptops are 64 bit now. All of AMD's present offerings are 64 bit. All of Intel's are too except for a few old Pentium M machines that are still selling in the low end. Most laptop chips are dual core today too.

Re:Does size matter?

[waferhead]

I dount a modern dicing saw would have issues with hex die.

Layout software and steppers OTOH...

Re:Does size matter?

[Riley Holmes]

Wow, I feel like a tard now. At least I'm a slightly more informed tard. Thanks guys!

Re:Does size matter?

[Calinous]

You could be selling more processors than you can produce - the problem is the fact that you are selling at rock-bottom prices. Just think that after the introduction of Intel Core2Duo, the prices for AMD processors were cut to half. Unlike any other theoretical (and not easily known) quantities as yield, cost per wafer and so on, a decrease in half of the selling price halves the bottom line. (Now I know not all AMD processors fell in price to 50%, but many of them did)

Re:Does size matter?

[Anonymous Coward]

longer design time

Nah, they just use the 'reduce' button on the copier.

Re:Does size matter?

[Calinous]

No, laptop users don't care a bit about the nm thingie a processor has. All they care is performance, heat, noise (sometime) AND battery life. As such, an laptop user would (better) choose a 90nm Athlon64 than a 65nm Pentium4, all else being equal.

Re:Does size matter?

[HuguesT]

Core duo is only 32-bit.

Core duo 2 is 64-bit.

Re:Does size matter?

[PlusFiveTroll]

Don't you mean number of transistors is heat? 486 processors had a very large area, but they pretty much ran cool to the touch.

Re:Does size matter?

[aminorex]

All the T2xxx series Core 2 chips are 32-bit, and will never address large flat memory.

I for one...

[Anonymous Coward]

I, for one, welcome our new Intel 65-nanometer overlords!

Re:I for one...

[man_of_mr_e]

Thing is, Intel is already shipping 45nm chips, though 45nm CPU's won't be shipping for a while yet. It's already working on sub-45nm technology.

Apple

[Gotung]

To all those AMD fanboi's that cried "Why not AMD"? when Apple choose Intel, this is why.
 
Disclaimer: I have nothing against AMD, I like there fact there is healthy competition in the chip world. Makes for better/faster/cheaper products for us consumers.

Re:Apple

[binary paladin]

Bah, I was hoping with only like three posts up I could be the first to mention that. My PCs use AMD processors almost exclusively and I have been using AMD since my bloody 286. My main work machine, however, is now a Mac (a PPC one currently).

Since Apple first announced Intel I thought that it was pretty obvious why they went that route rather than AMD. This, right here, was one of the main reasons. Supply programs have haunted Apple for quite some time. Why switch to a new architecture just to get more of the same?

Re:Apple

[Aadain2001]

Yup, this was one of those "bullet point" items why Apple went with Intel over AMD. Only Intel (and IBM) have the kind of manufacturing capacity that Apple needs in order to keep the supply chain moving while still providing to other customers. It would/will take many many years of constant fab building for AMD to compete with Intel on production scale.

Re:Apple

[MightyYar]

At least do a good job... Apple was like the #4 PC maker in the US last quarter. Make fun of Apple users if you want a flame war.

Re:Apple

[MightyYar]

Dell, too. You can see why they might have been hesitant to go with AMD.

Re:Apple

[99BottlesOfBeerInMyF]

As if Apple sells enough computers that it would really matter anyway...

Hmm, Apple's US market share is about 5%. AMD's US market share is about 18%. So had Apple gone with AMD, they would be accounting for about 1/4 of AMD's sales in the PC market. That would almost certainly make Apple AMD's largest customer.

Re:Apple

[nine-times]

Sure, especially when you consider that one of the problems Apple had with IBM/Motorola is that they were always in short supply of chips. Apple was constantly delayed in releasing new models because their processor suppliers weren't meeting their estimates.

Re:Apple

[certain death]

HAHA!! Thanks for the info there, I will keep that in mind the next time I require flamage :o)

Re:Apple

[Anonymous Coward]

Why the fuck did someone throw away mod points on an obviously "funny" post, or at least an attempt as such?

It was a fucking joke. If you don't find it funny, it does not mean it was a troll post, nor flamebait. It just means you did not find it funny. It's not like he called Jobs a racial epithet, or claim "only fags would use macs" or something equally deserving of a troll moderation.

Idiots. Stop throwing away mod points. Go mod up a more worthwhile post, please, and don't mod posts like this down. Some of us actually LIKE Macintoshes and still found it funny.

Re:Apple

[Niebieski]

Another reason they went to Intel instead of AMD is also power i.e. performance per watt), an area in which Intel had a better roadmap than AMD (and obviously IBM)

http://en.wikipedia.org/wiki/Apple_Intel_transitio n

Re:Apple

[toddestan]

Supply programs have haunted Apple for quite some time. Why switch to a new architecture just to get more of the same?

Except that it wouldn't be the same. Apple was dependent on Motorola and later IBM because they were the only suppliers of the PPC chips they needed for the Mac. However, since AMD and Intel both make x86 chips that function the same, they wouldn't be dependent on AMD if they went that route because they could switch over to Intel chips at any time without much trouble (and vice-versa I suppose - Apple could switch over to AMD right now if they wanted to).

Re:Apple

[jo42]

I'd hazard to guess that AMD is in this situation because of Dell...

Re:Apple

[Nimey]

Except they'd also have to use different motherboards, since the sockets and chipsets aren't compatible, and sometimes the memory isn't either, as before AMD switched to socket AM2. Better for them to have a single supplier, and Intel's the one with fab capacity (and a better chip, finally).

Re:Apple

[TubeSteak]

It would/will take many many years of constant fab building for AMD to compete with Intel on production scale.

You realize that when AMD moves from 90 to 65, they're going to get ~25% increase in yield? (possibly less, since more chips = more potential defects)

AMD has been using the previous generation of fabbing & is still strongly competing with Intel.

I just hope AMD has a better transition plan for the 65nm to 40nm switch, which will most likely be the industry's final step down in size.

Re:Apple

[jmorris42]

> To all those AMD fanboi's that cried "Why not AMD"? when Apple choose Intel, this is why.

Wrong. Had Apple signed AMD to supply them they would have certainly been in a position to have demanded first dibs on production. Seeing as there is zero likelyhood of Apple consuming the entirety of AMD's fab capacity......

Apple would have went with Intel regardless of quality of the good, delivery problems or anything else. Apple isn't about hardware or software or even technology, they sell a brand experience. To do that any subcomponent that is seen by the buying public as a seperately branded item must also be perceived as a 'premium' brand. Intel has spent billions creating that impression in the general public while AMD has courted the tech savvy customer.

Think of Apple as more like Nike. Like Apple, Nike does R&D and innovates in their products, but that ISN'T what they are selling. Same with Apple, they are selling the brand, not individual products. The Nike swoosh increases the value of a baseball cap, obviously there isn't anything 'better' about it other than the logo itself. Apple hasn't quite managed that feat... yet.

Re:Apple

[Overly Critical Guy]

To all those AMD fanboi's that cried "Why not AMD"? when Apple choose Intel, this is why.

Also, the fact that the Core 2 crushes AMD's chips while consuming less power, which was probably the most important factor. The Powerbooks were dying for a processor upgrade.

