This article really ------ my ------! I'll never -------------------------------------------

---------------------------- the same way again!

Yeah, I'll -------- my --------. It's the one with ------------ on its ---------------.

This looks very simple. If AMD was _____ by Intel while ____________ and ____ then it is clear that ________________ was the cause of it all and __________ should be severely _____.

____ should therefore be _____ to a sentence of ______ and fined _____.

An open and shut case I say.

OK - how long do we think this case will take to reach it's final conclusion? Two metrics please:

1) Duration in months

2) Legal fees (sum of both parties costs) in $US

I think El Reg should offer a prize to the person with the best estimate in each category...

Intel really has a great deal to thank AMD for. Had AMD not taken at least some notional lead in PCs then the giant that is intel might never have cottoned on to its oversights.

AMD worked, the giant awoke and now AMD has a bit of a struggle.

The dynamics of at least two organisations working constructively independently and competitively is reward enough.

How much will the lawsuit cost?

Does it amount to similar figures mentioned in the court case?

Could those resources be better spent as part of an out-of-court agreement?

There's a lot of very funny stuff on the register, but very few things actually make me burst out laughing - the lego pictures, about one in ten BOFHs, but that third image down ended up with me dropping my laptop I laughed so hard :)

Intel is just out to royally [blank] Intel. Obivously [blank] with its [blank] [blank] and [blank] [blankety] [blank] because [blank] banana [blank] Paris [blank] Hilton.

They should just [blank] and [blank] themselves.

...[blank] that [blank], boy! We drink Blue Riband!"

I love the subbing here :)

I predict around ___ months and $______________________________US, which works out to roughly £__ in real money

Tux because ______________________ on _______ up his ________________ with paris hiltons _________

AMD and Intel _________________________________________________ walrus _________________________________________ black cats ____________________________ and the priest and the lemon left hand in hand.

No contest, Chipzilla will walk away the winner. Bigger corps can hire more legal guns depite that fact that it was open knowledge for quite some time that Chipzilla was playing fast and loose.

Hearts with AMD, heads with Intel.

:(

AMD -------- --- ---------- --- Intel --------- - ------- --- ---------!

So heres what I was thinking if Intel was really _____________________________________________________________________________________and if AMD was really_____________________________________________________________________________________________then I guess what is at stake here is________________________________________________________________________________________________________

Heh Yep I guess that looks awful close to some of those briefs.

As for the subbing common people lets see who can get the most wasted stapce on Reg today :)

/mines the___________________________________

AMD could be in trouble, unless Intel hires the same bad lawyers that Microsoft did.

----|

| O

| --|--

| /

|______

Intel ____ ___, AMD _____

P please bob

Could it be that HP turned down free chips because they thought it would harm their image, and that it would impact the sales of other machines? If HP has lots of Intel chips on hand that they paid for, they might want to use those up before using the AMD Chips. If HP thought that by the time their supply was up of Intel chips the AMD chips would be obsolete, they might not choose to go with them.

Design considerations might have also played a part in this. If a machine is designed with a certain custom motherboard, the retooling of making a new motherboard or new case to work with the chip might be too expensive. The same with testing software with the new AMD chips to make sure that it works (HP diagnostics and other tools that work specifically with specific chipsets. The new chipsets might require new programming or the rewriting of the utilities on hand)

I think that Intel has shown predatory practices, but they are not at the same level as Microsoft. AMD realizes that they are slowly loosing the footing they had two years ago, and think that pursuing legal measures will either slow Intel down, or be a better investment then getting their new chips to market.

Uh, ben... They took 160,000 freebies, which is much higher than needed for R&D, so they did obviously sell some of the freebies.

If the chips are free, surely the cost of the product is lower, so they should be able to sell more while retaining the same profit margin... So they should have been able to easily move at least half a million boxes.

I hope AMD get something done here... If it's a big waste of money and time, it's going to be really bad for them.

I looked at the page's source, as I was convinced Mr. Modine would have hidden a joke in the [blank]-[blank] subtitle (see all the PDF-censoring cases, where an existing PDF document gets an added layer of blackness; this works fine in printing, but re-editing in Acrobat lets you uncensor).

