Antitrust Lawsuit discussion

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
I started this thread because I'd like to weigh in and respond, but the mods are correct that we should take it to a new thread...

Stated so far...
Nemesis
I think I will completely disagree with this notion. Rector brought a law suite against Intel . Thats what opened up the OEM market to Rector ++++ the fact that K8 was that good. The trueth is AMD didn't have the resources to supply the oems When this law suite is over I believe the courts will come to same conclusion . Intel will be able to prove this with recent developments.
Intel's rebates may have been based on certain conditions . But in the end Intel will beable to prove AMD simply did not have the capicity to supply the OEMS with guartenteed deliveries. It is that simple. Why do so many places sell either Coke or Pepsi but not both . Now that is very very odd. It is also ilegal by AMDs standards


Viditor
Then we shall have to agree to disagree...
Though I will add that most all of the OEM contracts were in place well before the lawsuit (Sun, HP, and IBM all had OEM contracts before the lawsuit).
As for capacity,
1. Fab 30 had a capacity of 5500 WSPW (Wafer Starts Per Week)
2. Even the largest K7 on .18u was only 129mm2
3. On a 200mm wafer, that's just over 200 candidate dice per wafer or 1.1 Million candidate chips/week
4. If we assume even an 80% yield (and most reports at the time wer well over that after ramping), that's 880000 chips/week or 45.76 Million chips/year.
5. I believe that the world market sales in those days was close to 120-130 Million chips/year, so they had the capacity to deliver 35% with Fab 30 alone. (though the actual amount was higher as the average die size was MUCH smaller)
Also, remember that Intel has voluntarily shut down the rebate program (though they can volunteer to start it again) since the lawsuit, so whatever has happened since that time carries far less weight.
I personally think that Intel doesn't have a hope in Hell of actually winning the lawsuit (especially after they lost or destroyed the e-mail evidence they were ordered to retain by the courts)...the only 2 scenarios that make sense to me are
1. Intel settles at the last moment, thereby sealing a lot of the records that will be displayed in court and avoiding a very high potential award
2. Intel tries to appeal for as many years as they can in hopes that AMD will settle out of frustration at a later date.
My bias is towards option 1 as it seems to me to make the most sense for Intel and will most likely be the far cheaper option. JMHO


Coldpower27
Well in Coke and Pepsi's case I believe they have very comparable marketshare so yeah exclusivity in some situations isn't anti-competitive because the impact isn't threatening when both are of comparable marketshare.

http://www.nytimes.com/ref/bus...0527_COKE_GRAPHIC.html

Nemesis 1
I see your point but its also removing my right to get the drink i want at the food place i preferr.
What happened To Digital equipment . Didn't IBM squezee them out of the market.
I suppose that the IBM pc using intel x86 squezed them out. Why isn't ms a monopoly by law . Shouldn't it be broken up like AT&T was.
To many differant standards going on here . Its USA for crying out load.
How do we know mike dell didn't approach intel and say if we sell Intel exclusively and we get a cut on cpu cost we can under sell everone in the market.
I don't see why Dell and Intel couldn't agree to such a deal . After all its Dells choice . Did dell ever sell AMD cpu's till recently?
I just don't see what the problem is.
Perfect example is Apple Intel now . Intel exclusive for Apple it was Apples choice and apple is grabing market share now with intel.
apple is also getting 1st products from intel is that illegal ? Dell is know longer intels cpu's first to market vender . when c2d first appeared it was apple that got first shipments from intel . if i was pc maker i would go exclusive intel in the hopes of getting first products ahead of dell hp. first to market is no small deal. so i think amd is just cring like a baby .


coldpower27

Really? You don't really have a right to buy whatever drink you want at a restaurant, you have a right to buy whatever they carry. There is a no one pointing a gun to your head and saying you must go to that restaurant. If you want that drink badly enough, you will simply go to another restaurant. They are also potentially losing some customers due to the reduction in choices, so it's not all advantages for a restaurant to only carry 1 drink.
In Apples case it's fine that particular vendor has a small enough marketshare that exclusivity with one manufacturer isn't going to cause enough damage worth caring about.
Since now that Dell is both AMD/Intel as well as most other Major OEM such as Gateway and HP, AMD has most of the largescale OEM carrying it's products, so it's up to AMD getting out good products now.
With Intel and AMD, it can be classified as critical damage if Dell doesn't carry AMD because they have significant OEM marketshare, simply because Intel's marketshare is at about a 4:1 ratio with AMD's give or take and not a 1:1.
At any rate AMD doesn't have any more excuses with major OEM now as they all carry AMD, as well, besides Apple that has so little marketshare that they aren't worth mentioning.
Microsoft is a monopoly are you talking about the Operating System sector right?, Well there are some alternatives now, like Mac OS for x86 and Linux, so I wouldn't complete agree.
Though I do agree that AMD loves to use the anti-competitive flag a little too often


