Nvidia or IBM to acquire AMD ?

Page 3 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Idontcare

Elite Member
Oct 10, 1999
21,110
64
91
Originally posted by: Nemesis 1
Somebody has to stop these false analyst reports.

LOL, is there any other kind?

Seriously unless you personally pay for the report you are receiving "information" for free, and no one works for free, so they are all releasing the info to legally manipulate the stock market.

From what I have gathered the SEC is pretty much as useless to individual shareholders as FEMA was to Katrina victims.

When they do strike a finding they fine the company or evildoers and take receipt of the fines for themselves. I can't recall the SEC ever demanding the offending party put some coin back into the victims pocket.
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
Here is the waiver agreement. I bolded the part that relates to this discussion. As you can plainly see 6.2 both a&b relate to material breaches.
Well the bold didn't work. Read b
3.9. Waiver of Indirect Infringement Liability.
-----------------------------------------

(a) For purposes of this Section 3.9, "Indirect Infringement"
means a claim for infringement where the accused infringer is
not directly infringing the subject patent rights(s), but is
in some manner contributing to a third party's direct
infringement of the subject Patent Rights(s) by, for example,
supplying parts or instructions to the third party that as a
result of such parts or instructions enable such third party
to infringe directly the subject patent rights(s). Indirect
Infringement includes without limitation contributory
infringement and inducing infringement.

(b) Each party agrees that, unless the licenses it has granted
hereunder are terminated pursuant to Section 6.2
, for any
Patents licensed hereunder and/or subject to Section 3.8(d) it
will not assert a claim of Indirect Infringement against the
other party ("Licensed Party") where such a claim would be
based in any part or in any way upon (a) any activity for
which the Licensed Party is licensed under this Agreement, or
(b) the Licensed Party providing instructions regarding or
sample designs related to its Licensed Products. The parties
agree that the foregoing sentence does not and shall not in
any way limit their respective rights to assert direct or
indirect claims of infringement against third parties.

3.10. No Other Rights. No other rights are granted hereunder, by
---------------
implication, estoppel, statute or otherwise, except as expressly
provided herein. Specifically, (i) except as expressly provided in
Section 3, nothing in the licenses granted hereunder or otherwise
contained in this Agreement shall expressly or by implication,
estoppel or otherwise give either party any right to license the
other party's Patents to others, and (ii) no license or immunity is
granted by either party hereto directly or by implication, estoppel
or otherwise to any third parties acquiring items from either party
for the combination of Licensed Products with other items or for the
use of such combination. Nothing in this Section 3 shall be deemed
as an agreement or prohibition against the manufacture, use, sale or
importation of any product by either party.
 

bunnyfubbles

Lifer
Sep 3, 2001
12,248
3
0
Originally posted by: lopri
Originally posted by: bunnyfubbles
because Intel will have their own solution as will AMD and both will have trump card system-on-a-chip solutions that nVidia currently cannot compete with as they are now...
Not if NV somehow successfully kisses MS' arse. Didn't you guys see APX 2500 Application Processor launch? If MS writes OS suited for NV, without touching the X86 line, then NV don't need X86 license. They're starting with small things of course.

wow, extreme niche vs. true mainstream...sorry, not impressed or encouraged at all

kissing ass wouldn't be enough, they'd need to go all out providing hj's, bj's and zj's
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
Thanks Idontcare Viditor had me so turned upside down I couldn't recall SEC as being proper usage. But you still understood. I agree with what ya said to . But one must try.
 

lopri

Elite Member
Jul 27, 2002
13,314
690
126
Originally posted by: bunnyfubbles
kissing ass wouldn't be enough, they'd need to go all out providing hj's, bj's and zj's
I get hj and bj, but what is zj? PM plz.
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
Viditor get your butt back here . You did read section 8 right . For those who haven't . I will post it . Now I know that viditor believes in his own mind that I can't comprehend the wording in this contract . But it really doesn't get in plainer than this. Even Viditor can't deny what it states. Well he can deny it . But its hopeless viditor.

8. ASSIGNMENT

8.1. No Assignment. This Agreement is personal to the parties, and the
-------------
Agreement or any right or obligation hereunder is not assignable,
whether in conjunction with a change in ownership, merger,
acquisition, the sale or transfer of all, or substantially all or
any part of a party's business or assets or otherwise, either
voluntarily, by operation of law, or otherwise, without the prior
written consent of the other party, which consent may be withheld at
the sole discretion of such other party. Any such purported
assignment or transfer shall be deemed a breach of this Agreement
and shall be null and void. This Agreement shall be binding upon and
inure to the benefit of the parties and their permitted successors
and assigns.