That said, maybe was little hidden: when asking for their diagnosis of my motorbike's problem, a technician helpfully explained that "the f*cking f*cker is f*cking f*cked!". Note that with or without censoring, the same information is conveyed...

Intel ---------------------------------------------------------------------------------------------------------------------------------------------------- with a horse and three eggs ----------------------------------------------------------------------------------------------------------------------------- Paris Hilton -----------------------------.

The old way of doing this with black marker pens usually left the text readable with an IR camera. You needed an original and not a photocopy though.

Just thought people might like to know.

... Of course he was very, very drunk.

Is it just me, or does AMD offering HP a million free chips seem a bit contrary to their argument against intel?

More to the point, wouldn't you have thought that the embarrassment of not even having a 25% take up by HP would be something to keep schtum about?

I think it unfortunate for AMD that they have levelled accusations about hidden microcode. If this practice is embarked upon then I would expect other covert microcode features to be embedded in the dies Chipzilla produces. Let’s suppose that national security depends on these features to allow cXXXXXX and XXXXXXXXXing then fighting this will be difficult and would undermine AMD’s position when the evidence is not corroborated by the opposition (because their position is protected by higher authorities). Moreover, if the authorities wish to push these features in to mega-corps across the globe what better than by using high-end products blue-chips would tend to rely upon or are encouraged to purchase through other corporate tie-ups! Then, o’course, Chipzilla might use the same for their own nefarious purposes... ;-)

I think the funniest aspect of this story is the image of a whole roomful of El Reg hacks drooling in anticipation of their first look at the various submissions only to discover that the black marker pen got there first.

HP turned down a shed load of processors that were being given away, does seem odd that they thought it was fine to use some and not the rest of them.

Most folks would have taken the lot and found a use for them later.

Money for nothing -- and your chips for free

Coat - got

... that highlighting the blanked out words doesn't reveal it in all its glory then?

;)

I may be missing something here, but are AMD saying that:

"We know Intel are behaving anti-competitively because we couldn't give our processors away for free."

Isn't giving your product away for free a tad anti-competitive?

It seems to me that AMD has little to lose here. Even if they are unsuccessful in being awarded damages, the court case and publicity means that the OEMs have to be much more careful and transparent in their transactions with suppliers, and the discovery process gives AMD an insight into how Intel does business, and with whom. In addition, if individuals at the OEMs are singled out in the brief, doesn't it mean that these individuals may be under scrutiny and as such be forced to work more openly so that AMD will get a fairer chance with these OEMs in future? Any damages or even further discovery opportunities (such as the judge taking a closer look at the compiler) are all bonuses.

Black helicopters, because after this case, Intel we'll be watching you..

I dunno, ask all the Linux Distros.....

Apparently -

1. ???????

2. ???????

3. ???????

4. Profit!

Paris, becase she can ???? my ???? any day of the week!

Why AMD was over a year late with its quad core server chipset?

It appears to me that AMD's market share grew well when they brought out Opteron - because it was a strong product and superior price-performance to the Intel chips at the time.

I may sound idealistic, but if AMD comes out with superior chips once again, the market will come to them, regardless.

AMD will let me have the free chips HP didn't use so I can knock up some cheap boxes to flog on the eBay tat Bazzar?

One good reason for refusing a gift of free processors is if the OEM manufacturer (HP) thought that the processors are junk.

Another good reason is if HP didn't expect to sell one million units of the product that incorporates that processor.

It seems to me that this particular argument in AMD's case has no merit. Maybe it is in the briefs simply to add bulk to the filing.

Oh those silly sales people. Going out and trying to strong arm sales for their company. Telling vendors that if they sell any of the competitions' kit they won't be their friend any more.

'Course that's about the same as giving away a million of your product. Different sales techniques for different folks I reckon. :)

Cheers,

John

It is sad to see that two giants in the microprocessor design and fabrication business are subject to the same bloody annoying "black band" problem that plagues so many photocopying machines even in this day and age.

Something Must Be Done.

...I think someone's been watching too many 'X Files' episodes again...

Didn't Mouldy fall foul of this when he found the Majestic 12 documents?

