AMD Breaking up

AMD didn't sell x86 clones until 1991, when they released their Am386DX (after a long legal battle with Intel, resulting in the first x86 cross-license. It was renewed in 2001, and that one is still valid as far as I know).
AMD did produce x86 processors before that, but that was Intel outsourcing part of the 8086 and 80286 production because of a clause in the deal they made with IBM for their PC line (IBM insisted on having multiple suppliers for CPUs to reduce the risk). Obviously AMD didn't design these CPUs, but just manufactured them, and they were still sold under the Intel brand (although most of them do have AMD's name on them).
AMD didn't design the 386 and 486 they sold either, they reverse-engineered the Intel parts (Intel was never keen on outsourcing production and giving others their designs, they could get out of the deal with IBM in the 386 era when many alternative IBM PC-clones had arrived).

my point still stands... when exactly was it that Intel jacked up prices discounting the natural progression of cheaper to make and faster electronics?

besides, the way the market is now... with high level programming languages being used *almost* universally, an industry wide shift to another instruction set... while it would be painful, is possible.... both windows and osx have ports


intel's x86 ownership is *not* a trump card...
 
Creating a new CPU that can compete with Intel cant be done by a few nerds in the basement.

Still there are various alternative CPU designs. Some of them are very successful, such as the ARM which is used in many mobile devices, such as phones.

The main problem is how the world is dependent on x86. Even Intel faced those problems head-on when they tried to introduce the Itanium.
Chances are that when AMD goes belly-up, people will be more motivated to pry themselves loose from the claws of x86, which is better for everyone.

I'd like to compare it to the GPU-world. The shaders aren't tired directly to a physical instructionset, so both nVidia and ATi can design their own architectures, and this allows for more variation, more efficient designs and faster evolution of GPUs.
 
my point still stands... when exactly was it that Intel jacked up prices discounting the natural progression of cheaper to make and faster electronics?

Well, theoretically that should have been between 1978 (introduction of the 8086) and 1991 (the introduction of AMD's Am386).
But there doesn't seem to be any proof supporting that notion, so I agree with you.

besides, the way the market is now... with high level programming languages being used *almost* universally, an industry wide shift to another instruction set... while it would be painful, is possible.... both windows and osx have ports

I'd go further than that and say it's *required*.

intel's x86 ownership is *not* a trump card...

Well, it is, but it shouldn't be. In fact, if it wasn't for AMD and its x86-64, we'd probably be using Itanium now, and we'd already gone through the proces of shifting to another instructionset (and probably not making the same mistake of painting ourselves in the corner yet again).
In which case it'd be a good example of why competition isn't always good. You can throw Betamax in as an example aswell. Sometimes the better product doesn't win because of too much competition.
 
Well, it is, but it shouldn't be. In fact, if it wasn't for AMD and its x86-64, we'd probably be using Itanium now, and we'd already gone through the proces of shifting to another instructionset (and probably not making the same mistake of painting ourselves in the corner yet again).
In which case it'd be a good example of why competition isn't always good. You can throw Betamax in as an example aswell. Sometimes the better product doesn't win because of too much competition.


the rise of Linux, smartphones, Google, and OS X have significantly changed the landscape from 2003 when AMD64 was release en masse.

these technologies were not nearly as resurgent then as they are now, and while the business world remains a solidly Windows world, it is not necessarily a Wintel world. Intel and MS have had a strained relationship as of late, with Intel's large backing of Linux for low powered devices (see it's MID concepts).... and given the proper incentive, even the business world could break with tradition... especially if going forward the biggest leaps in technology (from a software standpoint) come from systems running a non-x86 instruction set...

in this respect, google android/linux netbooks/*nix servers/cell phones in general/"cloud computing" (though I hate it) are all very interesting, and make an apples to apples comparison between 2008 (almost 2009!) and 2002/3 impossible.....
 
the rise of Linux, smartphones, Google, and OS X have significantly changed the landscape from 2003 when AMD64 was release en masse.

