intel compatibility and open architecture

Bob

Distinguished
Dec 31, 2007
3,414
0
20,780
Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

Is intel's microprocessor architecture open one?
If not How come cyrix and via and amd make their processor
*compatible* with the intel microprocessor and moreover market them
with an adjective "compatible with intel"?
don't intel take objection to that?



greetings,
Bob
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

On 4 Oct 2004 09:24:11 -0700, bob_peterson@rediffmail.com (Bob) wrote:
>
>Is intel's microprocessor architecture open one?

How do you define "open"?

>If not How come cyrix and via and amd make their processor
>*compatible* with the intel microprocessor and moreover market them
>with an adjective "compatible with intel"?
>don't intel take objection to that?

Not so long as AMD and VIA keep paying Intel their licensing fees.
There are rather large and complicated cross-licensing agreements
between the companies that involve a fair money and intellectual
property changing hands.

Of course, Intel isn't 100% happy with the agreement, but they don't
have much choice. One of the legal rulings that came out of a rather
lengthy court battle they had with AMD back in the late '80s/early
'90s was that Intel was REQUIRED to license their x86 technology to
other companies at reasonable rates. There is some speculation that
this is part of the reason why Intel and HP set up a rather
complicated separate entity for their IA-64 instruction set (used in
the Itanium), such that they would not have to license this
instruction set to other companies if it took off like they had
expected (of course, given how badly Itanium has bombed so far, this
hasn't been an issue... no company in their right mind would WANT to
build an Itanium-compatible processor!).


So.. does that make it "open"? Are the ARM or MIPS architecture's
"open"? There are dozens of manufacturers for each of those
architectures, but they still require some licensing fees. Really
SPARC is probably the only truly open architecture out there today,
and even that has some restrictions (in particular you can't use the
"SPARC" brand name unless you pay some sort of royalties).

-------------
Tony Hill
hilla <underscore> 20 <at> yahoo <dot> ca
 

Bob

Distinguished
Dec 31, 2007
3,414
0
20,780
Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

> Not so long as AMD and VIA keep paying Intel their licensing fees.
> There are rather large and complicated cross-licensing agreements
> between the companies that involve a fair money and intellectual
> property changing hands.

AMD and Via both are leading competitors of Intel and AMD is not
enchroaching the intel's market.
How intel sells its technology to its leading competitors?
what is the business motivation behind this?


greetings,
Bob
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

Bob wrote:
>> Not so long as AMD and VIA keep paying Intel their licensing fees.
>> There are rather large and complicated cross-licensing agreements
>> between the companies that involve a fair money and intellectual
>> property changing hands.
>
> AMD and Via both are leading competitors of Intel and AMD is not
> enchroaching the intel's market.
> How intel sells its technology to its leading competitors?
> what is the business motivation behind this?

In some cases, it's due to a court-order (i.e. AMD). In other cases, it's
due to a historical alliance between Intel and another company (i.e. IBM).
Or in other cases, it could be just a straight exchange for royalties.

Yousuf Khan
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

On 5 Oct 2004 06:04:56 -0700, bob_peterson@rediffmail.com (Bob) wrote:
>
>> Not so long as AMD and VIA keep paying Intel their licensing fees.
>> There are rather large and complicated cross-licensing agreements
>> between the companies that involve a fair money and intellectual
>> property changing hands.
>
>AMD and Via both are leading competitors of Intel and AMD is not
>enchroaching the intel's market.
>How intel sells its technology to its leading competitors?
>what is the business motivation behind this?

As mentioned in the previous message, I don't think that Intel is too
happy about the arrangement, but they don't have much choice. See, we
have these pesky things called laws that prevent anti-competitive
behavior among companies. If Intel were to cut off all licensing to
other companies the FTC and various other 3-letter organizations
around the world would come down on them like a ton of bricks (much
like what should happen with Microsoft if these organizations had the
balls to do so).

Note that the flow of intellectual property is definitely not a
one-way street between Intel and AMD. There are a number of
innovations that AMD has made over the years that Intel licenses from
them. Of recent interest is their AMD64 extension to the old IA-32
instruction set. However even before this there have been various
technologies (big or small) that Intel has licensed from AMD. It may
surprise you to know that AMD has actually been granted more patent
each year for the past 3 or 4 years than Intel has, and those patents
give them a bit of leverage.

VIA is a slightly more complicated issue, and some of it ties in to
agreements that Intel has previously with IDC and Cyrix (both of whom
were bought out by VIA), as well as S3 (who kind of merged with VIA).
For the most part though I think Intel is happy enough to not get too
involved in legal battles on this one though since VIA is such a small
player (less than 1% of the market). Same goes for Transmeta, who
will almost certainly soon go under/be bought out by someone anyway.

-------------
Tony Hill
hilla <underscore> 20 <at> yahoo <dot> ca
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

Tony Hill wrote:
> player (less than 1% of the market). Same goes for Transmeta, who
> will almost certainly soon go under/be bought out by someone anyway.

