backtop


Print 74 comment(s) - last by Penti.. on Mar 23 at 10:57 PM

Just wants to protect itself from the world

Intel is not trying to shut down AMD or GlobalFoundries, a source inside Intel Corporation told DailyTech in a phone conversation. The chip giant is simply trying to protect its intellectual property, which it doesn't want to see distributed without its control.

"Whenever a company uses technology developed by Intel, it reinforces Intel's technology leadership position", our source stated, speaking in an unofficial capacity.

AMD recently spun-off its manufacturing fabs and support staff into GlobalFoundries. This new joint venture with ATIC (Advanced Technology Investment Company) is expected to become a major player in the foundry market, which produces semiconductor chips from integrated circuit designs for many technology firms, including Texas Instruments and NVIDIA.

Intel believes that GlobalFoundries is not a subsidiary under the terms of a 2001 patent cross-licensing agreement, and is therefore not licensed to produce CPUs that use key technologies licensed under the agreement.

In a filing with the Securities and Exchange Commission, AMD stated Intel "purports to terminate the Company's rights and licenses under the Cross License in 60 days if the alleged breach has not been corrected."

In the phone conversation with our source, references were made to IP licensing deals with companies such as Taiwan Semiconductor Manufacturing Corporation (TSMC), which recently signed an agreement with Intel allowing manufacturing of Atom derivatives, albeit on a limited scope and scale.

Our source also noted that although Intel is suing NVIDIA over its chipset license in a separate case, that doesn't mean that Intel isn't willing to negotiate a deal allowing the graphics company access. He stated, "The first step is for NVIDIA to recognize our IP. Then we can sit down and work out a deal".

Apparently Intel is open to all sorts of deals, as another source told DailyTech that a licensing deal with GlobalFoundries is possible, if only they would acknowledge that they are a third party and not a subsidiary of AMD, and that they need to broker a separate agreement.



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

The plan
By skagon on 3/19/2009 11:07:26 AM , Rating: 2
So, that 's what they're trying to do, after all...
They just want to cut a new deal with GlobalFoundries, so that they'll end up having to pay double license fees to Intel; once from AMD for designing x86 CPUs and yet again from GF for actually manufacturing x86 CPUs.
I guess that would pretty much make it impossible for AMD to market any cost-effective CPU, since its performance would have to be stellar, compared to Intel products, to actually be economically viable.




RE: The plan
By just4U on 3/19/2009 11:20:34 AM , Rating: 4
OR, it's possible that Intel just wants to ensure that GF doesn't make X86 cpu's for anyone but Amd (without brokering a new deal that is) If that's the case, you'd think it would be a given that all sides understand that.. just maybe it needs to be in writing. (dunno)


RE: The plan
By omnicronx on 3/19/2009 11:32:44 AM , Rating: 4
I doubt it, AMD's x86 license is non transferable. Meaning that Nvidia cannot obtain a license from AMD, nor will the license be transfered if Nvidia were to buy AMD outright. The only way this could happen is if AMD were to buy Nvidia, but what would be the point of starting up an x86 line when you already have one..


RE: The plan
By TheFace on 3/19/2009 11:45:39 AM , Rating: 3
Or if Nvidia buys VIA.


RE: The plan
By omnicronx on 3/19/2009 12:19:41 PM , Rating: 2
Which they have been trying to do for the past few years unsuccessfully. My guess is they are also in the same boat as AMD, they probably can't transfer their license.

The only way I could see it happening is if there was an Nvidia/Via joint venture.


RE: The plan
By omnicronx on 3/19/2009 12:26:49 PM , Rating: 2
Forgot Via acquired their license upon the acquisition of Cyrix.. so I guess an Nvidia acquisition is still possible


RE: The plan
By phu5ion on 3/19/2009 2:25:27 PM , Rating: 2
I don't know, something tells me they are trying to "double-dip that chip".


RE: The plan
By Spectator on 3/19/2009 3:22:44 PM , Rating: 2
I thought Intel and AMD were all working together in the background anyways.

Intel shut down/sell fabs. Amd suddenly open a new Fab. Backed by Saudi.. Hmmm where did intel borrow capital from Saudi?. lol

Its all the same sht. May as well be price fixing LCD screens. if you want fastest you buy Intel. If you want cheapest buy Amd.

Im more curious to see if likes of Nvid change this x86 old guard strangle hold.

If Saudi has billions invested in both. its.. win/win for them and they have a say in both(non direct price fixing).


Didn't AMD have a third party Founder make cpus?
By Adul on 3/19/2009 1:39:54 PM , Rating: 2
Was it TSMC and/or IBM that AMD had make CPUs for them when they where at capacity and couldn't satifisy market demand. How come intel didn't have a problem with this then?




By FITCamaro on 3/19/2009 2:06:13 PM , Rating: 2
Because those foundries probably have a license....