Re:Apple

[Aadain2001]

Intel has demonstrated working transistors all the way down to 1nm. The big question though, is can those transistors be manufactured in high quantities reliably? 40nm will not be the final step down, but we are getting closer.

For clarification, what do you mean by AMD competing with Intel? Are you referring to units sold to consumers, units sold to OEMs, manufacturing capacity, or design innovations? In some of those areas, Intel and AMD are neck-and-neck and trade places (who is in the lead) every few years. In other places, such as manufacturing capacity, Intel has always had the clear advantage. Intel is much further ahead in manufacturing processes, with AMD playing catchup. While AMD is transitioning to 65nm, Intel will be moving to the next step down. And when AMD finally moves to 40nm, Intel will have already transitioned to the next step down.

Re:Apple

[toddestan]

I suppose that would be an issue for a company that loves propriety motherboard designs. Though companies like HP manage to juggle both an AMD and Intel line up for their notebook computers, so I'm sure Apple could handle it. Besides, they wouldn't have to deal with the really tough stuff, as all their code for their Intel Macs would run on an AMD Mac just fine (well, except for that stuff that deals with the TPM, which is certainly another reason Apple went with Intel - the included DRM). Basically, switching between AMD and Intel would be a lot easier than the switch from PPC or even the switch from the 68k line - so it's really pretty silly to say that Apple would be tied to either AMD or Intel like they were to Motorola and IBM.

Re:Apple

[TheRaven64]

There were three main reasons for the transition:

• Intel could produce a large volume. Apple is quite a small customer to Intel. They would have been a major customer for AMD. They were a small customer for IBM and Motorola too, but Intel is selling the same product to lots of other people so Apple isn't shouldering all of the risk (if Intel gave up making x86 CPUs then the market is big enough for it to be profitable for someone else to sell Intel-compatible chips and provide an upgrade path for Intel customers).
• Intel had a good performance-per-watt roadmap. IBM still can't produce a low-power 970 that has decent performance (their portable version is something like 1.2GHz).
• Intel had the whole platform. Apple had to design their own north bridge chips for the PowerPC systems. Intel ad the entire Centrino system, including low-end GPUs and WiFi chipsets, keeping Apple's costs right down.

Re:Apple

[Bing Tsher E]

You mean they're in this trouble because Dell isn't buying enough AMD processors?

I heard the Ford Motor Company was in trouble, too. Is it because Dell isn't buying enough Ford trucks?

Re:Apple

[Bing Tsher E]

The Nike swoosh increases the value of a baseball cap

I've gone out of my way so long to call it 'The Adidas swoosh' that that's now what I think of it as. It's almost time to start calling it 'the New Balance Swoosh' for awhile.

And actually, the 'Nike swoosh' just increases the probability that a dumbass will purchase a baseball cap. Which effectively lowers the value of the brand, unless selling to dumbasses is a profitable venture. (oh, wait....)

Re:Apple

[modecx]

For clarification, what do you mean by AMD competing with Intel? Are you referring to units sold to consumers, units sold to OEMs, manufacturing capacity, or design innovations?

It's pretty obvious that in the context of "competing with Intel", he meant in the "outright processor performance" category, as in that AMD has continually used/s lower tech to compete with, if not frequently surpass the performance intel processors demonstrated--and usually at a lower price, as in the only thing normal people pay attention to: price/performance. He didn't mean megaflops/kWh, nor did he mean squirrels per cubic hectare, nor any of the stuff you quoted, which are things only boring old smelly investors would be interested in, anyway.

Re:Apple

[Aadain2001]

NetBurst based processors sucked. The new Core 2 processors kick AMD's ass up and down the block AND consume less power. AMD isn't competing much on price, performance, or availability.

Re:Apple

[mgblst]

One of the reasons for apple success, is that they don't give you a 1000 choices of laptops, or ipods. You get 3 or 4 choices, which actually makes things easier for the customer. If Apple were to do this, it would needlessly complicate the choices for a lot of people. I can see why they only wanted to choose one company.

Re:Apple

[Ilgaz]

Why not AMD OPTION, some of us said. It was clearly an exclusive agreement with Intel locking up Apple to Intel brand processors.

Why XServe can't use AMD server CPUs? They don't have any kind of supply problem and every serious server manufacturer gives the option.

Think about why...

While moving to X86, Apple should get rid of single CPU manufacturer problem. They have chosen not to do it. X86'es power comes from choice.

What if Intel has some problems in the future? Impossible? Well, while using Amiga 500, if anyone told me Commodore will go out of business years later, I would laugh to his face. Same for Atari too.

There isn't a single invulnerable company in this business.

BTW, I have a good guess what will fix AMD problem, IBM will rent a factory as old days. That's all.

I am a quad G5 owner and I would buy AMD only if I setup a huge server or rendering farm. So, please don't call people "fanboy" while they speak about choice.

Disclaimer: I have too much things against Apple, especially our great Cult leader, Steve Jobs to be correct. Making Apple a Intel posterchild right after switching CPU Arch, tricking entire Mac community that "universal binary" can be done with single click (yes, those stupid developers can't find that switch), giving bad name to PowerPC while only their lame ass G4'es had FSB etc. problems, manufacturing low quality Intel based laptops tricking people with brand, making an entire community Intel puppets starting to bitch about AMD too, stating never-heard lame "4x faster" allegations comparing Xeon based servers to G5s...

Anyway, entire gaming planet has chosen Power architecture based CPUs, on their 1920x1080 3d output monsters even.

Re:Apple

[Ilgaz]

Apple supply problems. FreeScale said only 5% of their processors were purchased by Apple. I have never heard any problem with G5 (PPC970) from IBM too.

Apple can switch to AMD right now? If there isn't an exclusive agreement with Intel,why not? I can't remember any Intel PR person promoting Intel CISC CPUs like Steve Jobs did...

Re:Apple

[Anonymous Coward]

Actually, Intel copied the DRM from AMD.

Re:Apple

[aminorex]

You are mistaken. If Apple had contracted for certain chip deliveries, AMD would have fulfilled those contracts.
No, the well-established and fully-understood and in no-wise mysterious reason is that Intel is subsidizing Apple
to switch over from PPC to x86. Intel had deeper pockets (and was more motivated to get Apple on the x86 ship-of
fools). As a result, you don't get 64-bit 8GB RAM Turion laptops from Apple, but Apple gross margins are higher,
and the shareholders are as happy as pigs in slop.

Re:Apple

[aminorex]

You are mistaken. If Apple had contracted for certain chip deliveries, AMD would have fulfilled those contracts.
No, the well-established and fully-understood and in no-wise mysterious reason is that Intel is subsidizing Apple
to switch over from PPC to x86. Intel had deeper pockets (and was more motivated to get Apple on the x86 ship-of
fools). As a result, you don't get 64-bit 8GB RAM Turion laptops from Apple, but Apple gross margins are higher, and the shareholders are as happy as pigs in slop.