Not really though.
They didn't really replace the current Windows/x86 systems. Windows is still on more than 90% of all PCs, and even most linux desktops/servers are x86-based (quite a lot of linux software is also pretty locked in the x86 world, and the linux kernel itself is far from a good example of a portable design. Its x86-roots are quite apparent. FreeBSD or NetBSD would be a better example, but somehow they never caught on).
 
AMD didn't sell x86 clones until 1991, when they released their Am386DX (after a long legal battle with Intel, resulting in the first x86 cross-license. It was renewed in 2001, and that one is still valid as far as I know).
AMD did produce x86 processors before that, but that was Intel outsourcing part of the 8086 and 80286 production because of a clause in the deal they made with IBM for their PC line (IBM insisted on having multiple suppliers for CPUs to reduce the risk). Obviously AMD didn't design these CPUs, but just manufactured them, and they were still sold under the Intel brand (although most of them do have AMD's name on them).
AMD didn't design the 386 and 486 they sold either, they reverse-engineered the Intel parts (Intel was never keen on outsourcing production and giving others their designs, they could get out of the deal with IBM in the 386 era when many alternative IBM PC-clones had arrived).

When you license something from some one, in this case, where AMD was licensed to create x86 chips by Intel, you don't need to reverse engineer it. Thats the whole point in licensing. You are correct the reason Intel licensed them was because of pressure from IBM to have more than one manufacture for their CPU's. All of the "clone" CPU's AMD manufactured were not reverse engineered. After the 386/486 AMD decided to design their own chips under the x86 license and then became a true competitor. While operating x86 instructions, AMD's chips have very little else in common (in terms of architecture) of Intel counterparts of that era.

Whats more important is Intel's reaction to Nvidia's Tegra
 
Not really though.
They didn't really replace the current Windows/x86 systems. Windows is still on more than 90% of all PCs, and even most linux desktops/servers are x86-based (quite a lot of linux software is also pretty locked in the x86 world, and the linux kernel itself is far from a good example of a portable design. Its x86-roots are quite apparent. FreeBSD or NetBSD would be a better example, but somehow they never caught on).

well part of the reason why everything is running x86 right now is that Intel and AMD have done such a damn good job with it....

but, this hypothetical situation we are talking about... a dead AMD, and intel reacting to that by charging higher prices, slowing innovation, etc....

that would be the missing catalyst for a big migration away from x86 for sure.... and like I said, the landscape today is so very much more dynamic and different than it has ever been.... a catalyst like this would have consequences that nobody can really predict, other than that it would *not* be business as usual if AMD died....


and honestly... i think the linux kernel is a fine enough example of portability :) I personally have compiled it for an ARM processor and it worked quite well in the capacity that I used it for....
 
Whats more important is Intel's reaction to Nvidia's Tegra

I eagerly await Nvidia's Tegra with bated breath.... I hope they pull off a killer platform that is hugely successful.... we need Nvidia to do well, and we really need the non-x86 world to keep pounding x86 into the ground like it is today when it comes to mobile devices
 
When you license something from some one, in this case, where AMD was licensed to create x86 chips by Intel, you don't need to reverse engineer it. Thats the whole point in licensing.

Wrong. The license only allows AMD to use the x86 instructionset. It doesn't give AMD access to Intel's designs.
If it did, AMD would be making Core2 processors now, wouldn't they?

All of the "clone" CPU's AMD manufactured were not reverse engineered.

Not true, the 386 and 486 were pretty much a transistor-for-transistor exact photocopy of Intel's chips (although AMD later made its own variations on the 486 independently of Intel). This is widely known and well-documented on the net.

After the 386/486 AMD decided to design their own chips under the x86 license and then became a true competitor. While operating x86 instructions, AMD's chips have very little else in common (in terms of architecture) of Intel counterparts of that era.

This is true, although ironically enough AMD rarely designed a complete chip.
Their K5 was their own design (although some parts, like the poor FPU, still based on the 486). This was not a very successful CPU.
They acquired NexGen and launched the architecture that NexGen worked on as the K6.
The K7 used a lot of technology from the Alpha processor bought from Digital/Compaq when they went belly-up.
The K8 was little more than a K7 with the memory controller moved onto the die. And the Phenom is a quadcore variation of that architecture.
 
and honestly... i think the linux kernel is a fine enough example of portability :) I personally have compiled it for an ARM processor and it worked quite well in the capacity that I used it for....