RE: Didn't AMD have a third party Founder make cpus?
By VeauX on 3/19/2009 2:57:14 PM , Rating: 2
and that's what everyone wants to know , do they have a license or not.


By omnicronx on 3/19/2009 3:34:41 PM , Rating: 3
They do.. Both TSMC and IBM have manufacturing rights to the x86 process. Although I think TSMC only has manufacturing rights, they can't design chips.

I think UMC and Fujitsu also have manufacture only rights also.

Also, IBM, SGS-Thomson and Texas Instruments all had the ability to manufacturer and sell products designed by others with x86 licenses under their own name.

There are a handful of companies that can design x86 processors, but so many companies have stoped production and sold off their licenses that it is hard to keep track anymore.
I know companies like NEC and Transmeta used to have this ability, but they no longer produce x86 chips. I wonder if their licenses are still valid.


By JumpingJack on 3/19/2009 8:21:25 PM , Rating: 3
The answer to your questions is
a) Chartered Semiconductor in Singapore.
b) because the license agreement provides for a certain amount of outsourcing to be manufactured in fabs that are not owned by AMD, this does not ipso facto provide license to that part to make x86 CPUs for anyone.

The clause is in the publicly available redacted version:

(c) have made ***** AMD Licensed Products by another manufacturer for supply solely to AMD for use, import, sale, offer for sale or disposition by AMD pursuant to the license granted above in Section 3.2(a).

The key here is "solely to AMD for use", i.e. AMD can have **** % of their total production (what ever **** is) made by a third party, solely for AMD. THe third party cannot make any CPU for anyone else but AMD.

Intel's primary concern is that if the IP rights are transfered to GF, that any design house can design any x86 compatible CPU and have GF make them under license of GF.

My guess is that Intel will ultimately secure agreements that either a) GF only makes x86 CPUs for AMD under AMD's current license or b) grants GF licenses to make anyones x86 GPU for royalties provided by GF.

Who knows, just my guess anyway.


By JumpingJack on 3/19/2009 8:23:52 PM , Rating: 2
OOopos, sorry, :to make anyones x86 GPU for royalties provided by GF:

I mean "to make anyone's x86 CPU for royalties provided (or paid out) by GF.


Found something VERY interesting
By omnicronx on 3/19/2009 12:24:53 PM , Rating: 5
I don't know if this is the actual agreement, but I think it is.

Here is what a Subsidiary is outlined as in the contract.

1.22. "Subsidiary" shall mean any corporation, partnership, joint venture,
limited liability or other entity, now or hereafter, in which a
party

(a) owns or controls (either directly or indirectly) or originally
contributed (either directly or indirectly) at least fifty
percent (50%) of the tangible and intangible assets of such
entity; and

(b) owns or controls (either directly or indirectly) either of the
following:

(1) if such entity has voting shares or other securities, at
least fifty percent (50%) of the outstanding shares or
securities entitled to vote for the election of
directors or similar managing authority and such entity
is under no obligation (contractual or otherwise) to
directly or indirectly distribute more than seventy
percent (70%) of its profits to a third party, or

(2) if such entity does not have voting shares or other
securities, at least fifty percent (50%) of the
ownership interest that represents the right to make
decisions for such entity and an interest sufficient to
receive at least thirty percent (30%) of the profits
and/or losses of such entity.

(c) An entity shall be deemed to be a Subsidiary under this
Agreement only so long as all requisite conditions of being a
Subsidiary are met.


Now I am not a lawyer but it definitely seems like GlobalFoundries falls under the category of Subsidiary according to these conditions.

http://contracts.corporate.findlaw.com/agreements/...




By afkrotch on 3/19/2009 1:49:19 PM , Rating: 2
It's capable of being a subsidiary, if AMD says they want it to be one. If not, then it's not one.


RE: Found something VERY interesting
By tmouse on 3/23/2009 2:08:48 PM , Rating: 3
I see your point however it is possible that although AMD meets section "A" by originally contributing directly at least fifty percent (50%) of the tangible and intangible assets, AND they meet part of section "B" by having voting shares or other securities, of at least fifty percent (50%) of the outstanding shares. They could be failing the rest of part "B" "such entity is under no obligation(contractual or otherwise) to directly or indirectly distribute more than seventy percent (70%) of its profits to a third party" IF the future 6 billion cash infusion to GF from ATIC is actually in the original contract, since this probably would result in more than 70% of the profits going to ATIC (their already at 65.8%). I don't know what language is in the actual contract.


Must've missed something....
By nixoofta on 3/19/2009 1:44:42 PM , Rating: 5
....what does a gerbel tootin' on a meat trumpet have to do with the story?




RE: Must've missed something....
By FITCamaro on 3/19/2009 2:19:04 PM , Rating: 2
It's cute and distracting.


RE: Must've missed something....
By afkrotch on 3/19/2009 2:57:08 PM , Rating: 2
Gerbil = Intel

It's Intel tooting their own horn.