Re:Apple

[aminorex]

Core 2 Duo CPUs suck donkey balls when you try to run them in 4-way and 8-way systems.
Also, they will *never*, ever, ever integrate the core architecture with GPUs, like
AMD is doing with ATI (see http://www.internetnews.com/ent-news/article.php/3 640251">here for press on the topic).
But yes, they are wonderful for single-socket desktops under 4GB. I wouldn't consider buying
anything but Core 2 for a single-socket desktop, today. Nor would I ever consider a
Core 2 Duo for an SMP enterprise server, a blade system, or a laptop or a desktop using
more than 4GB of RAM. All of these applications belong to AMD 64-bit CPUS.

Re:Apple

[Aadain2001]

Wow, you are such a moron I'm not even going to take the time to walk you through, point by point, where your assumptions are completely wrong. Have fun being an idiotic AMD fanboy.

No wonder AMD is having problems...

[Anonymous Coward]

From the News.com article: "AMD's current pickle ..."
Well, *No wonder* AMD is having problems... they should NOT be making pickles, they should be making chips!

TDz.

Re:No wonder AMD is having problems...

[HermMunster]

You mean Dill chips?

Re:No wonder AMD is having problems...

[Anonymous Coward]

yes. more productivity per cucumber.

Re:No wonder AMD is having problems...

[dacaldar]

Why do you think AMD just bought (Canadian company) ATI?

AFAIK, Ketchup and Dill Pickle chips are only sold in Canada, not the U.S.

Re:Balls in the air?

[MyEyesTheyBurn]

Obviously we are outside the box if our balls are in the air.

No Matter What...

[hurting now]

the consumer wins. I was an AMD fan boy for the past few years, but like a true Chicago fan, I am rooting for the other team because they are up. AMD may strike back again, maybe not, but this price war has really benefited many of us.

Re:No Matter What...

[kimvette]

No kidding. Thanks to the price war I scored a 2.4Ghz Core 2 Duo really cheap, and have it happily overclocked at 3.06Ghz. I wanted to go AMD, but I didn't because:

1. They are no longer cost-effective in comparison to Core 2

2. Compatibility issues; athough the chipset I chose is not 100% supported on the kernel rev I'm running, it's still a far sight better than getting an ATI or NForce chipset to run acceptably well. Also, The AMD-ATI merger does not bode well for Linux users, given ATI's abysmal track record. I refuse to buy ATI products and am now avoiding AMD until I see whether or not ATI cleans up its act. (insert a rabid "fuck ATI" right about here)

3. at stock clock speeds, it is 80% faster (according to benchmarks) than the Pentium D I sometimes use at the office, and well over 100% faster than the Pentium 4 (the other box) my primary box at the office. At 3.06Ghz, it's (obviously) much faster than even that. :)

I wanted to go AMD, I really did, but with Intel's quad core processor coming out Real Soon Now and with my board's already being certified to run it, it was the logical choice. Quad core upgradability was the clincher.

Re:No Matter What...

[Chris Burke]

Also, The AMD-ATI merger does not bode well for Linux users, given ATI's abysmal track record. I refuse to buy ATI products and am now avoiding AMD until I see whether or not ATI cleans up its act. (insert a rabid "fuck ATI" right about here)

Naively, since AMD, the larger company, bought ATI I would figure AMD's culture of openness would dominate. At least, I hope this is the case, because I'm sick of having only one choice for graphics cards (though I'm used to it; ever since 3dfx).

Anyway, it doesn't seem fair to refuse to buy AMD's processors (as opposed to graphics cards/chips) because of ATI's poor linux support when AMD has been a huge supporter of free software, in particular of gcc and to a lesser extent the Linux kernel.

Though I will still not buy AMD/ATI (or DAAMIT as the Inquirer calls it) graphics products until they do actually improve. And I will continue buying AMD processors until they start becoming more like ATI.

Re:No Matter What...

[TehDuffman]

I don't know what kind of Chicago fan you are...

History repeats itself

[Sebastopol]

AFAIK, this has always been AMDs problem: my earliest recollection is when they bought NexGen's K6 and sold it to Compaq in the sub-$1000 segment in 1995. Since then, anytime the get a good product, they blow it on production, leaving Intel to fill the void they created.

It is where they have failed again and again and again. I can't believe they haven't learned yet.

Re:History repeats itself

[zensonic]

they blow it on production, leaving Intel to fill the void they created

It's a bit harsh to say that they blow it on production. You do realize how much it costs to construct cleanrooms, right? All while still being profitable and pour lots of money into R&D.

I think they are playing their cards to the best of their abilities.

Re:History repeats itself

[Libertarian001]

Increased supply requires increased production capabilities, which more or less requires a new fab (factory) being built. No, generally you can not just add additional tools to an existing factory. Nor can you just make your workers work harder/faster. Most of the steps in the process are automated and require definite process times. So it's going to cost you. A state of the art fab today runs ~$4-5 Billion. It also takes well over a year to bring it up to full production status (depending on how many tools you have. Intel's Fab 12 had 60 diffusion furnaces, Fab 22 across the street had 38). And that's part of the problem. A fab doesn't make money unless it is running at capacity.

I worked AT (not FOR) Intel for 3 years.

Re:History repeats itself

[kfg]

Tell it to the investors who withdraw funding when demand goes up and they need to put more fabs online to meet it. Demand is up, but margins are down. So is the stock price. There goes the money to build new fabs.

It's all about capitalization. Not being able to meet the captital requirements of an excess of success is the second most common cause of business failure. Outright stupidity only comes in at number three.

KFG

Re:History repeats itself

[Sebastopol]

...or do a better job with managing profits.

Can't blame the investors for a company's failure to supply customers.

Re:History repeats itself

[Sebastopol]

I'm well aware of the costs of a fab.

Well, they have a ton of fab capacity coming onlin in the next 18-24 months. Dresden was largely due to their K7 success, but it takes years to make a fab: they planned dresden 5~7 years ago.

Re:History repeats itself

[kfg]

Can't blame the investors for a company's failure to supply customers.

Yes, you can. Why? Because they are often the cause.

Why do you think they call it "capitalism"?

KFG

Re:History repeats itself

[Sebastopol]

Um, what?

Stay in school, kid.

Re:History repeats itself

[kfg]

Um, what?

Horticulture.

KFG

Re:History repeats itself

[ocbwilg]

AFAIK, this has always been AMDs problem: my earliest recollection is when they bought NexGen's K6 and sold it to Compaq in the sub-$1000 segment in 1995. Since then, anytime the get a good product, they blow it on production, leaving Intel to fill the void they created.

Intel isn't in much better shape right now. Sure, the Core 2 Duo rocks. Everybody wants one. Unfortunately, the Core 2 Duo CPUs are still less than 25% of their desktop production. The rest of it is still the old Pentium D and Pentium 4 CPUs, which no one wants. Which is bad news, because Intel has large stocks of the older gen chips still unsold.

So if you can't get as many Core 2 Duos as you want, your choice is either the Pentium 4/D or the Athlon 64/X2 (which will outperform the Pentium 4/D by a large margin). Apparently most companies are choosing AMD. That's one of the things that the Intel fanbois kept forgetting. It doesn't matter how much faster the Core 2 Duo CPUs are if Intel isn't manufacturing enough of them to meet demand. AMD's capacity problems are at least partly related to Intel's (and the fact that Dell is eating a significant portion of AMD's production).

How long?

[MyEyesTheyBurn]

How long before we can expect a 5-nanometer based CPU?

