OPTIMIZER WANTED FOR A SIX CORE AMD PROCESSOR

mikey
mikey
Joined: 22 Jan 05
Posts: 12,087
Credit: 1,834,325,298
RAC: 12,906

RE: RE: In a fair race, I

Message 99844 in response to message 99841

Quote:
Quote:
In a fair race, I wouldn't expect the lead runner to be spreading a trail of nails and bombs behind them to obliterate all competition... Or do "Dick Dastardly" rules rule?

Sorry to point it out Martin, but you keep dodging the key question by divergent constructions. In your chosen analogous terms the racetrack has an Intel banner over it, an Intel logo painted on the ground, everyone has an Intel T-shirt on ( except the AMD guy ) and the entire area is plastered with Intel advertising. And Intel paid for it all to boot. I knew all of that many years ago when I bought an Intel compiler for myself. It said, and still does, that it is designed for optimising code for Intel products. There are plenty of other racetracks for AMD to go running at, including their own should they wish to pay for the construction of one. So why oh why does AMD believe, and hence yearn in the breach, in 'fairness' at this running location - specifically expecting shareholders of another company to pay for some later manufactured obligation to them, within a product they would never have contributed one iota to? What blindfolds, or pretense thereof, do they wear?

This is why the apparent definition of 'fair' seemingly used in this instance actually looks quite screwy from a distance. Maybe one wrong is being used to correct another perhaps, or somesuch, but it looks like a very dangerous precedent indeed.

AMD has competed to some degree - but by going to court and diverting effort away from exercising their core business technical skills. If this legal process has rewarded a business that manufactures legal fouls rather than better mousetraps, then such a system is truly doomed.

Cheers, Mike.

( edit ) If you do think that shareholders of one company should be accounting positively for the interests of shareholders of another company, then just say so! I won't think any the worse of you. I just can't work out what greater evil is apparently avoided by doing so, though. :-) :-)

I think the point is that Intel did not only not help AMD, which they have no obligation to do, but they actually worked to impede AMD's progress. THAT is what they did wrong! I have no problem going to a competitors place of business and being treated like a 2nd class citizen, but to have my reservation lost on purpose, my bags run over by the bus again on purpose, etc, etc, is inexcusable! Competition is one thing, but to intentionally write code to actively sabotage a competitor is another. It means that you are not confident that your product can stand on its own and beat the other one head to head!

MS did this very thing in the 'browser wars' many years ago when they wrote code so that Netscape's Navigator ran slower and slower and slower until the users eventually chose IE because it didn't do that. MS ended up paying a HUGE fine and even being forced to unbundle IE from Windows, in Europe, for a time. MANY things have changed since then and IE is back in, but MS has agreed that they will not do anymore coding to actively slow down other companies browsers!

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6,552
Credit: 289,676,357
RAC: 108,078

RE: I think the point is

Message 99845 in response to message 99844

Quote:
I think the point is that Intel did not only not help AMD, which they have no obligation to do, but they actually worked to impede AMD's progress. THAT is what they did wrong!


OK, let's follow that then.

So AMD shareholders either have now, or should have had all along, the financial advantage of being gifted, from their closest competitor, equivalent use in the market place of a product they never outlaid/risked a cent on developing, nor ever intended to?

Please, not too many high fives here! More sober reflection may enlighten that a golden goose has been shot ....

Cheers, Mike.

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

transient
transient
Joined: 3 Jun 05
Posts: 62
Credit: 115,835,369
RAC: 0

Would it have been enough if

Would it have been enough if Intel had stated in their documentation; "AMD products will be inhibited, when compiling source using the Intel C Compiler."? Should AMD offer financial support in the development of the compiler? Or was that already implicit in the purchasing of the license to use the SSE instructions set. Does AMD pay a licensing fee for the use of the SSE instructions set to Intel? Does Intel even own the rights to SSE?

Bikeman (Heinz-Bernd Eggenstein)
Bikeman (Heinz-...
Moderator
Joined: 28 Aug 06
Posts: 3,522
Credit: 692,140,060
RAC: 18,951

RE: Would it have been

Message 99847 in response to message 99846

Quote:
Would it have been enough if Intel had stated in their documentation; "AMD products will be inhibited, when compiling source using the Intel C Compiler."? Should AMD offer financial support in the development of the compiler? Or was that already implicit in the purchasing of the license to use the SSE instructions set. Does AMD pay a licensing fee for the use of the SSE instructions set to Intel? Does Intel even own the rights to SSE?