That's because someone already did the hard work of porting it to ARM for you.
That doesn't mean that the porting itself was easy to do, because it wasn't.
And even once you have ported the kernel, that doesn't guarantee that all linux software will work.
 
Their K5 was their own design (although some parts, like the poor FPU, still based on the 486). This was not a very successful CPU.
They acquired NexGen and launched the architecture that NexGen worked on as the K6.
The K7 used a lot of technology from the Alpha processor bought from Digital/Compaq when they went belly-up.
The K8 was little more than a K7 with the memory controller moved onto the die. And the Phenom is a quadcore variation of that architecture.

bulldozer will be a very interesting chip to watch....

pretty much all of the design will come "from AMD", and if it is as fundamentally new as they would lead us to believe, then it's success/failure would speak volumes about the company...
 
That's because someone already did the hard work of porting it to ARM for you.
That doesn't mean that the porting itself was easy to do, because it wasn't.
And even once you have ported the kernel, that doesn't guarantee that all linux software will work.

well I wont disagree, as I haven't really delved too deeply into the Linux source (that stuff is way over my head...)

but... lets just agree, in it's current state, it is portable "enough" even if the source might be a bit ugly...

on the applications end.... many frameworks are cross platform, and much open source work is done with high level programming languages.... so in fact, many (i'd say most...) programs could be easily ported... especially projects that are actively being developed (i.e. the biggest and most widely used ones...)
 
De-Rail.jpg


Back on topic, I think it's a good move. AMD has always unloaded extra inventory at prices they can't be turning a profit on, if they could slow production down to meet demand and sell their extra fab capacity they still win.
 
De-Rail.jpg


Back on topic, I think it's a good move. AMD has always unloaded extra inventory at prices they can't be turning a profit on, if they could slow production down to meet demand and sell their extra fab capacity they still win.

I guess that most investors agree with you as the share price jumps really high right now.
 
All you have to do is look at how much Nvidia had to drop their prices on their video cards when AMD came out with the 4000 series to see that yes Intel un-checked would charge more. That is pure capitalism and an industry standard.
 
All you have to do is look at how much Nvidia had to drop their prices on their video cards when AMD came out with the 4000 series to see that yes Intel un-checked would charge more. That is pure capitalism and an industry standard.

And likewise AMD dropping their processors so aggressively in the past 2 years because they were outclassed by the Intel chips. Not to mention, if AMD wasn't trying to undercut someone to begin with, there would be no need to make the 4000 series such a bargain.

Josh
 
All you have to do is look at how much Nvidia had to drop their prices on their video cards when AMD came out with the 4000 series to see that yes Intel un-checked would charge more. That is pure capitalism and an industry standard.

The turnaround rate of GPUs is far higher though (at least with high-end enthusiast cards like that).
You generally need to upgrade your GPU every 2 years or so, else you're hopelessly behind the times and unable to play any game properly.
CPUs can last much longer, and are also sold to a much bigger audience.
Simply put: You can get away with high GPU prices for enthusiast high-end parts, but that won't work for more mainstream parts, especially not with CPUs.

Likewise nVidia couldn't have sustained such prices... Once the initial enthusiast market is saturated, they need to lower prices, else they won't sell anything anymore. Also, they need to develop a new generation of GPUs, because you can sell people a GTX280 only once. 2 years later their card will still work, and they won't buy a new card if it's no faster than their current one.

You see the same with CPUs by the way... The high-end parts of a new architecture are expensive, but after a few months the prices drop. We call that "afromen", translates to "scooping off the cream" or something. You just take advantage of people with money to burn on the fastest and newest available stuff in the first few months.
 
And likewise AMD dropping their processors so aggressively in the past 2 years because they were outclassed by the Intel chips. Not to mention, if AMD wasn't trying to undercut someone to begin with, there would be no need to make the 4000 series such a bargain.