Nemesis 1
So your saying that its allright for apple because of small but rapidly growing market share to be exclusive to Intel . So when penryn is released and Apple gets the first shipped processors that its OK . But what if Apples market share keeps growing at present rate when will it not be alright? With VM it is possiable for Apple to grow by intel exclusitivity. Just like Intel grew Dell because of its exclusivtivity. Dell was first to recieve Intel new products . This isn't a small matter.
If I owned HP and I decided to go Intel exclusive. Your saying that is wrong. Iwill never understand that kind of thinking not as an American.
Its my company so I can put in the box what ever I want.
If Intel gave me a better deal because I was exclusive your saying thats wrong . I say its OK and very smart. When AMD had the performance lead that could hurt my sales. But isn't it really my choice. Intel wouldn't care if Dell went exclusive AMD I doubt intel would care. But Dell would go bankrupt without Intel . Thats A fact


Stoneburner
Huh? As an american you do what the gubmint says!
IN all seriousness, there are reasons for such regulations so falling back on simplistic libertarian notions isn't helpful here. It's a point alot of people misunderstand about AMD fanboys, alot of them arent' fanboys they just want some competition

coldpower27
Apple, isn't big enough to be a threat yet in the x86 world, when they actually become a threat, I doubt it will ever occur, since they cater to the more wealthy of the market. Your situation is only hypothetical, we will deal with that when we come to it.
Like I said already, with something as large HP/Dell, it is enough to cause significant damage. This is not acceptable to AMD's continued existence. Not so for the moment with Apple.
Not so when your actions, will cause the destruction of a company that was preventing a monopoly. AMD was first introduced as a second source so that if the super unlikely event Intel died we would have second source for x86 processors, AMD's continued existence, is of paramount importance. If what Intel does endangers that, that is not acceptable


Nemesis 1
So because you believe apple is highend use only or high $$$. You believe that they have limited market penatration. OK thats fine. But avoiding the hypothetical based on that is just dancing around the question I asked
It is unacceptable to who? AMD! Surely their is NO law saying they have to use AMD products
So you are saying that if Intel came out with a product so good and so cheap that Intel by law couldn't produce it . Come on!
I really laughed at this viditor. Who were the E-Mails sent to and who sent these E-Mails? Shouldn't they have these e-mails also? Why say Intel lost destroyed these e-mails if you place a burden on Intel for losing said e-mails shouldn't you place said burden on sender reciever also? They should have copies of these e-mails

Stoneburner
Nemesis, destruction of evidence can give rise to an inference under rules of evidence (at least in california) It's not a question of what side has it

Viditor
Ummm...I'm surprised you haven't heard about this. The e-mails were from Intel's top executives and there were over 1000 missing!
The whole point is that you can't get copies from the recipient if you don't know who it was.
And besides, the recipients weren't under court order to preserve them!

Original Story

Nemesis 1
I have heard of it . Do you have a link about which e-mails are missing. If their is no record how do we know there missing. Are these e-mails before the court order or after the court order.
If they know theres 1000 missing e-mails than someone has copies of said e-mails or they wouldn't know about them .
We really have to wait on this one to find trueth.
Why aren't you and others upset about the already settled case were amd reverse engineered intel cpu's and basicly stoled x-86 . I know IBM connection here . Doesn't change the fact amd stole intels tech and used it as its own


coldpower27
I don't "believe" Apple is high end, they are a high end vendor, Apple's cheapest IMAC is 1199 USD/1299 CND and 1099USD/1249 CND for MacBooks. I don't answer questions that have no relevance to what is current occurring, they are a waste of my time. As long as Apples continues with the style they have shown so far, they aren't going to continue to sustain their growth, as the upper mainstream to enthusiast just doesn't command that much marketshare.
If the decision is made without outside competitor influence then it is fine, if however the larger company is using it's influence due to sheer size, then no it's is not permissible.
Nope that is not what I am saying at all. The thing is Intel won't do that, as they know having AMD around on critical life support, is a better situation, compared to the alternative which is no AMD at all


lopri
AMD's ability of supplying enough chips only matters when Intel is involved in the equation. Without a threat from Intel, an OEM could sell both AMD/Intel CPU equipped systems per market needs (for example, Pentium M for laptops and A64 X2 for desktops). What OEMs feared was that Intel would completely cut off the chip supply on every front. Only in that situation AMD's capacity (and their inferior mobile chips) would become an issue, because then an OEM will have to rely on AMD for its complete line-ups.
I had (and have) no intention to join the argument in this thread but I wanted to comment on this common misunderstanding. 'AMD didn't have enough capacity to supply chips' can make sense only when considering a certain effect that Intel could have on an OEM.