(Or was that what the CIA, oops sorry [i]some unknown, overzealous office clerk[/i] did to the "M12" docs in Real Life... together with most of the files from a certain US air base down Mexico way.)

"Gort, Klaatu ********** **** *** **** ******* *** ******* *"

The ones that I used for looking through people's clothes !! I knew they'll come in handy one day !!

The interesting thing is that both had problems with clean rooms at the same time (going down to that spec. requires a first of clean), which was the main reason for the delay on the quad cores. If I remember correctly AMD had to (temporarily, I think) shut down one factory because of an A/C problem.

As far as the xxxx blanks...just think of an executive's mind every time you see one.

--Glenn

Kinda seems like this case comes down to a couple of things.

a) Intel has deals with customers that reward volume purchasing including clauses that offer significant discounts where the volume of sales reaches the degree of an exclusive supply deal. AMD doesn't like this and fails to see the irony in them giving away a million processors for free, presumably as an incentive? Sounds like a potential multi-million dollar discount on a contract to me. Which seems to be what Intel were offering. Perhaps HP didn't like selling chips that had terrible problems, or perhaps the product line incorporating them simply didn't sell. Who knows? Who cares. It seems that AMD is pissed off be cause they failed to successfully use the same sales techniques that Intel succeeds with.

b) AMD claims that Intel's compiler doesn't work well with their chips. Short of there literally being functionality inserted in an executable compiled by the compiler that looks at the CPU and borks the program if it's an AMD, they haven't got a leg to stand on. If Intel makes a compiler for x86 processors, they are naturally going to make it work best on their own chips. I would go so far as to say they will make no effort to optimize for AMD processors, nor would they have much interest in fixing any issues in the compiler that relate specifically to AMD chips, especially if the issues relate to errata in the AMD processor itself. So AMD complaining that an Intel compiler doesn't work that well for their processors seems rather an empty complaint unless they can point to some kind of legally binding contract or court order that forces Intel to make a compiler that is as good on AMD as it is on Intel chips. Frankly that's ludicrous and as long as I have worked with PCs and AMD has been trying to take on Intel there have always been instances where compiler optimizations favored one brand or the other, just as certain bench marks favor one or the other.

AMD fails as far as I am concerned. Their arguments seem weak, and their motivation appears to come from the simple fact that their own poor decision making in the past few years has left them trailing Intel significantly. Legal action is a poor substitute for good products and good business decisions.

Given the number of processors shifted by these orgs 1 million isn't all that many. Also, bare in mind that AMD isn't the larger of the 2 companies.

Giving a product away for free to say " Look over here guys!" is perfectly normal and decent practice.

Giving a product away for free to make sure people continue to by your product and not another (typically new-comers) product is with out a doubt anti-competitive (MS tactics).

Telling a company that they will get a discount if they buy greater than X of your products is perfectly normal and decent practice.

Telling a company that they can only get the discount if they use only the one company's products is, again, without a doubt anti-competitive (Intel tactics).

Remeber - buying only from one company doesn't mean you will be buying more of their chips. Two very different beasties.

18 months from now, $1.2b AMD wins.... (I have been known to be wrong) And, I'm not saying that AMD's troubles are purely Intel's fault. I just reckon this is how it'll go.

How much is AMD asking for?

Hi, I have been a computer reseller for over 10 years. I have witnessed alot of Intel's antitrust behavior. I deal with most of the major computer distributors, and when AMD producted the first 64 bit server processor, I was shocked some distributors didn't sell AMD at all! I was told by their sales force that we have an exclusive contract with Intel for better prices so we wouldn't sell AMD or we will pay more for Intel processors! This distributor is owned by a Far East company that also owns 1 of the major server motherboard companies! Then you have all the Far East companies that wouldn't sell AMD laptops at all. But this all changed when AMD filed the antitrust suit. The distributors all started selling AMD processors, the OEM started making AMD computers to cover theirselves.

Since news of this lawsuit I have stop actively selling Intel processors, and promote only AMD processors. I encourage other resellers to do the same, boycott Intel! I recently learned that the Phenom quad core is faster on any Unreal 3 engine game, beats the fast Intel cpu! Also wasn't late with their discrete quad core cpu, its the first of its kind! Intel has nothing on the market like it! Intel only has a dual dual core calling it quad, just two duo tied together on FSB! Rumor has it Intel will produce a discrete hex core cpu but AMD is likely planning a 8 core cpu next!!!