X64??
By SlyNine on 3/19/2009 3:37:09 PM , Rating: 1
Can you develope a X64 based chip without any of the X86 pipelines and have it be compatible with Windows 64 and other 64bit OS's? Or is the X64 pathways too dependent on X86?




RE: X64??
By GodisanAtheist on 3/19/2009 4:16:18 PM , Rating: 2
While I can't be 100% sure, the current X64 implementation of just about anything consumer based is merely an x86 extension. x86 is an instruction set rather than any particular way to design a processor. Anyone can make a CPU, whether or not it can carry out instructions based on x86 is a totally different story.

For example, Intel has their own 64 bit instruction set known as IA-64 which is used exclusively in their Itanium processors. It has virtually nothing to do with x86 and cannot run any sort of x86 (or x86-64) applications without a heavily emulated virtual environment.


RE: X64??
By wifiwolf on 3/19/2009 11:45:12 PM , Rating: 2
A bit more complicated that that.
The 64bit x86 is just a wrapped up x86 instruction set so they could extend it with new 64 bit instructions that can make use of true 64bit capabilities just like what happened when there was a transition from 16bit to 32bit.
I really would like to see anand making some "pc architecture anthology" too just so that people can realize how far the instruction set wars don't mean anything in practice but only on paper.


respect the IP
By mattclary on 3/19/2009 11:18:52 AM , Rating: 2
[cartman]You will respect my IP!!![/cartman]




RE: respect the IP
By FITCamaro on 3/19/2009 12:05:20 PM , Rating: 1
He's The Coon now. But don't tell anyone. Or what happened to Mysterio will happen to you.


RE: respect the IP
By nixoofta on 3/19/2009 1:36:11 PM , Rating: 1
Yeah,..couldn't believe they unmasked Mysterio. Even more surprising was who Mysterio actually was. Boy,...never saw that comin'.


Maybe my little fantasy, but...
By bupkus on 3/19/2009 3:44:52 PM , Rating: 2
I think I just had a small epiphany.

I have for so many years thought that capitalism has an Achilles' heel-- that being redundant research and resources. Now I see that through "cross licensing" competitors can share achievements just like software developers can share code modules.
Licensing fees and reciprocity bind two economic entities just as two nations that share trade invest in their peace.
The relationship between AMD and Intel although competitive seems a healthy one. Each trying to out trump the other but simultaneously cross licensing technologies in order to cheaply and expeditiously fend off or leapfrog the other.
I am certain, of course, that there are elements in each camp that would like to "engage in a battle of extinction."[Reference from the History Channel]
The resulting carnage would leave both wounded and vulnerable to another capital entity perhaps of foreign investors. [China comes to mind]

Is there a lesson here for our two political parties?




By BansheeX on 3/19/2009 8:30:05 PM , Rating: 2
quote:
Is there a lesson here for our two political parties?


No, they both deficit spend. They both have preemptive war adventures. The Dems have Vietnam, Repubs have Iraq. They both expand welfare schemes, Bush passed an 8 trillion prescription drug bill even though medicare was already underfunded by 20 trillion. They simply have no fiscal discipline, they largely work together and posture and grandstand when it suits them.


Why is it so hard to understand?
By TA152H on 3/19/09, Rating: -1
RE: Why is it so hard to understand?
By Fenixgoon on 3/19/2009 11:57:07 AM , Rating: 4
quote:
Intel makes the best processors in the world, excellent chipsets, and fine motherboards.


Do you happen to remember what processor line preceded the Core2 series? That's right, it was the Pentium 4, which *wasn't* the best processor in the world. The Athlon64 was.

I'm highly interested in AMD at this point because they offer a *complete solution*:

CPU
Motherboard/Chipset
Video (both discrete and integrated)

I really think/hope DAAMIT will start coming together in the next few years, instead of being just ATI and AMD under the roof.


By FaceMaster on 3/20/2009 10:46:45 AM , Rating: 2
quote:
Do you happen to remember what processor line preceded the Core2 series? That's right, it was the Pentium 4, which *wasn't* the best processor in the world. The Athlon64 was.


What does that have to do with it? Yeah, we all know tha the Pentium 4's sucked. But that was 3 years ago. That's like still slating the Nvidia series for its FX 5800. Or Rocky Balboa for losing a match ages ago. Stay with the times, maaannnn


By mackintire on 3/19/2009 11:58:16 AM , Rating: 4
I'm speechless that was the biggest fanboy post I've read in quite a while.

Anyhow, Intel and AMD do have a symbiotic relationship. granted Intel is in much better position. AMD does keep Intel innovating, as they know if they do not AMD will be happy to take more market share.


By FITCamaro on 3/19/2009 12:04:09 PM , Rating: 2
Is there even a seem anywhere between the edge of Intel's ass and your neck?