golem
If AMD's ability to supply an adequate number of chips for an OEM's needs is based on Intel's sales policies, then I don't see how Nemesis 1's original comment is wrong

coldpower27
I will try to explain what I think lopri is getting at, AMD's ability to supply only comes into effect only if the OEM becomes AMD exclusive which can't occur due to AMD's small size.
The problem is that it becomes a situation of choose Intel only and do business well, or choose to have both and Intel withdraws because it doesn't like you carrying the competitor's product. This latter choice can't be made by the OEM, because it would basically send them to oblivion, if they were completely reliant on AMD. So basically OEM's have to choose to Intel, there isn't a choice.
This is obviously a bad thing for AMD, as it doesn't matter how great it's product is, if it can't sell to that vendor, because of the competition.
This wouldn't be a problem in a Pepsi vs Coke environment where both companies are of comparable marketshare, as either one has enough capacity that a vendor can choose one and not go bankrupt because of it. Not so in the AMD/Intel competition


bryanW1995
also, as lopri mentioned, amd was/is inferior in mobile chipsets, so even the most die-hard A64 junkie company would want to be able to use intel for those. Intel's forcing other companies to not do business with amd was an abuse of monopoly-like power

golem
I don't know then, but from what you and coldpower27 say, then Intel is not allowed to choose who they sell to or to give priority to larger or more loyal customers or else they are accused of abusing monopoly power

bryanW1995
we don't make the laws in the US, congress does. ask them why it is illegal to abuse monopoly powers

golem
Are you sure that's the correct definition of abuse of monopoly power, or just what AMD wants it to be? It still seems ridiculous to me that once you reach a certain market share, you lose the ability to decide who you do business with (as long as the the reason is legitimate and legal).
 

bryanW1995

Lifer
May 22, 2007
11,144
32
91
reply to golem:


once again, that is the law here in the US. Large companies with dominant market share like MS and intel have to very careful that they don't get broken up into smaller pieces ala AT & T or Standard Oil. These days it is much more difficult to be a "monopoly" since you could theoretically have 100 % of the US market but have a strong competitor in a foreign market keep you from being a monopoly.
 

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
OK...now for some of those replies...

Nemesis...
"Do you have a link about which e-mails are missing. If their is no record how do we know there missing"

Because there are logs that state how many and who sent e-mails on the servers. The content and destination isn't included in those logs, but they exist. When the logs for number of e-mails don't tally with the e-mails turned over in disclosure, it's evident that there has been a breach of court orders.
BTW, Intel has already admitted under oath that 1000s of e-mails were destroyed in breach of court orders...and over 1000 of these were from the highest executives.

"Why aren't you and others upset about the already settled case were amd reverse engineered intel cpu's and basicly stoled x-86"

AMD was asked by Intel to be a second source (so that Intel could meet their contractual obligations to IBM)...they didn't steal, they were given x86.

But does it really matter? All companies reverse engineer (where did you think EM64T came from)...and we're is getting away from the point of THIS thread.
 

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
Golem...
"Are you sure that's the correct definition of abuse of monopoly power, or just what AMD wants it to be? It still seems ridiculous to me that once you reach a certain market share, you lose the ability to decide who you do business with (as long as the the reason is legitimate and legal)."

Well, there's doing business and doing business...let me explain with a hypothetical.

1. Let's assume (remember that this is a hypothetical!) that AMD is offering chips that are better (faster, use less power, and cost less) to OEMs.
2. Also assume that an OEM requires x number of chips to keep their customers supplied and that their competitors also require x number of chips.
3. Assume that AMD can only provide x/2 chips and Intel can produce 2x chips

Obviously, all OEMs MUST do business with Intel in order to survive...what the anti-trust laws protect are the decisions of OEMs to include AMD in their lineup without the threat of either
1. A lack of timely supply of necessary Intel chips
2. A comparative cost on Intel chips which is higher than their competitors

In other words, retribution for using AMD chips by a monopoly whose supply is necessary for their survival.

In many businesses, this wouldn't be an issue as a company like AMD would merely increase their capacity...but in the semiconductor industry, the cost and time to increase capacity is so extraordinarily high, that this is a practical impossibility in the short term (this is called having a "high barrier-to-entry").
Therefore, in order to protect and encourage smaller companies to innovate and compete (thus improving product and decrease prices), the anti-trust laws seek to limit a monopoly's ability for eliminating competitors.
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
As long as MS is around in its present state I can't see how Intel can be considered a monoply.

Let us take Micheal Dell as an example. He started building PC ' in a dorm . Using Intel cpu's . He stayed exclusive to intel and built the company into the no. 1 selling maker on the market. Why should he use anything else . Only after he started selling AMD did he lose that number 1 position.

Why didn't some PC maker form out of the Dust that was AMD64 to become the no1 vender of pc's?

Is it a monoply tactic that intel has advertizing dollars and AMD doesn't?