I understand they have a cross-licensing deal. AMD doesn't pay for SSE, and Intel doesn't pay for using the amd64 instructions (the 64 bit instruction set is AMDs invention, from times when Intel still thought that 64 bit would be the domain of their non-x86-compatible Itanium CPU line)

Was a golden goose shot? noo, just had it#s wings clipped a tiny bit.

IMHO it's not about fairness or criminal charges or even shareholders. It's about us, consumers. Do we want to live in a world wheer we HAVE to buy Intel because there's noone else left? AMD isn't so strong financially...so I'm fine with Intel getting its wings clipped a little bit from time to time.

CU
HB

ExtraTerrestrial Apes
ExtraTerrestria...
Joined: 10 Nov 04
Posts: 770
Credit: 546,490,677
RAC: 193,995

The way I see it there were 3

The way I see it there were 3 options for Intel:

(1) Neither can we guarantee that the SSE code will work properly on the AMD nor do we want to invest any money in making sure there are no bugs. Allowing SSE on AMD may place our compiler in a bad light in case something goes wrong and thus reduce our revenue from selling it.

(2) We can't guarantee that our SSE code will work properly on AMDs, so we'll make it clear that if you feed our code to an AMD everything past that point happens at your own risk.

However: the issue raised in (1) applies not only to SSE, but also to normal code. So by "allowing" AMDs to run their normal code they already agreed to (2) in case of not-SSE instructions. So they can not convincingly argue according to (1) in the case of SSE, as a failure on non-SSE code would be just as bad.

What other justification attempt is left?

(3) It's our software. We're free to decide which code we put in there. If you don't like it, don't use it.

Option (3) may sound fair enough at first glance. However, within a society there are limits to freedom. In Germany these start when your actions harm or trouble others. So going back to that race track analogy: even if the track is owned by Intel (and clearly marked as such) and the AMD racer decides to use unappropriate tires Intel wouldn't be forced to adapt their track to make him fast as well - but they're not allowed to pop his tires.

MrS

Scanning for our furry friends since Jan 2002

Metod, S56RKO
Metod, S56RKO
Joined: 11 Feb 05
Posts: 135
Credit: 812,344,933
RAC: 68,586

It's all lovely and

Message 99849 in response to message 99848

It's all lovely and everything, it's just the way Intel did it. As Dan described in this post, check was not made as Intel VS the rest, but rather against AMD.

So if it was option 1. or 2., did Intel test other vendors' products for proper execution? I highly doubt it. And even if it was option 3. which bothers me (not too much though), it was still directed against AMD.

My guess is that this fact (targeting AMD procs, not protecting their own procs) was what pissed most people.

Metod ...

mikey
mikey
Joined: 22 Jan 05
Posts: 12,087
Credit: 1,834,325,298
RAC: 12,906

RE: RE: I think the point

Message 99850 in response to message 99845

Quote:
Quote:
I think the point is that Intel did not only not help AMD, which they have no obligation to do, but they actually worked to impede AMD's progress. THAT is what they did wrong!

OK, let's follow that then.

So AMD shareholders either have now, or should have had all along, the financial advantage of being gifted, from their closest competitor, equivalent use in the market place of a product they never outlaid/risked a cent on developing, nor ever intended to?

Please, not too many high fives here! More sober reflection may enlighten that a golden goose has been shot ....

Cheers, Mike.

I am not very good at analogies but let me try it this way: you and someone else are both competing for a grant to do research on something. They write their paper and send it in, you write your paper and send it in, but in the process of writing your paper you see what they said in their paper. So you write how your process will be better than theirs and actively campaign against their way of doing things. Is that fair? I wouldn't think so, each process should be compared on its own merits, if it fails it fails, if it succeeds it succeeds! And that is the whole point of writing the grant paper, to let someone else decide which process they want to fund. Now if they ask for a comparison of the different way that is a different story, but to write the original grant request and to slam your competitors process, is just not ethical, is it? Each paper should show how it is the best of the best and not even bring up the other persons process as a possible alternative, shouldn't it?

Let's say I want to build a telescope so I can see Mars and find the spacecraft Beagle and be able to physically see it like thru a pair of binoculars at a ball game. I write a grant proposal for someone to give me the money to build my telescope, not even bringing up any other alternative ways as options to my plan. But during the process I find out you are trying to do the same thing but use the Hubble telescope to do it. should I slam your idea or promote mine, or both. IMO I should promote mine to the fullest and not even mention yours as a viable option. This example is in reality a very silly one which is why I chose it.