Sure if you have a contract the demands exclusivity like that. However if your contract simply promises certain discount levels based on your volume and your annual sales are say 50 million units and you can achieve maximum discount by taking a volume of at least 49 million, that is a perfectly fair volume discount program, isn't it? If Intel knows what a customer's likely volume is in the coming year, then they can structure their volume discounts to be most attractive at the point where the business is effectively exclusive. I don't see how that can be a problem, it's simply using the volume discount to reward high volume customers. Giving away a million processors would seem to be something of a gift, perhaps even an incentive. Possibly even anti-competitive.

>"a) Intel has deals with customers that reward volume purchasing including clauses that offer significant discounts where the volume of sales reaches the degree of an exclusive supply deal."

Did you actually RTFA? They're explicitly accusing Intel of offering a deal to Sony with an exclusivity clause in it - an "all-or-nothing 'conditional' rebate". Why did you write "reaches the degree of" an exclusive supply deal? That's a pretty strange way to phrase anything, unless you want to conceal that you're deliberately distorting the truth. It's not about them just selling more to the extent of 100% vs 0%, there's a whole world of difference between that and actually demanding the customer sign up to a contractual obligation not to buy from the competitors. You seem to be trying to hide that difference by your unusual and twisted wording, when it's the very epitome of illegal anticompetitive behaviour.

>"b) AMD claims that Intel's compiler doesn't work well with their chips. Short of there literally being functionality inserted in an executable compiled by the compiler that looks at the CPU and borks the program if it's an AMD, they haven't got a leg to stand on."

Did you actually RTFA? That's exactly what AMD are claiming.

Could you have misrepresented what the argument is about any more if you tried? I doubt it. Do you want to declare a financial or career interest in this matter, by any chance? It's inconceivable you could have misread the article by a hundred and eighty degrees like that, so you must have some motive.

In the first case, I think he's simply pointing out that a perfectly reasonable volume discount plan could look like an anti-competitive exclusivity agreement. In that case, AMD would be making the accusation on reasonable grounds, but there would turn out to be no substance to it.

In the second case, I think he's simply expressing doubt that Intel would go so far as to insert such nerfing-functionality into their compilers. I'm less sanguine about that one, but it should be easy to test. AMD must have copies of the compiler floating around - pick one and reverse-engineer it. If the bad functions are there, they'll be found.

Sounds like you are the one with a financial interest in the case. Did you bet the farm on Intel losing? I have neither financial, nor career interest in either company.

Regarding discounts and volume contracts, Diane got the point, but you didn't. I wrote 'reaches the degree of exclusivity' because if I were framing a contract and wanted an exclusive deal but didn't want to spell that out in the contract (because I have smart lawyers and what not). Then I would build the contract to apply an escalating discount based on the likely 100% volume where the maximum discount was significantly better than the next best and therefore encouraged a volume purchase that was functionally identical to an exclusive deal without the necessity of a contract stipulating exclusivity. Which is what I'm suspecting Intel tried to do with it's contracts. AMD for it's part appears to want to have the court see it's own attempts at buying business as simply another example of Intel's manipulation because HP didn't take their 'free' chips. Frankly 1 million AMD processors is a pretty big chunk of change to hand out for free, so I fail to see how AMD can bitch the big one about Intel when they're engaging in the same aggressive business tactics, but hey, you obviously have your side picked out on this one. And I'm certain that you think you know which side I support - as it happens I could care less about either of them

Regarding the compiler functionality, I'm not sure that it would matter if Intel had written a function into the compiler to make executables flip on their back with their legs in the air on an AMD chip. Intel writes a compiler for it's product, not that of someone else. the fact that AMD have an x86 chip of their own design does not mean that an Intel compiler will produce code that works well (or at all ) on it. Unless the AMD chip is functionally identical to the Intel chip(s) that the compiler is targeted at there can be no expectation that it would work unless it's explicitly stated that it will work.