Is it a monoply that the time were talking about that AMD only had 1 fab and intel had many .

On intel suppling X86 to IBM . Intel didn't go to AMD as a second source on their own IBM forced the issue . SO intel choose AMD as second source. But Intel never gave AMD x86 as their own . That only came latter as a settlemrnt in another law case. After AMD reversed engineered the 386 I believe . That when I believe the P pro was born not sure on that. Than intel used a new socket.

Intel had the resources to advertize were AMD didn't . AMD had one fab intel had several. It only stands to reason that intel built its business off the likes of HP COMQUAKE DELL and others. Intel had the resources AMD didn't now AMD cries foul . AMD is like a child who doesn't respect its father who spawned it. Shame on AMD for being so disrespectful of its parent. THe son will never be the father.

I am so happy about SSE4 and AMD's announment of SSE5 . AMD knows that Intel is moving away from a pure X86 cores and they know what SSE4 is all about.
TO bad for AMD but Intel has to get this monkey off their backs.

The havok announcement has a lot off people saying Intel will optimize for their CPU's and not AMD's . This is true because AMD won't have sse4 untill after 09 or sse5 as was written. Intel will optimize havok for sse4 since amd isn't using sse4 all are correct but it isn't intels fault .

What year was this lawsuite started? AMD can only go back to the last lawsuite they brought against intel . If you check the dates AMD had only 1 fab . and the window between the lawsuites wasn't that many years.

AMD has never had the fab capicity until recently to supply all the chips dell requires . AT time the lawsuite was filled. AMD's problem really wasn't Intel. In K8's hayday it was the lack of resources. FABS and ADVERTIZEMENT.

If you think other your just fooling yourselves. Coupons and rebates are normal its not a monoply tool as many make out.
 

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
As long as MS is around in its present state I can't see how Intel can be considered a monoply
I think you're making the same mistake that a lot of people do here...you need to know that monopolies are NOT illegal! Microsoft is and has been ruled to be a monopoly, but that only effects they way they are allowed to compete.
Only after he started selling AMD did he lose that number 1 position
I think if you look again that the opposite is true...Dell has only been a vendor for AMD since last Dec, but already they have turned their server sales around (revenues increased 20% over last year when they were Intel only).
As to losing the number 1 slot, that was in Q3 06 when they were beat by HP (who boasted AMD as their most successful server chip as far as gains went).
The havok announcement has a lot off people saying Intel will optimize for their CPU's and not AMD's
Intel has already announced that they won't do this...and they'd be pretty stupid to. All that would do is give developers a very good reason not to use their runtimes (why would anyone develop with a product that doesn't work on all popular chips?)...Intel isn't that stupid.
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
Originally posted by: Viditor
As long as MS is around in its present state I can't see how Intel can be considered a monoply

Really why doesn't the courts force MS to share source with another company as they forced Intel to share X86?

I think you're making the same mistake that a lot of people do here...you need to know that monopolies are NOT illegal! Microsoft is and has been ruled to be a monopoly, but that only effects they way they are allowed to compete.

So weres the competition? Why is vista so heavely optimized for AMD processors?

Only after he started selling AMD did he lose that number 1 position
I think if you look again that the opposite is true...Dell has only been a vendor for AMD since last Dec, but already they have turned their server sales around (revenues increased 20% over last year when they were Intel only).

Your spinning again point is they lost the no.1 position only after AMD came onboard thats all I said . Its a fact. On the server sells increase how many of those servers were Intel . If your going to make statements based on sells volumn supply the numbers of amd servers sold vs. Intel servers sold. I believe intel has gained servers sells since release of Merom is this not a fact?

As to losing the number 1 slot, that was in Q3 06 when they were beat by HP (who boasted AMD as their most successful server chip as far as gains went).

Supply link that it wasn't the 4th qt.

The havok announcement has a lot off people saying Intel will optimize for their CPU's and not AMD's



Intel has already announced that they won't do this...and they'd be pretty stupid to. All that would do is give developers a very good reason not to use their runtimes (why would anyone develop with a product that doesn't work on all popular chips?)...Intel isn't that stupid.

Ya so intels telling truth . But if the use SSE4 I guess that doesn't help AMD now does it . AMD can use SSE4 if they choose but only after intel releases the cpu's that have sse4. Why wouldn't developers use a runtime for a cpu's that have 80% of the market.
 

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
Originally posted by: Nemesis 1
Originally posted by: Viditor
As long as MS is around in its present state I can't see how Intel can be considered a monoply

Really why doesn't the courts force MS to share source with another company as they forced Intel to share X86?

They didn't force Intel to share x86 because of any anti-trust suit...they did that because according to the original contract with AMD as a second source, they found that Intel didn't have the right to pull the contract arbitrarily when it became inconvenient to them.