Intel uses some of AMDs technology too. And its not like AMD uses x86 because its some kind of god-like technology. It's just what is still the dominant processor architecture. Could AMD or anyone else come up with something better? Of course. But unless Microsoft supports it, it'll never gain widespread acceptance.

So don't act like AMD, or anyone else for that matter, just sponges off Intel's genius.


RE: Why is it so hard to understand?
By c4xp on 3/19/09, Rating: -1
RE: Why is it so hard to understand?
By FITCamaro on 3/19/2009 2:09:08 PM , Rating: 2
While AMD had the first commercial dual core out, Intel was the first to bring them to us for cheap. And there were multi-core processors long before AMD came out with them.


By omnicronx on 3/19/2009 2:42:15 PM , Rating: 2
Well not long before ;) IBM was the first to do it with the POWER4 in 2001. In fact it was also 64bit. They also did not invent DDR or Hypertransport either, they were merely the first ones to implement it at the consumer level ;)

Not saying AMD did not force Intel to innovate, but most of their 'innovations' were merely previous ideas brought to the x86 process.


RE: Why is it so hard to understand?
By afkrotch on 3/19/2009 2:54:50 PM , Rating: 5
floating point bug for the original Pentium. Wasn't much of a problem. Hasn't been an issue since.

RDRAM. Not Intel. They simply choose to use it, as it was the fastest available memory. Had it 2 years before DDR even hit the market. I use to run a PC-800 and PC-1066 mobo. It wasn't until DDR-400 came to market, that it could perform close to PC-1066.

EM64T = Intel 64 = AMD64. It's made by AMD.

FSB. Seemed to work fine for everyone, including AMD. Latest Intel and AMD procs no longer use it of course.

x86-64. AMD and Intel were both working on their own 64 bit instructions. Intel's threw away compatibility with 32 bit and instead used software to provide the compatibility. Microsoft said they would only support one, they choose AMDs. I think cause it's less work for Microsoft.

DDR. Not AMD. It was used as video memory before it was ever used as system memory. AMD licensed both DDR and RDRAM. More ppl wanted DDR, so that's the way they went.

HyperTransport. It's a consortium. Partially AMD.

Multicore. AMD definitely invented it. About 4 years after IBM started shipping them.


RE: Why is it so hard to understand?
By omnicronx on 3/19/2009 4:06:17 PM , Rating: 5
quote:
RDRAM. Not Intel. They simply choose to use it, as it was the fastest available memory. Had it 2 years before DDR even hit the market. I use to run a PC-800 and PC-1066 mobo. It wasn't until DDR-400 came to market, that it could perform close to PC-1066.
Thats not really true, PC-800 RDRAM had a maximum bandwidth of 1600MBps, exactly the same as DDR-200 (in single channel mode). DDR266 has almost exactly the same bandwidth as PC-1066 RD-RAM (2100 DDR vs 2132 RDRAM). Thus DDR-333 is much faster than PC-1066 RD. Furthermore RDRAM was always dual channel, when dual channel DDR ram came along they wiped the floor with RDRAM chips. Furthermore Intel made an illadvised agreement with RAMBUS, they were not allowed to anything else but RDRAM for a certain period of time, regardless if its performance.
quote:
FSB. Seemed to work fine for everyone, including AMD. Latest Intel and AMD procs no longer use it of course.
AMD has not been using it since 2003, mostly because it was inferior technology. Intel actually owns the rights to the original HT design, but chose to develop their own method which is very similar called CSI, well 10+ years since the beginning of development and we are only seeing CSI now. Intel never thought they would be using FSB for this long.
quote:
More ppl wanted DDR, so that's the way they went.
More like people wanted nothing to do with RAMBUS, although I would tend to agree that people wanted the cheaper to produce DDR.
quote:
HyperTransport. It's a consortium. Partially AMD.
Yes but AMD was instrumental in the project ;) Without AMD there would have been no consortium.
quote:
Multicore. AMD definitely invented it. About 4 years after IBM started shipping them.
Hahaha, funny. Ya Multicore is anything but an AMD design, IBM's chip was even two core on die, and it was 64 bit ;)


RE: Why is it so hard to understand?
By afkrotch on 3/19/2009 4:43:06 PM , Rating: 2
quote:
Thats not really true, PC-800 RDRAM had a maximum bandwidth of 1600MBps, exactly the same as DDR-200 (in single channel mode). DDR266 has almost exactly the same bandwidth as PC-1066 RD-RAM (2100 DDR vs 2132 RDRAM). Thus DDR-333 is much faster than PC-1066 RD. Furthermore RDRAM was always dual channel, when dual channel DDR ram came along they wiped the floor with RDRAM chips. Furthermore Intel made an illadvised agreement with RAMBUS, they were not allowed to anything else but RDRAM for a certain period of time, regardless if its performance.