Re:How long?

[chrismcdirty]

Screw that. They'll come out with 5nm, then right after I buy it, they'll announce their 900pm chip.

Re:How long?

[Libertarian001]

IIRC...

45nm in 2009
33nm in 2012
25nm in 2015
12nm in 2018

I don't remember plans for sub-12nm, but I'd assume it'd be about the same 3-year cycle...

9nm in 2021
6nm in 2024 ...and that's assuming we don't hit any physical limitations. Power leakage is going to continue to be a major problem for the near future. Intel's 3D transistors should help quite a bit. But we'll see.

Re:How long?

[Anonymous Coward]

You forget, new chips most likely wont use silicon wafers.
Instead they'll go back to potatos!

Mmmm Intel potato chips.

Re:How long?

[Anonymous Coward]

By then, the real skynet will have become self-aware...

Re:How long?

[akuma(x86)]

Process steps move forward every 2 years.

45nm in 2008
32nm in 2010

more generally:

0.7 * gate_length(N) in year(N) + 2

Thermal limits will hit hard in 10+ years. No exponential can go on forever.

Re:How long?

[bazorg]

0.7 times something doesn't look very exponential to me...

Re:How long?

[akuma(x86)]

It's 0.7 in one dimension.
So in terms of area, it's 0.7*0.7 = 0.49 or roughly half.

Therefore you double transistor density every 2 years - a nice binary exponential.

Not new

[cheezedawg]

This is not a new problem for AMD. They have always had problems keeping up with demand, and they have been capacity constrained for a few years now, and they have nobody to blame but themselves.

That's the dirty little secret about the semiconductor industry- success depends just as much on manufacturing ability as the features of the chip. Intel didn't just get their 300mm wafers and 65nm process overnight- they invested 10s of billions of dollars in manufacturing R&D. The result is they have unparalleled capacity and a huge technological lead over competitors with manufacturing technology. When a large OEM comes asking for 5 million units in the next quarter with a defect rate of less than 500 per million, there are very few companies that can deliver.

Re:Not new

[Dastardly]

Yep, that huge technological lead of a few months. This is the first time since really K5 that AMD has had production issues. And, the production issues described in the article are really the good kind. The early ones were the bad kind where the chip design itself caused the production problems. The good kind is where yields are good, speeds are good and the design is manufacturable. Your only issue is figuring out how to push more die through the line.

Fab 36 will be online soon with 300mm wafers and 65nm. Just going to 300mm wafers pretty much doubles capacity. Going to 65nm gets you another say 50% (anyone got a confirmed number). Getting FAB 36 and FAB 30 going doubles capacity again. So, by my calculations that is 2 x 2 x 1.5 or 6x cpacity increase for AMD in the next couple of months.

Did Intel switch to 65nm and 300mm sooner than AMD? Yes. Did they switch to copper and low-K dieelectric before AMD? No. Did their 90nm production even work quite right for Intel ever? Not sure. When the 90nm P4s used more power than their 130nm brethren you have to wonder.

Of course Intel has something like 6 processro FABs all over the world that are likely larger than AMDs. Doesn't take much R&D just to build more capacity especially when you are the 800lb gorilla.

Basically, Intel and AMD, at this time, are quality processor maufacturing operations. Intel tends to make technology switches before AMD, but they also get to deal with first adopter issues. And, when they both buy there equipment form the same semiconductor equipment manufacturers like Applied Materials, Novellus, and others. How much is AMD benefitting from Intel working out the bugs.

Re:Not new

[Anonymous Coward]

Going online and shipping volume production are two completely different things. It can months to qualify new fab equipment, processes, etc. By that time, Intel will be shipping 45nm parts and will be ahead once again.
 
And to claim that this is the first production issue they have had since K5 is completely forgetting the SOI mess that AMD found themselves in when they were trying to push Clawhammers out the door a couple of years ago.

Re:Not new

[Mokurai]

Going from 90 nm to 65 nm results in 1.9 times as many chips.

(90/65)^2

Re:Not new

[Chris Burke]

Did Intel switch to 65nm and 300mm sooner than AMD? Yes. Did they switch to copper and low-K dieelectric before AMD? No. Did their 90nm production even work quite right for Intel ever? Not sure. When the 90nm P4s used more power than their 130nm brethren you have to wonder.

I think Intel's 90nm process was working just fine such as it is. In my opinion the fact that the 90nm P4 burned more power than the 130nm P4 was due to two things:

First, the Prescott core that was released in 90nm had more pipe stages than the 130nm Northwood core, in particular more pipe stages operating at the double-speed ALU clock frequency, so this means many more transitions per cycle. Other architectural changes probably added to an increase in dynamic power.

Second was the elephant in the room that all silicon manufacturers were trying to ignore, but which suddenly started stampeding around crushing things in a non-ignorable fashion at 90nm: Leakage current. All transistors conduct a small amount of current even when switched off. The problem is that this current rises as the length of the transistor decreases. Everyone saw that leakage was becoming a greater proportion of the total power consumed, but weren't doing much about it. Apparently 90nm was some kind of tipping point, and suddenly leakage was burning 25-50% of the power in the chip and you couldn't do anything about it other than power the transistors down.

Combine these two things, and Prescott was a power monster. Other chips that have less of a problem with dynamic power were better able to handle the increase in static power.

In computer architecture there's an age-old debate about the "speed demon" vs the "brainiac" which has gone on forever with no obvious answer. Roughly speaking the P4 is a speed demon, K8 and P-D are brainiacs. While I could argue about architecture and why one is better than the other, the outcome in reality has been decided. It is interesting that in the end the Speed Demon lost, and power was what struck the killing blow.

Re:Not new

[Chris Burke]

It's really not that simple. The process size is usually given as the minimum transistor length. Other parameters (minimum metal trace width, minimum distance between active regions, etc) may not shrink by the exact same amount. So you can't just scale your 90nm design by 65/90 and expect it to pass the design rules. I'd say that 1.5 as many chips is a reasonable guess. If I wanted a better guess, I'd look up the die sizes of previous chips that made process transitions with a minimum number of architectural changes to get an idea what real scaling factors you can expect.

On the other hand, the smaller your chip the better wafer utilization you get because the square chips better fit into the edges of the round wafer. This isn't a huge effect, though, unless you're talking really huge chips vs really small ones.

The big deal is that AMD's 65nm process is also coming with 300mm wafers. That's 2.25x the number of die/wafer right there, plus a little bit from being able to again pack chips to the edges of the wafer better.

Fab 36 is going to mean a hell of a lot more capacity for AMD. They're probably estatic to be having this "problem" right now.

Re:Not new

[onx]

Fab 36 is ALREADY online, albeit if at only 50% capacity, and using 300mm wafers.

Fab 30 is currenty running at full capacity, but next year it will drop to 40% capacity as it switches to 300mm wafers & gets echristened Fab 38 in 2008.

AMD is/will be outsourcing production to Chartered even though they just brought their new fab 36 online.

I think Nikkon is also a big semiconductor equipment manufacturer. http://www.nikonusa.com/template.php?cat=3&grp=79

Re:Not new

[default luser]

Just going to 300mm wafers pretty much doubles capacity.

Yup. The total area more than doubles moving between the two.