Josh

The concept remains, competition forces price cuts which benefit us.
 
The concept remains, competition forces price cuts which benefit us.

Competition requires products of similar capabilities from two or more parties however.
This has not really been the case in the CPU world for over 2 years.
AMD can do nothing to drive the prices of Intel's quadcores down, and it has trouble competing against Intel's faster dualcores, even with triple core processors.
Yet many people have no problems at all affording Intel processors.
 
Competition requires products of similar capabilities from two or more parties however.
This has not really been the case in the CPU world for over 2 years.
AMD can do nothing to drive the prices of Intel's quadcores down, and it has trouble competing against Intel's faster dualcores, even with triple core processors.
Yet many people have no problems at all affording Intel processors.

Intel has dropped their prices every time AMD has or vice versa, therefore both sides are influencing both and that's competition. AMD's stuff is competitive at stock levels when you factor in the entire platform and not everybody overclocks which is why Intel is responding. Just google it.
 
Competition requires products of similar capabilities from two or more parties however.
This has not really been the case in the CPU world for over 2 years.
AMD can do nothing to drive the prices of Intel's quadcores down, and it has trouble competing against Intel's faster dualcores, even with triple core processors.
Yet many people have no problems at all affording Intel processors.

Because if Intels prices went skyhigh people would stop buying them, duh. (as long as there is an alternative)
 
AMD's stuff is competitive at stock levels when you factor in the entire platform
AMD's declining revenues and market share disagree with that. ;) AMD's only weapon is price, and it wasn't doing itself a favor by selling products at a net loss.
 
Because if Intels prices went skyhigh people would stop buying them, duh. (as long as there is an alternative)

As if they'd buy it if there was no alternative.
It's not like your life depends on it. You could just use your old system for a few years more. Just watch how quickly Intel drops its prices when it doesn't sell any CPUs for a year.
 
AMD's declining revenues and market share disagree with that. ;) AMD's only weapon is price, and it wasn't doing itself a favor by selling products at a net loss.
Intel has been lucky in the marketing area
 
A monkey wrench in the works?

Intel says will defend patent rights against AMD

Tue Oct 7, 2008 4:35pm EDT

SAN FRANCISCO, Oct 7 (Reuters) - Intel Corp (INTC.O: Quote, Profile, Research, Stock Buzz) said on Tuesday it has "serious questions" about a deal announced by Advanced Micro Devices (AMD.N: Quote, Profile, Research, Stock Buzz) to spin off its manufacturing business and focus on chip design, and is ready to defend its patents.

Intel spokesman Chuck Mulloy said that AMD and Intel have a patent cross-licensing agreement under which AMD pays royalties to Intel.

"Intel has serious questions about this transaction as it relates to the license and will vigorously protect Intel's intellectual property rights," Mulloy said of AMD's announcement.

Mulloy said Intel has asked AMD to make the agreement public but he said it would not, so he was not at liberty to discuss the matter in detail. (Reporting by David Lawsky)
So it looks to Intel like AMD may be trying to break the x86 license agreement.
 
So it looks to Intel like AMD may be trying to break the x86 license agreement.

Makes you wonder... Perhaps AMD thinks they have an ace up their sleeve with x64? It is a *cross*-license after all.
 
I don't believe that Intel uses AMD's implementation of x86-64. It's the implementation and other technology that is covered by the x86 agreement, not the instruction sets. As the years have gone on AMD has been less dependent on that, but it is still parts of current CPUs that AMD is shipping... under the terms of the x86 license. ;)

x86-64 is not an ace.

I had suggested several months ago that AMD might break the x86 license out of desperation. This could be it.
 
I don't believe that Intel uses AMD's implementation of x86-64.

Ofcourse they don't (it's an extension of their 32-bit implementation).
But it's also an extension of the x86 instructionset, and AMD developed that extension, so they own the rights to it.
My question therefore is this:
If AMD and Intel no longer have a cross-license agreement, can Intel still use x86-64? I would guess not, in the most strict sense.

It's the implementation and other technology that is covered by the x86 agreement, not the instruction sets.