DDR wasn't available when Intel went Rambus. There was no DDR-200 to compete with. Rambus was the fastest available memory during then. Everything else was PC-133 or less.

After DDR overtook Rambus, Intel went DDR.

quote:
AMD has not been using it since 2003, mostly because it was inferior technology. Intel actually owns the rights to the original HT design, but chose to develop their own method which is very similar called CSI, well 10+ years since the beginning of development and we are only seeing CSI now. Intel never thought they would be using FSB for this long.


Some inferior technology. The C2Ds seem to wipe AMD's face to the floor with this inferior tech.

quote:
More like people wanted nothing to do with RAMBUS, although I would tend to agree that people wanted the cheaper to produce DDR.


I wanted Rambus, til the performance crown was lost. Then I wanted DDR. Rambus was ungodly expensive, while DDR was way cheaper.

quote:
Yes but AMD was instrumental in the project ;) Without AMD there would have been no consortium.


Everyone would just be using Intel's QPI instead.

quote:
Hahaha, funny. Ya Multicore is anything but an AMD design, IBM's chip was even two core on die, and it was 64 bit ;)


While Intel and AMD have created multiple technologies, I'd hardly say they created everything. More like the driving force in adoption for technologies they didn't create.


RE: Why is it so hard to understand?
By omnicronx on 3/19/2009 6:04:38 PM , Rating: 2
quote:
DDR wasn't available when Intel went Rambus. There was no DDR-200 to compete with. Rambus was the fastest available memory during then. Everything else was PC-133 or less. After DDR overtook Rambus, Intel went DDR.
Whats your point? I know RDram came out before DRR, most post has nothing to do with how Intel came to choose RDRAM. He claimed it was faster, when it really is not (they were basically equal for a long time, except for the 15-20% higher price for RDRAM). Furthermore Intel stayed with RDRAM until their exclusive contract with Rambus ended, it had absolutely nothing to do with DDR overtaking Rambus. If it was up to them, they would have gone DDR far sooner as AMD had a huge advantage back in the original Atlhon days.
quote:
Some inferior technology. The C2Ds seem to wipe AMD's face to the floor with this inferior tech.
Yes? What do you think separates the original C2D from iCore Nehalems? CSI which is Intels implementation of HT (i.e no more FSB) is the main reason for the improvements. There would be no on die memory controller and the extra bandwidth of DDR3 would be next to useless if they were still using FSB.
quote:
I wanted Rambus, til the performance crown was lost. Then I wanted DDR. Rambus was ungodly expensive, while DDR was way cheaper.
My comment about Rambus is about the company, not RDram. Some people still consider them the kings of patent trolling.
quote:
Everyone would just be using Intel's QPI instead.
Ok?
quote:
While Intel and AMD have created multiple technologies, I'd hardly say they created everything. More like the driving force in adoption for technologies they didn't create.
Can't disagree with you here, although I have no idea where this comment came from. I was commenting that none of the 'innovations' were really AMD's at all, as you said they were the driving force for adoption.


RE: Why is it so hard to understand?
By afkrotch on 3/19/2009 6:43:09 PM , Rating: 2
quote:
Whats your point? I know RDram came out before DRR, most post has nothing to do with how Intel came to choose RDRAM. He claimed it was faster, when it really is not (they were basically equal for a long time, except for the 15-20% higher price for RDRAM). Furthermore Intel stayed with RDRAM until their exclusive contract with Rambus ended, it had absolutely nothing to do with DDR overtaking Rambus. If it was up to them, they would have gone DDR far sooner as AMD had a huge advantage back in the original Atlhon days.


I guess you're not understanding what I'm saying.

It was the fastest available for it's time, so Intel choose it. Since DDR was not available, there was absolutely zero competition to match it's performance.

By no means did I say that it did not match it's performance, I simply said that it was the fastest available when Intel choose it.

quote:
Yes? What do you think separates the original C2D from iCore Nehalems? CSI which is Intels implementation of HT (i.e no more FSB) is the main reason for the improvements. There would be no on die memory controller and the extra bandwidth of DDR3 would be next to useless if they were still using FSB.


And? C2D with it's inferior technology seems to easily beat AMD and their newer Hypertransport.

As for on die memory controller, I'm not a big fan. Before it, you can swap to new memory by buying a new mobo and memory. Now you're ass is stuck buying a new processor to boot.

quote:
My comment about Rambus is about the company, not RDram. Some people still consider them the kings of patent trolling.


I'm partly for and against Rambus. I'm fine for creating new technologies and licensing them out, but the whole going to court over DDR2, DDR3, GDDR, and some other crap I'm not for.

quote:
Can't disagree with you here, although I have no idea where this comment came from. I was commenting that none of the 'innovations' were really AMD's at all, as you said they were the driving force for adoption.


I just threw that in for others, not so much for you.