Going to 65nm gets you another say 50% (anyone got a confirmed number).

There are no confirmed numbers, because semiconductor makers play their hands close to their chests. All the interesting data is typically proprietary and treated as a trade secret.

But the better estimate would be a straight shrink of the feature size + 5-10% for optimizations. The straight shrink of a 90nm feature size to 65nm, assuming scalable design rules, is a %28 gain. I'd say with custom reworking, you could gain as much as %35 production capacity over a 90nm line.

Getting FAB 36 and FAB 30 going doubles capacity again.

Fab 30 already exists. It is a 200mm 90mn facility. It produces the vast majority of AMD's microprocessors. The rest are trickling out of Fab 36 (which has limited 90nm production, and is ramping up 65nm), or are supplied by Chartered Semiconductor. Chartered looks to be a long-term partner (much like AMD used to be an Intel second-source), since they've licensed AMD's Automated Precision Manufacturing methodology.

AMD does have plans to rework Fab 30 as a 300mm, 65nm fab (and rename it Fab 38), but that won't kick off full-swing until Fab 36 reaches full production (middle of next year).

When you Chartered into account, I figure Fab 36 will only double their capacity. But with the transition of Fab 30->Fab 38, and the opening of the fab in New York (planned for 2010, 2011), there is potential for much-increased production. It amazes me that AMD can hold a %17 marketshare against Intel despite only having "one and a half" fabs. But when you look into it, while Intel has FIFTEEN fabs, only four of them are cutting-edge 300mm, 65nm. By concentrating on one core business, AMD can do more with less, but it also makes them more vulnerable as a company.

Expect 65nm chips from Fab 36 in the coming months, but don't expect the full increase in production until middle of 2007. Quadcore will be available from AMD at that time as well.

Re:Not new

[evilviper]

They have always had problems keeping up with demand, and they have been capacity constrained for a few years now,

More urban legend than reality. It's odd that AMD gets tagged as having "shortages" while Intel gets the opposite, despite MORE actual instances of Intel chip shortages than AMD has experienced in the past ~5 years.

The result is they have unparalleled capacity and a huge technological lead over competitors with manufacturing technology.

Right, that's why there were widespread Intel chipset shortages, forcing many OEMs to switch from Intel chipsets, or else stop making new systems...

The result is they have unparalleled capacity

That's funny, because the demand for the Intel Core chips caused more of a shortage of Intel CPUs than AMD has seen.

Intel, like AMD, doesn't have significant excess capacity. They are both living with minor shortages, in other to undercut each other.

In addition, despite the story, there's no information in the article, or anywhere else, indicating AMD is actually experiencing CPU shortages. It's implied, but that's all. The source for all this seems to be The Inquirer, which is wrong as often as it is right. If anyone has some reputable information that this is the case and not just more rumor and myth, please link it, by all means.

Re:Not new

[akuma(x86)]

Yep, that huge technological lead of a few months.

The lead is 1 year. Intel shipped 65nm in December 2005. That is HUGE. You get to offer 2x transistor density for 1 whole year. Your cost per transistor is 50% cheaper for a year. In case you missed it, Intel made about 40 Billion dollars in its last fiscal year. That's no accident. They're able to generate that massive cashflow because they are one step ahead of everyone in the industry.

Now, if you screw up the architecture, it turns out that you can be really inefficient at using those extra transistors (see Pentium 4). But that's been fixed with Core2Duo.

IMO, AMD is pretty screwed now that Intel is making better use of their transitor density advantage.

The process also improves continuously. So while AMD will have a new 65nm process in December, Intel will have a 65nm process that has been improved for a whole year. Just look at the frequency headroom that overclockers are getting with Core2Duo. Now look at the AMD 65nm roadmap. The first 65nm AMD chips will clock LOWER than the 90nm chips. This is pretty normal for a new process as the parameters need to be tuned and refined over time.

Fab 36 will be online soon with 300mm wafers and 65nm. Just going to 300mm wafers pretty much doubles capacity. Going to 65nm gets you another say 50% (anyone got a confirmed number). Getting FAB 36 and FAB 30 going doubles capacity again. So, by my calculations that is 2 x 2 x 1.5 or 6x cpacity increase for AMD in the next couple of months.

90nm->65nm is approximately a 0.7x shrink in one dimension. (90 * 0.7) ~= 65
In 2 dimensions it's (0.7 * 0.7) ~= 0.5 so yes, a 50% shrink in area.

But what you forgot to mention was that the die size doesn't stay constant across processes.
Engineers tend to do useful things with this bounty of transistors - like double the cache size for example.

Of course Intel has something like 6 processro FABs all over the world that are likely larger than AMDs. Doesn't take much R&D just to build more capacity especially when you are the 800lb gorilla.

Uh...doesn't take much R&D???
How do you suppose that Intel gets to the next process step a full 1 year ahead of everyone else on the planet? Magic?

How much is AMD benefitting from Intel working out the bugs.

Not much considering that AMD is using a vastly different process. SOI vs. bulk is just one of many MAJOR differences.

Re:Not new

[Dastardly]

It really is not 1 year. This is the strange thing about Intel having so many more Fabs. They get the first chips out, but that is not every Fab putting out 65nm. It is most likely one Fab, which would be the smart thing to do, it would be incredibly stupid to switch all production to 65nm at the saem time. So, it is very likely that Intel will be 100% 65nm at about the same time AMD is. It is hard to make an real solid comparison. ANd, given that AMD does not have 4 cutting edge Fabs I think a good argument could be made that AMD is behaving correctly, even if Intel gains some advantage. You do not want production problems caused by a transition when you only have 1.5 Fabs. I just do not think it is as clear cut or as significant an advantage for Intel as it may seem on its face. Advantage: Yes. Big Advantage: ???

Re:Not new

[Dastardly]

Hate replying twice.

Of course Intel has something like 6 processro FABs all over the world that are likely larger than AMDs. Doesn't take much R&D just to build more capacity especially when you are the 800lb gorilla.

Uh...doesn't take much R&D???
How do you suppose that Intel gets to the next process step a full 1 year ahead of everyone else on the planet? Magic?

I meant their capacity is not necessarily a direct function of R&D. Of course it takes R&D to develop processes.

As I said I do not think it is as simple a matter as Intel is huge and spends, so much more money on R&D therefore AMD is hosed.

Oh and. IBM had 65nm ASICs in June 2005 6 months ahead of Intel.

Re:Not new

[akuma(x86)]

It is 1 year. AMD only has 1 fab. So if Intel has 1 fab first, then it has the advantage first.
Intel can undercut AMD on price since it benefits from the per-transistor cost advantage. Prices drop on the older N-1 process chips, and by definition, that means prices drop on AMD's N-1 process for 1 year, assuming Intel/AMD are equally efficient at using their transistors. So, for a whole year, AMD is forced into a lower-end market.

You also forgot to mention that the process isn't static. It improves on a daily basis. That's why you get faster chips released every few months on the same process. That's a 1 year advantage to tune the process as well. As more fabs come online, they also inherit this advantage. You can't port those tunings to another competitor (like AMD) when their process is so radically different. SOI presents a very different optimization problem vs. Intel's non-SOI process.

The whole business is predicated on building fabs on the latest technology as soon as possible.