I think you're making the same mistake that a lot of people do here...you need to know that monopolies are NOT illegal! Microsoft is and has been ruled to be a monopoly, but that only effects they way they are allowed to compete.

So weres the competition? Why is vista so heavely optimized for AMD processors?

I honestly have no idea what you're talking about here...could you be more specific?

Only after he started selling AMD did he lose that number 1 position
I think if you look again that the opposite is true...Dell has only been a vendor for AMD since last Dec, but already they have turned their server sales around (revenues increased 20% over last year when they were Intel only).

Your spinning again point is they lost the no.1 position only after AMD came onboard thats all I said . Its a fact. On the server sells increase how many of those servers were Intel . If your going to make statements based on sells volumn supply the numbers of amd servers sold vs. Intel servers sold. I believe intel has gained servers sells since release of Merom is this not a fact?

As to losing the number 1 slot, that was in Q3 06 when they were beat by HP (who boasted AMD as their most successful server chip as far as gains went).

Supply link that it wasn't the 4th qt.

Dell loses Number 1 spot in PCs

Besides...do you actually think a company loses that kind of lead so quickly? Dell had been losing marketshare for well over a year prior to that...

The havok announcement has a lot off people saying Intel will optimize for their CPU's and not AMD's

Intel has already announced that they won't do this...and they'd be pretty stupid to. All that would do is give developers a very good reason not to use their runtimes (why would anyone develop with a product that doesn't work on all popular chips?)...Intel isn't that stupid.

Ya so intels telling truth . But if the use SSE4 I guess that doesn't help AMD now does it . AMD can use SSE4 if they choose but only after intel releases the cpu's that have sse4.
But SSE4 doesn't help Intel OR AMD until they're implemented into apps...and that's well after AMD will have it (if required). SSE3 took over a year to be fully implemented...
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
True sse3 took off slow . But sse4 is a differant cat altogether. When you see performance increases of 100% or more developers are going to be all over SSE4 like flies on crap.

I won't debate this any longer I have my views you have yours which is a great thing.

But this is what I believe. True or not its what I believe. Intel is moving away from x86 . They are going in the direction of simple multicores using VLIC(EPIC) and the use of code morphing with a modded Elbrus compiler. Intel is going to true 64bit binaries. that can morph x86 instuctions. So as to get the x86 monkey off their backs known as AMD.

Basicly intel is saying to AMD here here you want x86 so bad here here take it . Intel is also moving against NV and ATI in that I believe intel wants to go raytracing now this will take a bit longer as the softwere is required but many are working hard at that as we speak. As penryn as demonstrated and nethalem will really demonstraight raytracings abilities.

Anyway by the time this lawsuite is over I believe intel will be using VLIC processors that morph x86 instructions at which time intel will also be using mitosis.

Call me crazy but thats what I believe and the info thats available on the web kind of supports what I am saying . So basicly befor this lawsuite is over Intel will beable to say to the courts . Were no longer making x86 processors which we invented and have given that tech to amd.

On the Dell thing I thought it was 4th qt. thanks for the link. But it doesn't matter as AMD was being sold by dell in the 3rd quarter. I don't know when the AMD pc came to market but dell was buying amd cpu's in the 3rd qt.
 

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
Originally posted by: Nemesis 1
True sse3 took off slow . But sse4 is a differant cat altogether. When you see performance increases of 100% or more developers are going to be all over SSE4 like flies on crap.

The problem is that it takes a year to actually develop the app itself...even if a developer wanted to, they probably couldn't implement any sooner.

I won't debate this any longer I have my views you have yours which is a great thing.

Fair enough and well said.

But this is what I believe. True or not its what I believe. Intel is moving away from x86 . They are going in the direction of simple multicores using VLIC(EPIC) and the use of code morphing with a modded Elbrus compiler. Intel is going to true 64bit binaries. that can morph x86 instuctions. So as to get the e86 monkey off their backs known as AMD.

(I believe you meant VLIW there...Very Large Instruction Word)
Even with code morphing (otherwise known as CF or Code on the Fly), there is still a VERY substantial performance hit (at least 50% and higher at current levels). I think that you'll find that the cost of reworking all current software to compatability is probably more than Intel has made in Gross Revenue over the last 5 years (~$50 Billion), and unless there's a good reason for developers to do so, I can't see them raising their hands just to help Intel out...
Intel and HP have already spent well in excess $5 Billion getting the 12-15000 apps currently available for Itanium developed...


Basicly intel is saying to AMD here here you want x86 so bad here here take it . Intel is also moving against NV and ATI in that I believe intel wants to go raytracing now this will take a bit longer as the softwere is required but many are working hard at that as we speak. As penryn as demonstrated and nethalem will really demenstraight raytracings abilities.

Anyway by the time this lawsuite is over I believe intel will be using VLIC processors that morph x86 instructions at which time intel will also be using mitosis.