RE: Why is it so hard to understand?
By omnicronx on 3/19/2009 7:16:37 PM , Rating: 2
quote:
By no means did I say that it did not match it's performance, I simply said that it was the fastest available when Intel choose it.
I agreed with you last time, I was only implying that low timed DDR333 performed very closely to PC1066 RDRAM. I should not have mentioned it in the first place, it adds nothing to the conversation. (and really you are correct, DDR400 was the first time that DDR truly eclipsed RDRAM in speeds)
quote:
And? C2D with it's inferior technology seems to easily beat AMD and their newer Hypertransport.
I never said C2D was inferior technology, I said FSB was. The C2D core is obviously miles ahead of AMD offerings, but it is not because of FSB, its because of great design. I've already explained that most of the performance increases with the iCore series is because of CSI.
quote:
As for on die memory controller, I'm not a big fan. Before it, you can swap to new memory by buying a new mobo and memory. Now you're ass is stuck buying a new processor to boot.
You've lost me.. why can't you still swap in a new mobo and memory?


RE: Why is it so hard to understand?
By cheetah2k on 3/19/2009 9:42:32 PM , Rating: 1
Hmmmm the days of Rambus memory...

.. And thank God DDR memory was created.

If it wasnt, we'd still be locked into paying US$1000 for a 1Gb stick of the Rambus shyte... Sure it was quick for its time, but by golly, it was just too expensive.


By afkrotch on 3/20/2009 12:32:07 PM , Rating: 2
Rambus was ridiculous. I paid $600 for four 256 meg sticks of PC-1066.

I think the reason for it being expensive is because no one bothered to make it. I only saw Samsung Rambus sticks, no one else.

We actually still have some Dell Precision 650s (or 630s, can't remember the number) here at work that use Rambus.


By afkrotch on 3/20/2009 12:25:55 PM , Rating: 2
quote:
I never said C2D was inferior technology, I said FSB was. The C2D core is obviously miles ahead of AMD offerings, but it is not because of FSB, its because of great design. I've already explained that most of the performance increases with the iCore series is because of CSI.


I don't consider FSB to be inferior. Doesn't work for certain purposes yes. FSB simply blows for multiple core processor, due to what they need. It works splendidly for single cores and doesn't need any replacement for them.

quote:
You've lost me.. why can't you still swap in a new mobo and memory?


I mean new type of memory. Let's say I wanted to swap to DDR3 on my C2D system. I buy a new mobo and new memory. I'm good to go. With something like AMD, you'd need a new mobo, new memory, and new proc. Unless someone makes a board with both memory slot types and those type of board are usually the crap.

I'm the type to make small incremental upgrades. That's why it took me forever to move off the 478 P4s. I ended up having to get a new proc, mobo, memory, and vidcard.

Once or twice a year I make minor upgrades. The $400 or less each time roughly. I hate blowing more than that.


RE: Why is it so hard to understand?
By Penti on 3/19/2009 4:27:18 PM , Rating: 2
Microsoft still supports Itanium. And it's in the marketplace where it was intended for. NT also ran on MIPS and Alpha initially. That there where working on different 64-bit processors is moot when other like those MIPS and Alpha was already 64-bit.


RE: Why is it so hard to understand?
By afkrotch on 3/19/2009 6:50:38 PM , Rating: 2
Intel was working on it's own x86-64, but it was scraped when Microsoft went with AMD's x86-64 instruction set.


RE: Why is it so hard to understand?
By Penti on 3/23/2009 10:57:35 PM , Rating: 2
And that's not unexpected as 64 bit Opterons had been out for almost a year when they even announced they where working on a extension. Wheres the AMD64 specs had been around for years, with software ready before launch (linux mainly). It's not really anything you can do in secret they didn't with IA-64! By that time it was already x86-64 or EM64T as we came to know it then called IA-32e internally and it didn't trow away compatibility with 32 bit mode as it was an extension.

It's not surprising AMD won as none not even an ISA/registers came out from Intel. IA64 didn't btw drop x86 support till 2006. When it moved in to software. So what makes you say that they planned to make a desktop chip not compatible with 32-bit x86? I know Otellini talked about not releasing a 64-bit compatible with AMD for a while, that wasn't in the context of desktop processors but that don't explain the bit about software emu.

You already had Microsoft 64-bit compilers for AMD64 and Win64 before Intel even admitted to theirs thus they was running the IA-32e development silicon with the software developed for AMD64. They had even released a public preview of XP 64-bit days before the IDF event.

Conclusion your still talking about IA-64 and it's emulated x86 (which where first in hardware later removed to be replace by software only emulation).