Re:Not new

[akuma(x86)]

Oh and. IBM had 65nm ASICs in June 2005 6 months ahead of Intel.

There's a huge difference between an ASIC process and a high volume processor process.
When did IBM ship a 65nm POWER processor?
Did they even ship one yet? Let me call Steve Jobs.

Re:Not new

[aminorex]

> 35%

No. (90/65)^2 ~= 2. The shrink doubles the available silicon area. (1.92x, taken to 3 digits.)

But, while AMD is making the transition to 65nm from now through 2008, with an increasing proportion of capacity being 65nm over that time frame, aren't others already ramping up 45nm processes? Thats a 70% advantage, again. I don't recall whether it was Intel or IBM, but one of those two, I think, are already bringing up 45 nm fabs. How they will manage to deal with leakage current at that line scale is another matter entirely.

Watch out for an Axe...

[Numbah One]

AMD has a lot of balls in the air right now

Hopefully, they won't get them chopped off.

Bad means good.

[k4_pacific]

For AMD Success Means Problems, and also, to shake your booty means to wiggle one's butt. Permit me to demonstrate.

Re:Bad means good.

[Anonymous Coward]

Your application for a permit to shake your booty has been denied.

Science behind the miniaturization

[MarcoAtWork]

when I took microelectronics courses in university about 15 years ago the lower limit for our process was around 2um (if I remember correctly) and my professor several times seemed to strongly believe that the lower limit for gate length was around 0.6-0.7um for various reasons. Nowadays we're way smaller than that, and it's getting even smaller as time goes on: is there a website somewhere that details exactly which theoretical advances have been made during the past 10-15 years to enable processes to continue getting smaller?

Re:Science behind the miniaturization

[Mokurai]

The advances have been practical, rather than theoretical. They include

Shorter wavelength light sources, now including ultraviolet lasers. An X-ray laser was recently demonstrated, but is years away from production.

Better resists, including resists designed to work with ever-shorter wavelengths.

Greater precision in micropositioning devices. I once worked on a system that had 0.25u horizontal precision with stepper motors, but 12 nm vertical precision with a piezoelectric actuator. Things have gotten a lot better since then.

Better masks. Sharp-edged mask features create fuzzy shadows by diffraction. The exact right fuzziness in the mask edges creates sharp shadows.

Better mask-making tools, including e-beam.

Better materials. Copper, silicon-germanium, strained silicon, silicon on insulator...Increasing electron mobility and decreasing parasitic capacitance.

There are other factors, of course, like chemistry and process equipment. But to answer the precise question you asked, Yes. Wikipedia is your friend, and will tell you where else to look.

Re:Science behind the miniaturization

[MarcoAtWork]

I seem to vaguely recall my prof mentioning theoretical limits (quantum tunnelling maybe? I seem to recall it was some sort of quantum effect, apologize for the vagueness, it's been 15 years and I've been in software since then) that would prevent things from working if the gate length became too short. I have been looking on wikipedia but I haven't been able to find anything specific.

Re:Science behind the miniaturization

[chrwei]

about 20 years ago, the company my dad worked for made flooring materials, like the epoxies and such. one the projects he worked on was for an electronics company's (may have even been Intel) clean rooms. the problem they had at the time as I understood it was that epoxy floors, which are fairly easily made impermiable and anti-static, caused small, conductive, particles to be released for years as it continued the "curing" process. these particles were just big enough to get stuck in the chips and short them out. I don't know that he ever told me the solution, or if he was even involved in solving it, and I was 15 at the time so maybe I understood it wrong. I do recall being impressed that tiny things you couldn't possibly see were flying up out of the floor and shorting out electronics.

Re:Science behind the miniaturization

[sunspot55]

The CMOS process has also undergone substantial changes to address the problems that you encounter operating devices with smaller geometries. For instance lightly doped drain implants to counter hot electron injection were developed then halo implants for short channel effects. The methods that are used to isolate transisors on an IC (a critical element) have evolved over the years. From the LOCOS process that was prevalent at about that technology node to trench to SOI people just keep picking away at the technological hurdles that stand in their way.

Just as the materials and tooling of photolithography have been improved as mentioned above the CMOS process itself has undergone shifts to help smaller transistors remain operational and reliable. The current CMOS process flow would look very different to a process engineer transplanted from 1989.

Re:Science behind the miniaturization

[wwwillem]

Big change since "the old days" in clean rooms is that in the past the whole factory was kept at let's say 1 ppm level. Nowadays it's just the encapsulated equipment that's kept at a low particle level, while the clean room itself can be more moderate. Still not your average neighborhood garage of course. :-)

Besides the huge cost savings in air treatment equipment, the big advantage is also that for workers in the fabs the conditions become much better. As an example, for 10 ppm simple suits and caps are enough. For let's say 100x better you suddenly need space-suits and plexiglas caps. You get the picture.

Re:Science behind the miniaturization

[Anonymous Coward]

I am taking a microelectronics course right now and the lower limit for us is 500nm. We are taught how small the limits are in reality, but our simulation software is a bit old and doesn't know all the new materials and manufacturing processes. That is OK with me, as there is no hope of getting the latest and greatest software, especially if you study at an university in not-too-wealthy Eastern European country. Moreover, our textbook is printed in 1988, so it doesn't even mention copper, for example. So we just have to believe our professor when it comes to developments of the last decade. But, as others have mentioned, some basics can be found on Wikipedia.

Re:Science behind the miniaturization

[aminorex]

Deep submicron is done with UVA and now X-ray photolithography. The optical process has kept up with the curve, but the electronics are going south as leakage current predominates and parasitic capacitance explodes out of control.

Poor grammer in parent subject

[pkbarbiedoll]

Okay kiddies, say it together: commas are our friends!

For AMD, success means problems.

Re:Poor grammer in parent subject

[bmin]

News Flash poor spelling! It's grammar

Oblig. Boondock Saints

[MrSquishy]

Symboligy? I think the word you're looking for is "symbolism"

Re:Oblig. Boondock Saints

[value_added]

Symboligy? I think the word you're looking for is "symbolism"

He lost me at "falic".

Re:Oblig. Boondock Saints

[bdonalds]

seeing the terms "symboligy" and "falic" does NOT make me feel like Riverdancing....

Re:Oblig. Boondock Saints

[nowhere.elysium]

That would be "phallic" - to whit, indicative of the penis.
I haven't the slighest clue why (s)he's so tuned into it, though: the word 'pickle' is not overly representative of Freudian imagery.

Re:Oblig. Boondock Saints

[Fallen Mongoose]

Isn't Symbology that crazy religion that Tom Cruise follows?

Re:Oblig. Boondock Saints

[Phu5ion]

That would be "phallic" - to whit, indicative of the penis.

Are you an expert in... nameology

Re:Oblig. Boondock Saints

[N3Roaster]

No, I'm pretty sure Symbology is what they called magic in one of the Star Ocean games.

Re:Oblig. Boondock Saints

[dreadclown]

the word 'pickle' is not overly representative of Freudian imagery

"Let go of your pickle!"
"I'm not holding my pickle!"
"Then who's holding your pickle?"
-- "Titties and Beer", Frank Zappa

Sounds pretty Freudian to me.