No sooner said, than voila, first rumours:

http://www.theinquirer.net/?article=18984

Yousuf Khan
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

On Sun, 10 Oct 2004 00:27:09 -0400, "Yousuf Khan" <bbbl67@ezrs.com>
wrote:
>
>Tony Hill wrote:
>> player (less than 1% of the market). Same goes for Transmeta, who
>> will almost certainly soon go under/be bought out by someone anyway.
>
>No sooner said, than voila, first rumours:
>
>http://www.theinquirer.net/?article=18984

There ya go, though I don't think this is the "first" rumor. People
have been speculating that Transmeta might be bought out ever since
their first Crusoe processors started getting dropped by all the
hype-riding customers. As soon as people figured out that the
performance of these chips was abysmal, they kind of lost interest in
their low-power characteristics.

Interesting that they mention nVidia is the most likely buyer though.
That's the first time I've heard this tidbit. I suppose it's not
entirely out of line, though it doesn't exactly strike me as the
smartest move. Mind you, maybe they have something up their sleeves
that I haven't heard of yet.

-------------
Tony Hill
hilla <underscore> 20 <at> yahoo <dot> ca
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

Tony Hill <hilla_nospam_20@yahoo.ca> wrote in message news:<2nvlm0p6ppp4li42ffaej9sr5ufgh4omql@4ax.com>...
> Interesting that they mention nVidia is the most likely buyer though.
> That's the first time I've heard this tidbit. I suppose it's not
> entirely out of line, though it doesn't exactly strike me as the
> smartest move. Mind you, maybe they have something up their sleeves
> that I haven't heard of yet.

It wouldn't make much sense to me either. It's not the first time that
a big-name company came to the aid of a struggling x86 manufacturer
(e.g. National Semi buys Cyrix, then VIA buys Cyrix and Centaur).
NatSemi never made Cyrix a success ever again. I can't see Nvidia
doing much to shore up Transmeta. Basically, Nvidia-Transmeta would
have to live in the shadow of AMD, much like VIA-Cyrix does right now,
both supporting AMD and competing against it using AMD-derived
technology (e.g. AMD64 & Hypertransport). Via has said that their next
x86 CPU will have both AMD64 and Hypertransport, as well as an
integrated RAM controller.

Actually, it's kind of interesting nowadays, people are really making
AMD-compatible processors, not Intel-compatible. Transmeta just
recently added the NX-bit support, and it already has Hypertransport
support. Probably because Intel is jealously guarding its technology,
and AMD is happily sharing its.

Yousuf Khan
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

On 11 Oct 2004 20:44:03 -0700, yjkhan@gmail.com (ykhan) wrote:
>
>Tony Hill <hilla_nospam_20@yahoo.ca> wrote in message news:<2nvlm0p6ppp4li42ffaej9sr5ufgh4omql@4ax.com>...
>> Interesting that they mention nVidia is the most likely buyer though.
>> That's the first time I've heard this tidbit. I suppose it's not
>> entirely out of line, though it doesn't exactly strike me as the
>> smartest move. Mind you, maybe they have something up their sleeves
>> that I haven't heard of yet.
>
>It wouldn't make much sense to me either. It's not the first time that
>a big-name company came to the aid of a struggling x86 manufacturer
>(e.g. National Semi buys Cyrix, then VIA buys Cyrix and Centaur).
>NatSemi never made Cyrix a success ever again.

NatSemi drove Cyrix into the ground from which it never recovered.
The only thing that survived for VIA to buy was the name, and even
that was dropped in fairly short order.

> I can't see Nvidia
>doing much to shore up Transmeta. Basically, Nvidia-Transmeta would
>have to live in the shadow of AMD, much like VIA-Cyrix does right now,
>both supporting AMD and competing against it using AMD-derived
>technology (e.g. AMD64 & Hypertransport). Via has said that their next
>x86 CPU will have both AMD64 and Hypertransport, as well as an
>integrated RAM controller.

I really can't see nVidia (or anyone for that matter) buying out
Transmeta and trying to continue pushing it as a direct competitor to
low-powered mobile chips for laptops and such. The only thing I can
think of is that maybe someone will be interested in some of the IP
and technology that they have. From an academic standpoint Transmeta
does have some neat stuff, it just happens to be rather useless in the
real-world. However perhaps they just haven't found the right niche
to push that technology into.

Who knows, maybe the JIT translation to VLIW code will end up being
much more effective in designing graphics chips?! I don't really
expect it to be, but you never know, stranger things have happened.

>Actually, it's kind of interesting nowadays, people are really making
>AMD-compatible processors, not Intel-compatible. Transmeta just
>recently added the NX-bit support, and it already has Hypertransport
>support. Probably because Intel is jealously guarding its technology,
>and AMD is happily sharing its.

There's a lesson to be found in here...

-------------
Tony Hill
hilla <underscore> 20 <at> yahoo <dot> ca