Wrong, since both companies have had their own independent implementations since the K5-era. And before that, AMD had to reverse engineer Intel's 386/486 because Intel never gave them any implementation details.
The current x86 agreement dates from 2001, so long after that era. It is purely the right to implement a CPU using the x86 instructionset (although there is the Transmeta loophole that x86 wasn't actually implemented in the hardware, but in software). Basically it has to do with all x86-related patented technology, but patents are conceptual by nature, not implementational.

Licensing implementation wouldn't make sense. The license would then mean little more than Intel outsourcing production to AMD (which they did in the 8086/80286 era).

x86-64 is not an ace.

Well, I guess they can tell Intel "If we can't use x86, you can't use x86-64". Which could mean that Intel has to disable 64-bit mode on future products.
Question is if Intel cares. It could be an excellent opportunity for them to kill off x86 and move Itanium into the mainstream, like they planned some years earlier.
(I don't think they would be allowed to just buy out AMD as a whole if AMD is not willing to license x64 to them).

I had suggested several months ago that AMD might break the x86 license out of desperation. This could be it.

Yes I'm quite sure the license at least has to be renegotiated. And Intel might actually be able to stop AMD from shipping x86-based products pending the renegotiations?
 
Question is if Intel cares. It could be an excellent opportunity for them to kill off x86 and move Itanium into the mainstream, like they planned some years earlier.

That's impossible at this point. x86 is far too entrenched to go away. MS + the world is not going to accept a ISA that can't run binaries going past 15+ years without big emulation overhead for a consumer cpu, and x64 is also too far into wide adoption for Intel to give up.
 
AMD should hit intel where i really hurts if they want to be like that. They should just stop producing chips and let the US and EU break Intel into bits. What kind of retarded company attacks the only thing between them and monopoly charges?
 
seriously, where do you get the idea that intel would charge 2-4x more for CPUs?
Past experience. Some of us were old enough to buy computers in the mid-90s. It's clear that you weren't; otherwise you wouldn't be asking this question.

THERE ARE OTHER CPU MANUFACTURERS OUT THERE.

Windows is NOT the be all and end all of the world, and Linux and OSX both run on different architectures and do so quite well....
This statement is a mishmash of confusion. You start off by saying there are other CPU manufacturers out there, then go on to talk about Windows, then finish off by claiming that Linux and OSX run on different "architectures." A majority of the Linux boxes out there are running on Intel or AMD hardware, as OSX.

Intel would be slitting its own wrists if it jacked up CPU prices while slowing innovation, because I assure you, IBM or any number of the ARM manufacturers would be licking their lips at the thought of breaking down the Windows wall and completely opening the personal computing space to processor instruction competition...

Absolutely untrue. In the mid-90s the PowerPC was arguably a better CPU than anything Intel had to offer, and yet Intel didn't seem to be hurting too much when they initially refused to give refunds on their CPUs due to flawed floating point units unless you could prove that you did floating point processing on your PC.

it is an idiotic assumption to think that a dead AMD means a slower pace of innovation + higher costs.... Intel is NOT the only company that makes microprocessors.....

No it's idiotic to claim that history is not a valid predictor of future behavior.
 
get over this silly argument? and you ask me how old I am? how exactly is that a silly arguement....

with the migration to web apps, the growing importance of netbooks, UNIX, and cell phones, how exactly is x86 the only player in town?

your naivete is amusing to say the least.....
Three words: pot, kettle, black.


edit: the only thing you x86 zealots have in your arsenal is Windows.... in it's current released state.

It sure sounds like you're marginalizing how important Windows is. It does afterall represent 90 percent of the market.

But I would like to remind you that Windows has run on other architectures int he past... and if Intel *did* turn into a bunch of bastards, I'm sure MS wouldn't hesitate to release ports as soon as it could....

Really? Kinda like when Microsoft held off on releasing a 64 bit version of Windows until Yamhill was released, and even then didn't release a 64bit compiler until Visual Studio 2005 (although it was possible via the DDK to develop 64bit stuff prior to then, but the technique was officially undocumented)
 
If only people would have picked an x2 over a pentium 4.:(
 
Back
Top