Moo

[Chacham]

It's better than Intel's Pentium problem. They simply couldn't do the math!

Q: Why did they call it a Pentium instead of 586.
A: When they booted up the first Pentium and added 100 to 486, it answered 585.32752365107239874

Re:Moo

[TheRaven64]

At that time, some people at TI wrote a script for a spoof of the 2001 scene with Dave stuck outside. The problem with HAL was, of course, that he was built with a Pentium. It ended with HAL singing:

Daisy, Daisy, give me your answer, do.
Getting hazy, can't divide three by two.
My answers, I can not see 'em,
They're stuck in my Pentium

It would be sweet, my answers fleet on a workable FPU

Grammar Troll

[Anonymous Coward]

Did anyone else have trouble reading that headline? Perhaps one of the following two options would have worked better (at least for me):

Success Means Problems For AMD

-or-

For AMD, Success Means Problems

New 65nm AMD fabs coming on line

[Animats]

AMD is converting Fab 30 in Dresden from 90um and 200mm wafers to Fab 38, with 65nm and 300mm wafers. This should come on line in 2007. Longer term, AMD is building a new fab in upstate New York for 32nm features on 300mm wafers. That should come on line in 2010.

Meanwhile, AMD's main fab, Fab 36 in Dresden, is starting to produce 65nm features on 200mm wafers. AMD is also outsourcing some production to a 65nm fab in Singapore.

Down at the user level, this means that first shipments of AMD CPUs made with 65nm technology should appear in December of 2006. Coming soon to Dell Dimension desktops.

Re:New 65nm AMD fabs coming on line

[Ignignot]

Intel has a better shrink timeline for each of those steps compared to AMD. Intel will ship the next step from 65 (45? I forget) in 2H 2007. AMD is looking at 5 more years of lagged shrink.

Slight gripe

[uarch]

... but its dual-core Opteron and Athlon 64 processors contain two processing cores integrated onto a single piece of silicon, or a die. This design has given AMD great performance during the past few years ...

That incorrectly implies Intel's designs aren't two cores on a single die. They've been putting two cores on a single die for quite a while now.

Re:Slight gripe

[Anonymous Coward]

First dual core attempts by Intel were MCM designs, not single die.
Core2 is finally dual-core single die.
Quad Core2 (or whatever) is another MCM.

AMD X2 is dual core, single die.
AMD Quad will be single die as well.
And, of course, AMD's cpus have the integrated memory controller.

AMD has finally broken into a server market that is not as fickle as enthusiasts and yields the majority of revenue. With big names like Dell, HP, Lenovo, and Bejing Founder Electronics (Dell and Founder recently inked) they are set up well to capitalize on Intel's recent slip-ups. Netburst opened alot of doors that should keep things interesting for awhile. The ATI takeover just shakes things on top of that.

Cost too

[EmbeddedJanitor]

Cost per square inch is reasonably constant across processes. If you can pack in more devices per area you save cost.

ehm... increase the price by 5%

[Colin Smith]

That's the traditional thing to do when demand outstrips your ability to supply.

Re:ehm... increase the price by 5%

[servognome]

That's the traditional thing to do when demand outstrips your ability to supply.

Then you potentially alienate your customers, especially in a very competitive environment.
"Hello Dell, not only are we not going to be able to give you enough chips to meet your orders, we're going to raise the price on you too." If that's not bad enough, because the CPU is just a part of the system, the customer is going to have to sit on a bunch of inventory (motherboards, network cards, cases) they can't build, which is going to cost them even more money.

Re:Balls in the air?

[drinkypoo]

Obviously we are outside the box if our balls are in the air.

This reminds me of a joke I read once about a guy named Dolcetti. Dolcetti's banging his old lady on the kitchen floor when suddenly he starts screaming and really thrashing around like a madman.

"What's the problem?", says she. "Are you trying to get your balls in?"

"No!" screams Dolcetti, "I'm trying to get them out!"

It was a lot funnier when I read it than when I typed it. Must have been a long time ago.

Re:If I had a nickel...

[painQuin]

would you have 1 nickel?

Are AMD processors more stable than Intel?

[dtjohnson]

AMD used to make Pentium clones. Now, though, the AMD architecture is completely different from Intel's although they both will run the same software. The 64-bit AMD cpus seem to have fewer software faults when running Windows XP compared with the Intel P4s. This is an observation based on only a few systems and a LOT of things besides the cpu can affect that but I wonder if anyone else has noticed this (or maybe the opposite)? The comparisons between cpu architectures are always based on speed and benchmarks but not stability. Has anyone ever compared the different designs for how many GPFs they throw off, other things being equal? I was thinking maybe that's one of the reasons why the AMD systems are still selling so well, even though the new Intel Conroe is faster.

Re:Are AMD processors more stable than Intel?

[TheRaven64]

Has anyone ever compared the different designs for how many GPFs they throw off, other things being equal?

A General Protection Fault is not a fault in the sense of a hardware flaw. It means that a process has attempted to access a segment for which it does not have the correct rights. Assuming correctly written software, the most likely cause of these is memory errors. If you are suffering from a lot of them, then I suggest you try switching to ECC RAM (or, at the very least, run memtest86 on the worst offenders for a day or two). You can typically monitor the number of errors fixed by ECC memory in software, so this should give you some indication of how useful it is being.

If some machines are getting GPFs more than others, it may just be that their memory errors are overlapping the code segments more often; other machines may be silently corrupting data more often...

Re:Are AMD processors more stable than Intel?

[bky1701]

You maybe seeing something different then what you think. Since I have gone from using a 1 core Intel to a 2 core AMD I have had far fewer system-disabling crashes, because most of the time only one core is crashed, leaving me the other to successfully kill the process or fix the problem. However, being a known fact AMDs run less hot then Intels, I would not be surprised if there were real processing errors occurring on Intels.

Pricing

[Tancred]

When you can't make your product fast enough for all the demand, you're not charging enough. If you charge more, you can use that to increase manufacturing capacity. I'm sure someone at AMD understands that, so maybe they were caught off guard and are backfilling orders and have decided just to not reduce the price as early as they would have.

It's not poor grammar

[Anonymous Coward]

An introductory prepositional phrase, unless unusually long, does not require a comma.

there is one thing i think is true

[crashelite]

AMD is just like apple though they have a fan base that ppl who like AMD will use AMD. they will tend to spend the extra 10-100$ for a AMD chip because it is a solid product... i my self have never had a AMD processor fail on the other hand i have fried one out of every pentium line that i can think of from centrino to the P1 to the P4 i think i have fried em all (not the individual speed increases but just the marketing names)... it is really neat i have them all hanging on my wall... including motherboards that were fried in the process

Re:there is one thing i think is true

[TheRaven64]

My experience is the opposite. The only reasons I have bought AMD chips has been that they are cheap, from the K5 to the Athlon. I have had Athlons fry periodically, while Pentium systems have kept going. The main reason for this is that Intel typically shipped better heatsinks and fans as standard; AMD chips with a decent third-party cooling system seem to be stable. Of course, with the newer Athlons, there is built-in thermal protection, so bad cooling won't make them die.

And?

[homey of my owney]

> This design has given AMD great performance during the past few years, but resulted in processors that were almost twice the size of its single-core chips." And?