Call me crazy but thats what I believe and the info thats available on the web kind of supports what I am saying . So basicly befor this lawsuite is over Intel will beable to say to the courts . Were no longer making x86 processors which we invented and have given that tech to amd.

A couple of points here...
1. The lawsuit goes to court this coming April...
2. Even if Intel no longer makes ANY processers, they are still liable for the damages (which theoretically can be as high as $100 Billion...remember that anti-trust specifically states triple damages)
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
Ya it was vliw were the C came from is :confused:

How Fast Will It Be?
Like the Transmeta devices, software will not run at it?s full potential until it?s been fully translated, you can pretty much bet Intel will make sure third party bench-markers will be made well aware of this. I suspect we may also see speculative translation running in the background so everything gets translated and saved as soon as possible. Once translated, the new binaries are saved to disc, they will run as native VLIW thereafter.

The forte of this processor will be multithreaded code and multitasking. If you are doing lots of things at one you?ll be well happy, servers in particular will benefit from this approach. Multitasking will benefit because different cores will get different tasks, a user switching between them will not cause them to halt so responsiveness of systems with this processor will be very good.

Single threaded performance on the other hand could be relatively weak although that?s not a given,

You should read more about elbrus compiler great stuff let us not forget about mitosis its real and its coming to a threater near you.
 

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
Originally posted by: Nemesis 1
Ya it was vliw were the C came from is :confused:

How Fast Will It Be?
Like the Transmeta devices, software will not run at it?s full potential until it?s been fully translated, you can pretty much bet Intel will make sure third party bench-markers will be made well aware of this. I suspect we may also see speculative translation running in the background so everything gets translated and saved as soon as possible. Once translated, the new binaries are saved to disc, they will run as native VLIW thereafter.

The forte of this processor will be multithreaded code and multitasking. If you are doing lots of things at one you?ll be well happy, servers in particular will benefit from this approach. Multitasking will benefit because different cores will get different tasks, a user switching between them will not cause them to halt so responsiveness of systems with this processor will be very good.

Single threaded performance on the other hand could be relatively weak although that?s not a given,

You should read more about elbrus compiler great stuff let us not forget about mitosis its real and its coming to a threater near you.

You might want to clean up those other 2 posts...:)

Edit: I posted too late...thanks! :)

The only major weakness I see here is that it would require ALL developers to compile using an Intel "Elbrus-based" compiler...no other compiler would work on the system.
I just don't see that happening...but as you say, we shall agree to disagree.

As to mitosis, I've been following it for many years...both Intel and AMD have current projects trying to get it to actually work (no luck reported yet).
I agree that the Elbrus was brilliant (on par with Transmetta), especially so considering the lack of resources the Russians had in developing it.
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
Come on Viditor mitosis is a lot further along than your willing to admitt . Because its been posted here many times. Now reverse ht is back to what AMD does best monkey see monkey do . But their is almost zero info on the net about it other than AMD saying ya were working on that also.

I have a feeling next week is going to be an eyeopener for many. Mitosis has come along ways my friend . But that for another day and thread.

I just turned on virus scan and now no popups . Whats this a smart trojan???

Or have I lost control of this PC?
 

CTho9305

Elite Member
Jul 26, 2000
9,214
1
81
Originally posted by: Viditor
The havok announcement has a lot off people saying Intel will optimize for their CPU's and not AMD's
Intel has already announced that they won't do this...and they'd be pretty stupid to. All that would do is give developers a very good reason not to use their runtimes (why would anyone develop with a product that doesn't work on all popular chips?)...Intel isn't that stupid.

I take it you didn't know that Intel's compiler has been known to check if a CPU is Intel before using SSE codepaths. SSE codepaths should be used on any CPU that supports SSE (because you get a 2-4x speedup), but Intel's compiler only used them on Intel CPUs. That pretty much cripples the performance on any competitor's CPU.

From here:
Looking at __intel_cpu_indicator_init, it goes and checks the various processor capabilities and sets the global variable at 0xedb375c to a value depending on the processor capabilities. If you can do SSE2, the value is 0x200. If you can only do SSE1, you get 0x80, and so on. If you can't do anything, you get the value 0x1. What is odd is that the Athlon chips can do SSE quite happily, but seem to end up with a value 0x1 anyway. Tracing through the code on a P4 and an Athlon revealed the culprit (in __intel_cpu_indicator_init).
<some assembly>
The 'cpuid' call here puts the CPU manufacturer's ID in ebx:edx:ecx. The 'cmp' instructions later then check that these values were 'Genu','ineI','ntel' (i.e. 'GeniuneIntel'). If not, then we jump off to a bit of code that doesn't even pretend to check for the CPU capabilities but instead returns 0x1 (i.e it's a 386, and nothing other than the bog-standard 386 instruction set is available).