ML1
ML1
Joined: 20 Feb 05
Posts: 347
Credit: 86,321,259
RAC: 132

RE: RE: I think the point

Message 99851 in response to message 99845

Quote:
Quote:
I think the point is that Intel did not only not help AMD, which they have no obligation to do, but they actually worked to impede AMD's progress. THAT is what they did wrong!

OK, let's follow that then. ...


Looks like no "Dick Dastardly" is needed... Looks to me like Intel's dirty deeds have been done, all to the monopolistic detriment to progress and to all our cost:

Dirk Meyer: AMD is NOT Up For Sale!

... "AMD is not for sale, but we are happy to listen to any proposal which is in the interest to our shareholders," Chief Executive Dirk Meyer said...

Regards,
Martin

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6,552
Credit: 289,676,357
RAC: 108,078

RE: RE: RE: I think the

Message 99852 in response to message 99851

Quote:
Quote:
Quote:
I think the point is that Intel did not only not help AMD, which they have no obligation to do, but they actually worked to impede AMD's progress. THAT is what they did wrong!

OK, let's follow that then. ...

Looks like no "Dick Dastardly" is needed... Looks to me like Intel's dirty deeds have been done, all to the monopolistic detriment to progress and to all our cost:

Dirk Meyer: AMD is NOT Up For Sale!

... "AMD is not for sale, but we are happy to listen to any proposal which is in the interest to our shareholders," Chief Executive Dirk Meyer said...


Well hubris and flawed analogies aside, I'm slightly stunned that no one seems to have come up with an even simpler explanation that fits all of this : Intel attempted to dislodge it's biggest competitor that was getting an extended free ride on one of it's very successful products, got discovered doing that on the quiet, the competitor complained ( rather than respond with an improved product ) and got a payout via legal action and not direct market presence from anything it risked outlay upon. Intel didn't do it nicely for sure, but I say again : so what? Does anyone believe AMD would have compensated Intel shareholders if they led the market?

That's what I don't understand, that derivation in preference to innovation has been rewarded in the matter of this compiler. Hence I query all this clapping on the back that progress has been served in some way .... it's quite conceivable that shareholders ( remember one competes on a world wide market for their dollar ) will probably withdraw from both companies ie. the US jurisdiction altogether, given that this type of ruling applies. Competition means you can actually lose. So applying an 'every child wins a prize' or some similiar emotional paradigm feels good for sure but is unlikely to give long term benefit IMHO. So maybe, just maybe, AMD simply just wasn't as good as Intel at the core technical business that they overlap in, and AMD ( management ) has resorted to the victim role.

Aside : In Australia such issues are very much more on the side of cartels - where 'opposing' entities are exceedingly friendly to each other. A recent notable case was price fixing in the concrete supply for building in the Sydney area. We could have very much used some pro-active competitor inhibition here! Here though, the issue is different from AMD/Intel - that being the ability to detect independence of business entities. AMD & Intel are undoubtedly separate.

Cheers, Mike.

( edit ) Is there a tarring of Intel with a Microsoft brush here?

( edit ) I suppose the other problem is false dichotomy. This is an X. We don't like X's. Here is a Y. A Y is not an X. Therefore I like a Y. So how would I have solved the issue? Well, it's all been done before really ( say AT&T -> the Baby Bell's ). Require Intel ( or any hardware producer by extension ) to divest compiler ( read software ) production. Sell off the compiler et al to a set of shareholders not beholden to either AMD or Intel. You immediately solve all of the concerns expressed here. This new entity will, to satisfy it's shareholders, aim to please all comers. This yields innovation and fairness to the IT industry as a whole as defined by the genuine concerns expressed here regarding ripple effects. Give the goose a new home/owner. I'd propose this as a constructive rather than a destructive solution ...

( edit ) @Bikeman : that's the trouble with analogies. My meaning of 'the goose' was Intel's compiler, not Intel as a whole. You're right about consumer choice, it is valuable. However there's a price somewhere, and you still need a shareholder group to risk the development toward a market to sell to. Disconnecting risk from reward has traditionally been the death knell of many a venture. I don't invest in any public shares by the way .... just my own personal business. Notionally I may choose risk and reward levels, but I do sometimes wonder!! :-) :-)

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6,552
Credit: 289,676,357
RAC: 108,078

RE: Dirk Meyer: AMD is NOT

Message 99853 in response to message 99851


Martin, you didn't mention this bit ( from Oracle CEO ) :

Quote:
.... the company was focused on building its intellectual property portfolio.


Please do realise the incredible irony within this statement. :-)

Cheers, Mike.

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.