Merger hopes

[DrYak]

The AMD-ATI merger does not bode well for Linux users, given ATI's abysmal track record.

On the other hand, the open-source drivers are starting to get quite decent specially in latest Xorg/Mesa release (7.1/6.5.1). And given AMD's track of openness, maybe they'll help the community by trying to release as much as possible specs to promote development of a true useful (non binary-blob) solution. Let's hope and see whether the buyer's or the buyee's mentality will prevail.

but with Intel's quad core processor coming out Real Soon Now and with my board's already being certified to run it, it was the logical choice. Quad core upgradability was the clincher.

Yeah and given Intel's track record, the next thing you'll see is Intel issuing a completly new set of Northbridge / Socket, like they've done with every much anticipated processor generation (except for the PII/PIII family). So either you'll be stuck to the few first generation entry level quad cores and won't be able to upgrade, or you'll have to buy a new mobo to have acces to the whole Quad Core serie. (I suspect that, besides technical reason like shortening the path, one of the main reason why Intel abandonned the Slot 1 format was to make easier to force people to buy newer motherboards more often, instead of all this money going into the pockets of "slotcket" makers).

Supply problems? No kidding?

[jcr]

So much for all the AMD fans who were foaming at the mouth over Apple going with Intel.

-jcr

Stupid article

[Anonymous Coward]

What great reporting -- "2 cores take ~twice the silicon as a single core"... thanks, Einstein...

AMD hasn't had this problem before.

[Chris Burke]

Since then, anytime the get a good product, they blow it on production, leaving Intel to fill the void they created.

There's a difference between a production problem and a capacity problem, though they both result in supply failing to meet demand. A production problem is when you theoretically could manufacture enough parts to satisfy demand, but your process, technology, or chip design is flawed and is not reliable enough to meet production goals. This is especially bad when you tell your customers that they can expect so many chips, and then you fail to deliver them. A capacity problem is when you couldn't possibly produce enough chips

AMD has had plenty of production problems in the past, when their market share was low enough they've had the fab capacity to produce enough chips, but they were simply unable to. When this happens, you have to fix the design, fix the process, tweak the technology parameters, whatever to correct the problem. While this has hurt AMD quite a bit at various times, they did recover, and it didn't fundamentally limit their marketshare.

Now, AMD has reached a point where they simply have a capacity problem. Their production process and yields are very good, but their market share is such that Fab 30 is physically incapable of producing enough chips to meet demand, even if it were working perfectly and every chip on every wafer was good. Their market share is now capped. When this happens, there's only one thing you can do: Build more capacity.

AMD hasn't had this kind of problem before. They would have loved to have this problem instead of the production problems they've had in the past.

It is where they have failed again and again and again. I can't believe they haven't learned yet.

Oh, but they have. In particular, they anticipated this problem years ago and started to build Fab 36, which is about to come on line right when their capacity issues are becoming a serious limiter to their growth. Fab 36 is going to mean a lot more capacity for AMD, and let them grow their marketshare by a lot. Now they may have production problems with the new fab, but once they get those worked out they won't be having any capacity problems for a while -- and if they do, they will giddily ply investors to help them build a new fab.

P.S. speaking of investors, the point kfg was trying to make is that a capacity problem can only be solved by increasing capacity, which requires investment in a new fab. If investors fail to invest in the extra capacity, and the company hence cannot meet demand and fails, then it is the investors who are responsible. Somewhat; after all it is the company's job to seek out investors to help them build a new fab.

Re:AMD hasn't had this problem before.

[Sebastopol]

I understand the risks of going public, but I also look at Intel dumping billions of profit into fabs.

I just can't sit back and say, "It's the investor's fault AMD is hurting." AMD has been public for a long, long time. If the investors are such a risk maybe the should do some cost cutting and fund new fab capacity with profit rather than speculative public investor funds. If you want to say that investor's caused this, then you are saying AMD "plays the market" by putting important eggs in the public's basket. Maybe AMD shouldn't be public anymore.

Re:AMD hasn't had this problem before.

[ocbwilg]

Oh, but they have. In particular, they anticipated this problem years ago and started to build Fab 36, which is about to come on line right when their capacity issues are becoming a serious limiter to their growth. Fab 36 is going to mean a lot more capacity for AMD, and let them grow their marketshare by a lot. Now they may have production problems with the new fab, but once they get those worked out they won't be having any capacity problems for a while -- and if they do, they will giddily ply investors to help them build a new fab.

The other key to remember is that AMD has also contracted with third-party fabs to make their CPUs. I believe that Chartered Semi is the one. At any rate, once those wafers start producing chips (it takes many weeks to go from raw materials to wafers to CPUs) then supply will increase further.

Re:AMD hasn't had this problem before.

[Chris Burke]

I understand the risks of going public, but I also look at Intel dumping billions of profit into fabs.

Yeah, Intel has unparalleled fab tech and fab capacity. I assure you they aren't paying for all their fabs in cash, though.

I just can't sit back and say, "It's the investor's fault AMD is hurting." AMD has been public for a long, long time. If the investors are such a risk maybe the should do some cost cutting and fund new fab capacity with profit rather than speculative public investor funds.

A new fab can easily cost $5 billion and takes years to build. You don't just "do some cost cutting" and find yourself with $5 billion in extra profits, at least not in time to build the fab so it is finished when you need it, or without cutting the R&D that you need so that you have good products that will sell well enough to need the extra capacity.

Also, my point is that AMD isn't really hurting. Their market share is capped by fab capacity at the moment, yes, but Fab 36 is nearly online because AMD went to investors to get the money to start building the new fab several years ago. They predicted back then that they would need the extra capacity now, and while it'd be nice if Fab 36 was producing wafers now instead of next year, they actually did a good job of anticipating the need and then executing on the building plan so that the window in which they could use more capacity but don't have it is relatively short. Besides, to make use of Fab 36 now it'd have to be tooled for 90nm then re-tooled for 65nm which would take a long time and a lot more money than just targetting Fab 36 for 65nm.

If you want to say that investor's caused this, then you are saying AMD "plays the market" by putting important eggs in the public's basket. Maybe AMD shouldn't be public anymore.

I assure you Intel does the same thing, and in fact any similar manufacturing company does too. And the point was that investors -- which doesn't necessarily mean holders of publicly traded stock, btw -- are responsible for a company not having enough capital to grow as they need to, because providing a company with capital needed to grow is what an investor is for. This is not what has happened to AMD, because they did get the investment money to build a new fab, so if I say that investors caused this as in AMD's current situation, that would be praise not blame.

yep

[YesIAmAScript]

Think of apple with IBM/Motorola for PowerPC.

IBM/Moto would make a new chip, Apple would announce it and IBM/Motorola wouldn't be able to make enough. Once they begin catching up, the product would be near end-of-lifed already.

With Intel, Apple's sales are a drop in the bucket. No more first-day shortages for Apple.

Notorious A.M.D.

[Dorceon]

For AMD, Mo' Money, Mo' Problems.

Nikon

[wwwillem]

It's a while back I worked actively in semiconductor. But at that time it was Nikon, Canon and ASML who were the biggest wafer stepper manufacturers. Given the investments needed, I would be surprised if that has changed a lot.

That would explain the need for...

[fellip_nectar]

...this technology.