8.2. Limited Exception. As a limited exception to Sections 6.2 and 8.1,
-----------------
either party (but not any of its successors or assigns) may assign
the licenses granted under this Agreement pursuant to *****, subject
to the following conditions:

(a) The assigning party ("Assignor") must *****.

(b) *****.

(c) The licenses granted to Assignor hereunder terminate
immediately upon any such assignment and only the assignee
shall be entitled to the benefit of such licenses.

(d) *****.

(e) Prior to any such assignment becoming effective, the assignee
must:

(1) *****, and

(2) *****

(3) *****.

 

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
Originally posted by: Nemesis 1
Viditor get your butt back here . You did read section 8 right . For those who haven't . I will post it . Now I know that viditor believes in his own mind that I can't comprehend the wording in this contract . But it really doesn't get in plainer than this. Even Viditor can't deny what it states. Well he can deny it . But its hopeless viditor.

What's your point?
Obviously the patents aren't assignable, that's very standard...but it has nothing to do with the issues raised.

Not being assignable means that AMD can't give away x86 to a 3rd party, and Intel can't give away AMD's patents.

How is this relevant at all?

Edit: Please note the part that says "without the prior written consent of the other party, which consent may be withheld at the sole discretion of such other party"

Let me try again to construct this for you...

Let's say AMD is sold.
1. The purchasing company can no longer make CPUs unless they obtain permission in writing from Intel because AMD is not allowed to transfer those rights.
2. The purchasing company is now the sole owner of all of AMD's patents as they are a large amount of AMD's assets.
3. Intel can also no longer make CPUs without the purchasing company's permission because the former agreement is now voided (it was with AMD, but AMD no longer exists or owns the patents).

What actually would probably happen is that the new company and Intel would extend the previous agreement for a year or two to include the new owner while they hammer out a more permanent solution.

Edit2: I think you're hung up on trying to call something a breach of contract.
As it says, "Any such purported assignment or transfer shall be deemed a breach of this Agreement and shall be null and void"
But selling the company doesn't mean that anyone would be stupid enough to continue manufacturing without an agreement in writing...
In other words, the act of selling doesn't imply that they will assign or transfer Intel's IP without Intel's permission.
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
Viditor you do this all the time.


You asked why this is relevent. Than you go on to confuse the issue even more by writing this garbage

Edit2: I think you're hung up on trying to call something a breach of contract.
As it says, "Any such purported assignment or transfer shall be deemed a breach of this Agreement and shall be null and void"
But selling the company doesn't mean that anyone would be stupid enough to continue manufacturing without an agreement in writing...
In other words, the act of selling doesn't imply that they will assign or transfer Intel's IP without Intel's permission.



If intel says NO to assignment . It clearly says Breach of contract by said party. Which means that 6.2.a is in effect and intel can use AMDs patents. It clearly says this . Yet you try to confuse the issue with mumble jumbo. Viditor you do this type of thing all the time. It is not I who is confused .

Don't believe for one second I think your confused. Cause I know your not. You have an agenda . You have done this type of thing in the past many times.

You might confuse some of the kids here . and the idea your doing it on purpose is disgraceful.

People in on this topic ask some fundamental questions as to what AMD can do . So you assert that If intel doesn't give a third party trying to buy AMD approval to use x86 intel will lose its right to amd patents . Which is pure BS.

Simple trueth is If Intel says no. and AMD tries to do this anyway . This agreement is in breach of contract according to the agreement in section 6.2.a. Everthing that is written in section 8.1 is covered in section 6.2.b as to what constitutes breach of contract.

So the ans for people that wanted to know. Is this. If intel says no to amd selling out. and AMD does it anyway . The x86 is non transferrable without intels permission . Even if intel says yes . I new contract must be written as this agreement is null and void. Intel either way gets to keep using AMD patents. So basicly it comes down to this AMD either swims or drowns,

 

Idontcare

Elite Member
Oct 10, 1999
21,110
64
91
I wonder how muddy the waters are for the big-time investors (mutual funds) when they review these contracts.

I think we are safe to assume that to be sure both Intel and AMD have very different interpretations of the same contract in question.

I've never met two parties of the same contract who ever saw eye-to-eye on the terms, definitions, restrictions of the contract. Never ever. There is always plausible wiggle room, and teams of lawyers are paid big bucks to make that wiggle room large enough to drive a M&A thru.

We just won't likely ever know how the law would interpret this contract until it is tested in a court of law.

The good news, well maybe not so good, is that we stand pretty good chance of seeing it tested in matter of a few short years.