Think about what this means. The code produced by the Intel compiler checks to see if it's running on an Intel chip. If not, it deliberately won't run SSE or SSE2 code, even if the chip capability flags (available through the 'cpuid' instruction) say that it can. In other words, the code has been nobbled to run slower on non-Intel chips.
 

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
Originally posted by: CTho9305
Originally posted by: Viditor
The havok announcement has a lot off people saying Intel will optimize for their CPU's and not AMD's
Intel has already announced that they won't do this...and they'd be pretty stupid to. All that would do is give developers a very good reason not to use their runtimes (why would anyone develop with a product that doesn't work on all popular chips?)...Intel isn't that stupid.

I take it you didn't know that Intel's compiler has been known to check if a CPU is Intel before using SSE codepaths. SSE codepaths should be used on any CPU that supports SSE (because you get a 2-4x speedup), but Intel's compiler only used them on Intel CPUs. That pretty much cripples the performance on any competitor's CPU.

From here:
Looking at __intel_cpu_indicator_init, it goes and checks the various processor capabilities and sets the global variable at 0xedb375c to a value depending on the processor capabilities. If you can do SSE2, the value is 0x200. If you can only do SSE1, you get 0x80, and so on. If you can't do anything, you get the value 0x1. What is odd is that the Athlon chips can do SSE quite happily, but seem to end up with a value 0x1 anyway. Tracing through the code on a P4 and an Athlon revealed the culprit (in __intel_cpu_indicator_init).
<some assembly>
The 'cpuid' call here puts the CPU manufacturer's ID in ebx:edx:ecx. The 'cmp' instructions later then check that these values were 'Genu','ineI','ntel' (i.e. 'GeniuneIntel'). If not, then we jump off to a bit of code that doesn't even pretend to check for the CPU capabilities but instead returns 0x1 (i.e it's a 386, and nothing other than the bog-standard 386 instruction set is available).

Think about what this means. The code produced by the Intel compiler checks to see if it's running on an Intel chip. If not, it deliberately won't run SSE or SSE2 code, even if the chip capability flags (available through the 'cpuid' instruction) say that it can. In other words, the code has been nobbled to run slower on non-Intel chips.

I'm VERY aware of the compiler issue...it's part of the anti-trust case.
However, as soon as this part of the code was made public (and AMD filed), Intel re-wrote their compilers...they no longer shut off anything. That's why I feel fairly confident that they won't make the same mistake twice...
 

bryanW1995

Lifer
May 22, 2007
11,144
32
91
or they'll be smarter about it next time. Realistically, however, AMD in its current form is not much of a threat to intel and does act as a buffer to other entries into the cpu manufacturing segment, so it's probably not in intel's best interest to engage in any overt acts of monopoly power abuse. Now, if fusion ends up being a big hit then the gloves might come off again, but I don't think that intel is going to fall asleep at the wheel again.
 

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
Originally posted by: Nemesis 1
Come on Viditor mitosis is a lot further along than your willing to admitt . Because its been posted here many times. Now reverse ht is back to what AMD does best monkey see monkey do . But their is almost zero info on the net about it other than AMD saying ya were working on that also.

I have a feeling next week is going to be an eyeopener for many. Mitosis has come along ways my friend . But that for another day and thread.

I just turned on virus scan and now no popups . Whats this a smart trojan???

Or have I lost control of this PC?

Good luck with the Trojan, mate...

As to Mitosis, I don't think it's possible for it to show up before Nehalem, and even then may be too soon. Mitosis requires 2 things to work...
1. An aggressively tuned mitosis-specific compiler, and
2. Some large buffers specifically designed for speculative threading in the chip.

None of the Core chips have the buffers (they would show up in the die shots), and the compiler would have to be out there with developers for quite awhile before it would be effective. The compiler is also why it would have to be a joint Intel-AMD project...developers aren't going to accept a compiler that doesn't work with both companies.
 

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
Originally posted by: bryanW1995
or they'll be smarter about it next time. Realistically, however, AMD in its current form is not much of a threat to intel and does act as a buffer to other entries into the cpu manufacturing segment, so it's probably not in intel's best interest to engage in any overt acts of monopoly power abuse. Now, if fusion ends up being a big hit then the gloves might come off again, but I don't think that intel is going to fall asleep at the wheel again.

Actually, when AMD gains closer to 40% of the market, it's quite probable that Intel can stop worrying about monopoly status (though by no means is that a sure thing).
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
Originally posted by: Viditor
Originally posted by: Nemesis 1
Come on Viditor mitosis is a lot further along than your willing to admitt . Because its been posted here many times. Now reverse ht is back to what AMD does best monkey see monkey do . But their is almost zero info on the net about it other than AMD saying ya were working on that also.

I have a feeling next week is going to be an eyeopener for many. Mitosis has come along ways my friend . But that for another day and thread.