Well, yes, Intel DO make their compiler for their chips. However, if they are going to nerf AMD (or Cyrix/Via/...) performance, then don't sell the compiler as x86 compatible. In fact, you'd best SAY it's only for Intel machines and don't expect to get paid for a copy when a software house wants to compile software to sell on "x86 compatible" systems.

Well why waste the time getting 160,000? Surely just a handful would tell you they're crap.

Any competent developer would know almost immediately if a compiler was nerfing apps running on AMD chips specifically. It would show up immediately during any level of compatibility tests. This is ludicrous, if the compiler intentionally kills or breaks apps on AMD chips, a simple debug will show immediately where and why the app fails. It's not that easy to hide that kind of thing. The developer community is not so stupid that a bad compiler from Intel would continue to be used if it intentionally borks apps on AMD chips. This is easily the most stupid allegation, and the one that should be the easiest to prove. If all that AMD can do on this one is bleat and whine, then they should shut up now. If they have evidence then it's literally the IT equivalent of child's play for them to show it, and it would be almost as easy for any competent IT journal to do an expose showing the compiler in question and exactly how it farts in AMDs face.

I really wish people would think this kind of thing through before posting.

>"Sounds like you are the one with a financial interest in the case. Did you bet the farm on Intel losing? "

Nope, that's a false inference. Just because I don't immediately absolve Intel does not mean I have not said that I must therefore believe AMD's case.

>"I have neither financial, nor career interest in either company."

Absolutely likewise. That's why I don't choose to immediately side with /either/ one as a matter of blind faith. You /did/, and that's what made me wonder. In the spirit of fair debate, I will completely accept your declaration of disinterest, because that's actually the core of my point: why, if disinterested, are you so strongly (what seems to me) biased in your reading of the same set of basic information?

>"Regarding discounts and volume contracts, Diane got the point, but you didn't."

Well, that remains to be seen!

>" I wrote 'reaches the degree of exclusivity' because if I were framing a contract and wanted an exclusive deal but didn't want to spell that out in the contract (because I have smart lawyers and what not). Then I would build the contract [...] "

Yes, thank you, I'm not actually daft and I did understand perfectly that that was what you are suggesting had happened.

The point that /I/ made and that /you/ missed, however, is that that is *not* what they are being accused of by AMD. AMD are claiming that they *did* spell it out in the contract. That, I trust you would agree, is of *course* illegal anti-competitive behaviour if it *is* what happened, no? Well, /that/ is what I understand the story to be telling us that AMD alleges - specifically in the Sony case. I don't see anything in the story that suggests that AMD are claiming Intel did what you describe.

Therefore I suggest that there is a case to be heard here, that AMD must have their day in court and show *evidence* of what they claim, that Intel must *defend* against those claims, and that none of us can pretend to know from such a hugely redacted report what the strengths or weaknesses of the arguments are on either side until we get to see such evidence and defence. And I am surpised that you should produce this hypothesis based on nothing we can read in the story that it was a mere innocent sales tactic on Intel's part.

>"Regarding the compiler functionality, I'm not sure that it would matter if Intel had written a function into the compiler to make executables flip on their back with their legs in the air on an AMD chip."

Well, you'll have to forgive my misunderstanding there, because that's exactly what you said /would/ matter in your first post, and then came up with another unfounded-on-any-information-in-the-article-we-were-all-reading hypothesis why it wasn't the case anyway.

>"Intel writes a compiler for it's product, not that of someone else. the fact that AMD have an x86 chip of their own design does not mean that an Intel compiler will produce code that works well (or at all ) on it. Unless the AMD chip is functionally identical to the Intel chip(s) that the compiler is targeted at there can be no expectation that it would work unless it's explicitly stated that it will work."

There's a world of difference between the compiler not working for AMD because it, say, chooses the opcodes that work best on Intel CPUs which happens to be a deleterious choice for AMD CPUs, and actually having a segment of "if (cpuid == amd) then crash" code deliberately inserted into the compiler, and again, you're ignoring what the article actually tells us that AMD are claiming. They are claiming deliberate sabotage. It is, after all, anti-competitive to needlessly obstruct competitors from making compatible products. Again, there is a case to be heard here, and if what AMD claim is so, they should be able to produce some fairly convincing evidence in the form of disassembled code.