Whatever AMD does to crawl towards 32nm node will not help when Intel dives below the 32nm node with $80million EUV litho and 3D FinFETs while AMD is trying to figure out how to make 32nm yield with planar Hik/MG technology.

The inevitable downward spiral has begun, you can't expect to best a 4X R&D budget, you've got to go find your own niche.
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
Actually for us to determine exactly what this contract says we need to know *****.

But still this is fairky clear. But lawyers are good at twisting stuff. its true.

Idontcare I am just asking this . Because you confused me here a little.

Trigate or 3D is differant than FinFet 2gate I am sure you know this but the way its written is confusing.

I am pretty sure we will see FinFet on IBMs HighK/Metal gate process. I am pretty sure also that it will be better than Intels 45nm process.

Trigate 3D is what intel is suppose to bring next. I believe it will be at 32nm. Intel has pretty much in long term outlooks stated as much . The same as they did for 45nm HighK/ metal gates.

Time shall reveal all.
 

jameswhite1979

Senior member
Apr 15, 2005
367
0
0
Nvidia will and not Lenovo (IBM) as they have been selling off all their hardware and in the server arena AMD are nothing..
 

Idontcare

Elite Member
Oct 10, 1999
21,110
64
91
Originally posted by: Nemesis 1
Actually for us to determine exactly what this contract says we need to know *****.

But still this is fairky clear. But lawyers are good at twisting stuff. its true.

Idontcare I am just asking this . Because you confused me here a little.

Trigate or 3D is differant than FinFet 2gate I am sure you know this but the way its written is confusing.

I am pretty sure we will see FinFet on IBMs HighK/Metal gate process. I am pretty sure also that it will be better than Intels 45nm process.

Trigate 3D is what intel is suppose to bring next. I believe it will be at 32nm. Intel has pretty much in long term outlooks stated as much . The same as they did for 45nm HighK/ metal gates.

Time shall reveal all.

Intel won't throw away the HK/MG xtor's they developed for the 45nm node until they've been used for at least 2 nodes. No one builds one-node solutions, no one.

So 32nm is guaranteed to be the same HK/MG xtors from 45nm albeit with shallower implants, pocket extensions, etc and more strain engineering. I.e. a traditional shrink + optimize of the prior node's technology.

The absolute earliest opportunity Intel will take to change the xtor as radically as a migration to FinFET or 3D gates or 3D FinFETs will be the 22nm node. You can take that to the bank.

But the bottom line is it just doesn't matter what Intel does with their post-32nm tech as it will be out 2 years before IBM/AMD's and it will be at least technology node of development ahead of the curve.

The question for Intel is how to make this matter. Meaning they are about to find themselves up a creek trying to sell paddles where no one ventures to find themselves in need of a paddle. So they get to 22nm FinFET's 2 years before IBM/AMD, and they are pushing octal core desktop chips at 10GHz to Best Buy...uh is anyone gonna care or want to buy and octal core 10 GHz desktop for $500/CPU when they can get by with a 65nm tri-core Phenom at 3 GHz?

TI ran into the same issue with cellphones. When I left we were already making cellphone chips for 4G with 10 megapixel camera-phone capability and built-in DLP projector for live video-phone teleconferencing. Guess what, no one wanted it unless we sold the chip with the bundled features essentially for free. Folks are plenty happy with 3G for many years to come. So we shut the entire development efforts down. Sure TI will eventually bring the product to market, but in 10 or 15 years not in 1 or 2 years.

If no one wants it, unless you give it away for free, then why develop it. That is where Intel meets its own point of diminishing returns, with or without AMD biting at its ankles.
 

CTho9305

Elite Member
Jul 26, 2000
9,214
1
81
Originally posted by: jameswhite1979
in the server arena AMD are nothing..

In the multi-socket server market, AMD's processors are usually a much better choice than Intel's.
 

Idontcare

Elite Member
Oct 10, 1999
21,110
64
91
Originally posted by: CTho9305
Originally posted by: jameswhite1979
in the server arena AMD are nothing..

In the multi-socket server market, AMD's processors are usually a much better choice than Intel's.

True that, and the ASP for the 8xxx chips are proof of this. (generally >$1k on open market)
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
I agree with your logic. But to truely understand whats going on here . You must realize that Both were developing at the same exact time. IBM said Highk/Metal gates couldn't be done at 45nm. Intel said BS. Intel also wanted to intro Tri gate and High K/ metal gates at the same time. But they ran into some problems. Ist big problem was Intel had to go with gates last. Because the annealing process was metling the metal gates. So the annealing was done first than the gate material was added last. Something others said couldn't be done. So rather than bite off more than it could chew. Intel decided to use High K /Metal gates at 45nm. Work out any problems than do the project as they invisioned .