I just turned on virus scan and now no popups . Whats this a smart trojan???

Or have I lost control of this PC?

Good luck with the Trojan, mate...

As to Mitosis, I don't think it's possible for it to show up before Nehalem, and even then may be too soon. Mitosis requires 2 things to work...
1. An aggressively tuned mitosis-specific compiler, and
2. Some large buffers specifically designed for speculative threading in the chip.

None of the Core chips have the buffers (they would show up in the die shots), and the compiler would have to be out there with developers for quite awhile before it would be effective. The compiler is also why it would have to be a joint Intel-AMD project...developers aren't going to accept a compiler that doesn't work with both companies.

Agreed and next week intel is going to give us a lot more info on nehalem . Nehalem right now is less of a wait than K10 or K8L back than . Thats why I said After next week speculation about Nehalem is going to go way up in forum threads . We mite even get a die shot. I hope. Even tho nehalem is said to be 4 issue processor thats about all it will have in common with C2D with the capabilities to have 8 native cores or what ever combination intel chooses. Coprocessors Vertex engine whatever intel chooses

I think next week is going to light the forum threads on fire. Its our jobs to bring those poster here. The best forums on the net . Since the new policies its by far the best. We can debate without all the disrespect shown in other forums to the posters. Look at XS its turned into a joke and thats not me saying that its long time members at XS saying it. Read the forums their .

I think fugger and many others that post thier are great. I wish I could name them all but its not possiable. I could easily name the flamers and baiters but thats wouldn't be right either. All I know is since C2D came out the forum has gone to the dogs.

I use to debate rollo all the time at other foorums because I believed he was what we all know now he was. But I liked Rollo a lot he made me think and research. He made me angry also but so what. I read the stuff here back when the AEG thing came up . I was very disappointed in how he was treated. None should have been abgry with Rollo . It should have been NV that angered people. I personally gained a lot of respect for Keys at that time as he acted like a christian man should have . Or just maybe a good man. Debate is great thing but causing anger and personnal insults as a result is just plain childish. I know many forum members are young men but thats not an excuse for bad behavior.

If a forum thread upsets someone stay out of it and let a good debate on the subject continue . Book mark it and when the hardware is released see were the chips fall.

This thread about the lawsuite is a good one. I stand with intel on this one viditor doesn't. So we debate it. Thats good . But we still have to wait till its over befor we know how the courts view it.

As Viditor said 100 billion thats a joke . AMD was their own worst enemy . Thats my belief . His is Intel caused AMDs problems. But AMD still only had 1 fab and zero advertizing. Intel could supply DEll with all the chips they required and anyone else. AMD couldn't do that . I think in our minds we all know thats the trueth.

On the AMD 40% market share Viditor don't hold your breath as your to young to pass on.

 

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
Originally posted by: Nemesis 1
I know offtopic but here is a link for you viditor. Note the date. At the bottom by intel mole

http://techreport.com/articles.x/8721/4

Mitosis actually goes back to 1997 (Wang and Franklin out of Clemson in their paper "Highly Accurate Data Value Prediction Using Hybrid Predictors").
1997 paper
To my knowledge, it's still only about 82% efficient, and that's far too low to justify the resources on-chip for it...
 

CTho9305

Elite Member
Jul 26, 2000
9,214
1
81
Originally posted by: Viditor
Originally posted by: Nemesis 1
I know offtopic but here is a link for you viditor. Note the date. At the bottom by intel mole

http://techreport.com/articles.x/8721/4

Mitosis actually goes back to 1997 (Wang and Franklin out of Clemson in their paper "Highly Accurate Data Value Prediction Using Hybrid Predictors").
1997 paper
To my knowledge, it's still only about 82% efficient, and that's far too low to justify the resources on-chip for it...

Speculative Multi Threading is very different from simple value prediction. Speculative multi-threading goes back much farther than that. The first paper on it that I'm familiar with is from 1994. I don't know if the multiscalar group's earlier publications were specifically about SMT or not (I haven't read them).
 

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
Originally posted by: CTho9305
Originally posted by: Viditor
Originally posted by: Nemesis 1
I know offtopic but here is a link for you viditor. Note the date. At the bottom by intel mole

http://techreport.com/articles.x/8721/4

Mitosis actually goes back to 1997 (Wang and Franklin out of Clemson in their paper "Highly Accurate Data Value Prediction Using Hybrid Predictors").
1997 paper
To my knowledge, it's still only about 82% efficient, and that's far too low to justify the resources on-chip for it...

Speculative Multi Threading is very different from simple value prediction. Speculative multi-threading goes back much farther than that. The first paper on it that I'm familiar with is from 1994. I don't know if the multiscalar group's earlier publications were specifically about SMT or not (I haven't read them).

I actually found the Wang/Franklin paper because it's referenced in a number of SMT papers...for instance this one from UCSD.