RE: Why is it so hard to understand?
By SlyNine on 3/19/09, Rating: -1
RE: Why is it so hard to understand?
By Xerachiel on 3/19/2009 5:03:15 PM , Rating: 1
Psst, check this out
http://www.top500.org/lists/2008/11

1. (AMD) Opteron
2. Cray XT5. Powered by Opteron
3. Xeon. Phew, saved Intel's day
4. BlueGene/L. Powered by PowerPC
5. BlueGene/P. Powered by PowerPC
6. SunBlade. Powered by Opteron
7. Cray XT4. Powered by Opteron
8. Cray XT4. Powered by Opteron
9. Cray XT3/4. Powered by Opteron
10. Opteron *yawn*

Gee, that single Intel machine must feel lonely there, behind two opteron-powered machines and followed by 2 PowerPC IBMs and 5 more opteron-powered machines. AMD-Intel: 7-1. And unlike Intel, there's little difference between AMD's server and PC parts.


RE: Why is it so hard to understand?
By afkrotch on 3/19/2009 6:58:44 PM , Rating: 2
Umm...the number 1 spot is the Roadrunner. It's mostly Cell procs.

Course I'm not surprised the Opteron is uses most often. Hypertransport makes the Opteron a better proc for 4-way or more servers.

I'm thinking the list will change over time with tie i7's QPI.


By omnicronx on 3/19/2009 7:42:50 PM , Rating: 2
quote:
Umm...the number 1 spot is the Roadrunner. It's mostly Cell procs.
Theres a slash in there, it uses both.. Its a bunch of IBM BladeCenter LS21's and QS22 clustered.

The LS21 is Opteron based.. the QS22 is Cell based.


By Baladen on 3/20/2009 8:19:04 AM , Rating: 2
I doubt the Intel machine was feeling lonely, since it had 377 other Intel powered supercomputers on the top 500 list to keep it company in November 2008.

http://www.top500.org/stats/list/32/procfam


Yeah, right...
By Beenthere on 3/19/09, Rating: -1
RE: Yeah, right...
By omnicronx on 3/19/2009 11:36:42 AM , Rating: 1
So.. becoming a real monopoly will help them? Second of all, the case does not just disappear if AMD were to go out of business..


RE: Yeah, right...
By afkrotch on 3/19/2009 3:04:08 PM , Rating: 2
The case would disappear when they can't pay their lawyers anymore.


RE: Yeah, right...
By mmatis on 3/19/2009 4:54:28 PM , Rating: 2
You got that right! You sure... got that right!


RE: Yeah, right...
By finalfan on 3/19/2009 5:04:37 PM , Rating: 2
No, they can. They can spinoff some other useless stuff and make cash out of it by allowing others to access the stuff that they don't solely own.


RE: Yeah, right...
By afkrotch on 3/19/2009 7:01:50 PM , Rating: 2
If somehow AMD goes bankrupt and has no money to restructure, there's no way to pull money from anywhere.

Course not like that'll happen.


RE: Yeah, right...
By omnicronx on 3/19/2009 7:31:30 PM , Rating: 2
The EU in particular filed their own cases against Intel. They were merely following up on complaints by AMD. Same thing goes for the FTC's claims in the US. And Korea (25 million judgement against them has already been made).. the list goes on..

This will not go away if AMD were to go bankrupt. If anything it will add more fuel to the fire. This is an antitrust case, AMD is not suing Intel directly. (well they are such as the case that is set to begin in Delaware in 2010, but these antitrust cases are totally separate)


Third party
By omnicronx on 3/19/09, Rating: -1
RE: Third party
By ExarKun333 on 3/19/2009 12:08:24 PM , Rating: 3
You have it wrong; AMD only owns 34.2% of Global Foundries, that's what Intel is taking issue with.


RE: Third party
By Amiga500 on 3/19/2009 1:16:09 PM , Rating: 5
That does not mean they (Global Foundaries) are not a subsidiary...

Which is really what the argument is.

Intel on shakey ground IMO.


RE: Third party
By smegz on 3/19/2009 1:36:08 PM , Rating: 3
But, they retain 50% of the voting rights. And they contributed more than 50% of the assets. GF is a subsidiary by the definition provided in a previous post. I think we need to still wait to see what Intel's real motives are...because it's not about the IP and we all know that.


RE: Third party
By omnicronx on 3/19/2009 1:46:02 PM , Rating: 3
I'm pretty sure they own 34.2% when fully converted to common shares. While I was wrong about the 56%, AMD still holds 50% voting interest in GlobalFoundry. Which if you read the post above, is all they need according to my interpretation of the contract.


RE: Third party
By finalfan on 3/19/2009 2:40:07 PM , Rating: 5
See here http://www.amd.com/us-en/assets/content_type/Downl...

Page 7:

From an operating perspective, AMD will have a
commercial relationship with GLOBALFOUNDRIES, similar
to other third-party commercial foundry relationships

The voting rights AMD touted is just a paperwork. It will eventually conflict with the voting rights from common share ownership.