Not having any personal bias nor external sources of information to go on in this matter, I choose to wait and see the evidence.

Your immediate response was to immediately start grasping at unfounded hypotheses that absolve one side of guilt.

There not being any evidence for you to have chosen which side to absolve, why did you choose to absolve Intel?

I do not choose to absolve anyone of anything. I am however mindful of a few things. AMD is in competition with Intel. Their allegations of Intel wrong doing bear examination with a critical eye as they are hardly impartial. I simply chose to put forward a few rather obvious points that nullify most of the whining that AMD is indulging in. Frankly had AMD spent more time and money on R&D and less on questionable ventures such as acquiring ATI, then their designs would be more competitive with Intel's. Had they spent more time and money on materials science with their partners their manufacturing would be more mature. However, they did not, they rested on their Opteron laurels and the result is their impending failure in the market. AMD had several key markets in the palm of their hand, and frittered them away. They had a major deal with Dell that resulted in them screwing the retail channel and enthusiast market. AMD's own poor decisions are not the fault of Intel. Very little of what AMD accuse Intel of doing is anything that AMD would not do/has not done themselves. This case has the look and feel of a huge case of sour grapes. The compiler allegations are particularly stupid to me. it would be so easily proven beyond a shadow of a doubt, and yet where is the analysis in the Industry press to confirm it? Where are the headlines about it? The only time we hear about it is when AMD wants us too. That stinks. If Intel's compiler is crap on AMD is anyone actually surprised? Besides that, does anyone care? Is Intel the sole supplier of x86 compiler technology? Of course not. Doesn't AMD have it's own specially optimized compiler technology that makes it's chips look super ace? Yep, they do. So, does it matter if Intel nerfed their own compiler to make it work better for their chips? Not really. Oh, but let's all wave our hands over our heads like it's the end of the world instead of looking at the accusation critically and trying to figure out why AMD is even bothering with the allegation.

I would respectfully point you to the f00f bug.

Bad Via chipsets causing other problems.

Bad nVidia chipsets.

Bad Intel chipsets.

Running Windows.

With these (and many, many more) issues all colluding to cause your program to fail, it can be very difficult to find out.

Also, "nerf" doesn't mean "break", but can mean "use the slowest method".

So what? Well, AMD say they have proof. The proof that you say any programmer can find. Well, did you look? AMD did and they say they've found it.

So do you have any reason to believe AMD are lying?

Do you think that when it gets to court, if AMD has lied, that the court will accept it and let it ride?

>"I do not choose to absolve anyone of anything. I am however mindful of a few things. AMD is in competition with Intel."

Absolutely. I don't condone unsceptical belief in either side's case.

>"Their allegations of Intel wrong doing bear examination with a critical eye as they are hardly impartial."

Well of course, but we don't care if they're impartial or not, we care if they're /true/ or not. This is a simple matter of fact that can (and hopefully soon /will/) be solved by seeing the actual original contract they signed with Sony, or the disassembly of the generated code that AMD claims their compiler inserts. AMD will either provide this evidence and be right, or fail and be wrong. *We* have no privileged insight into how the future will go.

>"I simply chose to put forward a few rather obvious points that nullify most of the whining that AMD is indulging in."

Ah, but that's where your reasoning fails you. It goes like this:

AMD accuse Intel of doing X and Y.

You say "But there's nothing wrong with Intel doing A and B".

Do you not see the disjunct between your response and the allegations your response is supposed to be addressing? If you don't defend Intel against X and Y, you aren't nullifying AMD's accusation of X and Y; if you defend doing A and B, that has nothing to do with whether X and Y are legitimate marketing or illegal anti-competitive behaviour.

Looking at the hardware reviews for several years

it appears that the pseudo benchmarks used

by most reviews are adjusted to favor intel over AMD.

The "raw" or basic benchmarks like Everest

give AMD a much better rating or rank.

For example an AMD Socket 939 single core (later version)

ranks with and above intel duo-cores in raw benchmarks

like memory and processor bandwidth and basic Everest like units.

This is because the memory is connected directly to the CPU

and is "dual channel" and operates near the ideal 6.4 MB/sec.

AMD was "too good" and "too smart" so the mob had to put it down.