Tri gate was being developed at the exact same time as highK /Metal gates. The research money has been spent . The process is ready. We will see it @ 32nm.

As good as Intel has been to waite for 22nm could be a really bad move. I look for 22nm to be at least 1 year late maybe 2. 16nm I will be dust.
 

Idontcare

Elite Member
Oct 10, 1999
21,110
64
91
I have personally hands-on experience with HiK/MG as well as FinFET, while laborious to boil down how exactly this pre-disposes my bias towards release dates of certain technologies I will say the 2-node criterion for recycling developed "features" in technology nodes is one of budget and accounting reasons rather than one of "can we do it or not" type scientific reasons.

To put it differently, no shareholder of Intel's would be pleased with the capex numbers Intel would be chewing thru were they to migrate to 32nm and require new xtor toolsets after having just tooled up for HK/MG. Their cost-driven choices for how they implemented HK/MG at 45nm shows this to be true too. (<2% cost adder per wafer)

Further proof of how sensitive Intel manages their capex transitions is obvious when one considers the transition timelines for Al to Cu, TEOS to FSG, and 193nm scanner to immersion scanners. When Intel makes an investment into a particular process technology they milk it for as long as their design guys can live without the performance improvements that the new tech would provide.

I don't say this as if Intel is special or unique, 100% of all IDM's operate this way. And it is in this way that you can rest assured HK/MG planar CMOS is going to be 32nm POR (process of record) at Intel. It is incumbent upon the decision makers to recycle the technology for a minimum of 2 nodes.

If they have FinFET ready for production before it is needed for production then that will absolutely be viewed as management having wasted resources as it means the FinFET R&D team was over-resourced so much so that they delivered their milestones significantly ahead of schedule. (I am not making this up, this is how R&D timelines and budgets are reviewed and assessed)
 

sandorski

No Lifer
Oct 10, 1999
70,789
6,349
126
Originally posted by: larciel
Originally posted by: sandorski
If IBM were to buy AMD, I doubt they'd get back into the PC making Business. They could simply keep AMD as is, meaning as a Processor Design/Manufacturer for other PC Makers. IBM and AMD have had a close relationship for quite a long time now, each aiding the other especially regarding the Manufacture processes of CPUs. It would be a good fit.

Then we'd be back to days where paying $500-600 for mid-range CPU was normal.

I remember when I bought P2-350 for $170 back in, oh 98-99, next step up was P2 400 at $320 and P3 450 at $540. Right now, price diff between cpu models are mere $30-50 at most $120.

Why??
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
Intel hasn't worked with developing FinFet if so I would very much like a link to it.

So you are saying that Intel is going to do 32nm the same as 45nm. NO ultraviolet lithography. I thought either this process or immersion process was going to be employed at 32nm.
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
http://www.intel.com/pressroom...eases/20040802tech.htm

If the above is indeed true I don't understand what your talking about its simply a matter . If I were a stock holder of importance If I spent the money on this kind of equipment and development . I would dazel my investors with a new Tri gate processor and have my investors thinking and the hits just keep coming. I haven't seen were Intel is loosing $$.
 

Fallen Kell

Diamond Member
Oct 9, 1999
6,215
537
126
IBM wouldn't buy AMD. They are trying to get out of the hardware development market and into a pure services market. This is why they sold their hard drive lines, years ago to Hitachi, its laptop lines a couple years later to Lenovo, and have been scaling down thier PowerPC CPU lines (basically they told Apple to switch to Intel because they didn't want to continue making PowerPC chips in the future).

Cell is one of the last things they produce, and I am sure they are just waiting for Sony to make a reasonable offer and purchase it, however Sony doesn't want to be in the CPU business themselves either, so, IBM will continue holding onto that tech until someone makes a decent offer.

Again, IBM has slowly positioning themselves as a "services" company. They want to be your go-to guy for anything computer/IT related for your company, be it how to build a super-computer, to designing server room and data centers, to actually doing to work. Hardware has become too small of a profit margin now, where-as services continue to allow very large profit margins on just about everything you do.
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
Ya I read that bit about IBM telling Apple to shop somewhere else for Cpus. Problem is that story was after Apple inked the deal with intel.

Same as when Intel announced High K/ metal gates. IBM pops UP and announces YA we have that to. Anyone seen that chip. IBM thinks its somethinking special and they have caused so much more harm than good. Ya want to do some good reading . Read about IBMs activities durring WW2 that blow your sucks off.