If AMD succeeds in this case, any company that has cross-license agreement can choose to create a new shell company by partnering with another company that wants the access to the un-transferable licenses and in turn get cash from that. For example, VIA can just create another company and parter with NVidia. VIA gets cash and NVidia gets x86 license while Intel can say nothing about it.


RE: Third party
By omnicronx on 3/19/2009 3:08:06 PM , Rating: 2
quote:
If AMD succeeds in this case, any company that has cross-license agreement can choose to create a new shell company by partnering with another company that wants the access to the un-transferable licenses and in turn get cash from that. For example, VIA can just create another company and parter with NVidia. VIA gets cash and NVidia gets x86 license while Intel can say nothing about it.
The Cyrix license in which VIA holds states they are only allowed to produce their chips in a foundry that already holds an Intel license. I.e its pretty much useless in the way you described.

VIA can partner with Nvidia already regardless, this case changes nothing in that respect.


RE: Third party
By finalfan on 3/19/2009 3:42:30 PM , Rating: 2
Of course this means a lot. A compnay now can just write down "I own 50% of voting right" and then "sell" the cross-license to any other company without really "owning" the shell company.


RE: Third party
By omnicronx on 3/19/2009 4:17:59 PM , Rating: 2
Who is going to do this? AMD can't spin off of Globalfoundry, because they would control less than 50% of the second spin off.

Via could, but then whoever bought them would not be able to produce their own chips, they would have to use one of 4 foundries with the capabilities to do so. All of which have major ties to Intel. Furthermore they would be at a huge disadvantage as they will never be able to own their a foundry to produce x86 chips, nor would they be able to have partial control ala AMD/GF.

IBM? I don't see why they would want to re-enter the market (they could have bought out AMD a long time ago if they wanted to).

SIS? It would make more sense to just buy them out.

There really is nobody else.


RE: Third party
By finalfan on 3/19/2009 4:51:17 PM , Rating: 2
You are still assuming AMD is honestly owns 50% of the spinoff. It doesn't. It counts its patent as part of the investment. And at the same time, it doesn't loose the patent it invested in the new company. In that sense, it just copied its IP and license and count the copy as investment. And the value of that part of investment can be determined by their own.

If it can get away from that, it will be able to create another company and spin off any crap as investment along with its patent/license and make cash out of it. In theory, it can do this over and over again without actually investing much out of its pocket.

If you read my previous post, from their own presentation, AMD is telling people the new company is no difference than a thrid party company. Even it doesn't impose any legal obligation with that, it does show AMD is doing this without good faith by creating a loophole here.


RE: Third party
By omnicronx on 3/19/2009 6:37:12 PM , Rating: 2
quote:
You are still assuming AMD is honestly owns 50% of the spinoff. It doesn't. It counts its patent as part of the investment. And at the same time, it doesn't loose the patent it invested in the new company. In that sense, it just copied its IP and license and count the copy as investment. And the value of that part of investment can be determined by their own.
But it currently controls 50%, and will continue to do so. The second they do not control 50% they will breach their agreement with Intel and they will no longer be able to produce x86 chips if Intel so chooses.
quote:
If it can get away from that, it will be able to create another company and spin off any crap as investment along with its patent/license and make cash out of it. In theory, it can do this over and over again without actually investing much out of its pocket.
I've already explained why that can't happen. A spinoff of a spinoff won't give AMD 50% control it needs to consider them a subsidiary. AMD is already in shallow waters here, there is no way they could legally pull off what you are saying
quote:
If you read my previous post, from their own presentation, AMD is telling people the new company is no difference than a thrid party company. Even it doesn't impose any legal obligation with that, it does show AMD is doing this without good faith by creating a loophole here.
Really it does not matter what their presentation says, or if they consider it a third party. The agreement explicitly outlines what a subsidiary is, there is no getting around this. The contract trumps all, we all know that really GlobalFoundries is a really third party, but agreement states otherwise, which is exactly the card AMD is trying to play here. Their goal is not to give themselves more competition, it would hurt them far more than Intel.


RE: Third party
By CyberHawk on 3/19/2009 4:39:46 PM , Rating: 2
quote:
but ATI holds a 56% stake in Globalfoundries


Unless I have been sleeping in some sort of cave for the last couple of years you are totally wrong on this one.


RE: Third party
By omnicronx on 3/19/2009 7:20:12 PM , Rating: 2
You should really read at least the first responding post before commenting ;) (first commenter pointed this out)

Then again, I should have checked my sources before posting in the first place..


"Young lady, in this house we obey the laws of thermodynamics!" -- Homer Simpson




Latest Headlines
2/10/2012 Daily Hardware Reviews
February 10, 2012, 5:50 PM
2/9/2012 Daily Hardware Reviews
February 9, 2012, 11:54 AM
2/8/2012 Daily Hardware Reviews
February 8, 2012, 1:11 PM
2/7/2012 Daily Hardware Reviews
February 7, 2012, 12:23 PM










botimage
Copyright 2012 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki