Did Samsung just buy AMD?

[Osiris]

Possibly a good alignment for both companies.  AMD has a lot of potential just not the leadership or capital to do things properly.  With Samsungs weight behind them could be a very good thing.

 

Hope its true and it results in them actually getting their products to market quicker.

[FunkyMike]

I hope this happens.

 

AMD management and staff needs a good spring cleaning.

[The_Decryptor Veteran]

I'm not sure what you mean by "gut it".

AMD doesn't have fabs anymore.  They're basically a bunch of engineers who design CPU/GPUs.  The products already being manufactured would likely be allowed to continue even if they lost the license.  They'd just be prohibited from using the license in future designs and possibly from placing more manufacturing orders for the existing designs.  AMD is already designing ARM chips though so they'd just sell what they've got and then release their ARM chips going forward.  Losing the x86 license wouldn't effect their graphics business at all.  Just like ATI didn't make CPUs but sold graphics cards for PCs before AMD bought them.

 

It's not clear they'd lose the x86 license though.  It's true AMD licenses x86 from Intel but AMD designed x86-64 and it's a CROSS LICENSING deal not a one way street.  It's likely if Intel cancelled the deal due to the buyout they'd lose the rights to parts of the 64bit extensions.  Additionally having no major x86 competition may very well cause the U.S. government to launch an anti-trust investigation against them.  I wouldn't really see a combined Samsung/AMD entity as a major threat to Intel in existing x86 markets.  They'd likely keep around the same share AMD currently has it's the mobile segment and ARMv8 designed cores that would see the big competition IMHO.

No, it doesn't work that way. Companies don't get in legal trouble when their competitors get brought out and gutted, only when they actually break the law.

[Open Minded]

At the moment those Bay Trail CPU's are weaker than the equivalent Snapdragon series, in terms of both graphics and computational performance they're about in line with the Snapdragon 600 series CPU's. Which is reasonable, but hardly groundbreaking. Intel are nowhere near cornering the market yet, and not everyone needs a tablet that can run X86 programs, most people just buy them for content consumption.

 

 

Honest question- Do you think ARM can compete with Intel in the R&D department?

[JustGeorge]

This isn't in the WSJ so I'm very skeptical of if this is true...

 

Especially since Intel has said many times in the past that they don't believe that the US Court ruling that allows AMD to exist is transferable to another owner. As such, they have said that if AMD was ever to be purchased they would revoke the x86 license AMD needs to build x86 chips.

I'd bet any revocation would spur some court battles and Intel would lose. Intel would effectively eliminate its competition and gain a monopoly overnight. I just don't see something like that going unchecked, legal or not. This could be good for AMD. They'd finally have some much needed financial backing. Intel pulling x86 license is just admitting they're scared of the actual competition they might face.

 

Meanwhile, I hear current cost cutting policies are being suspended in light of this news. Every other florescent light is now authorized to be switched on, the heat/ac is back on, hot water tanks powered up and employees are now obliged to flush toilet paper vs. save/reuse. 

 

 

No, it doesn't work that way. Companies don't get in legal trouble when their competitors get brought out and gutted, only when they actually break the law.

True, but how shady is holding the revocation of a crucial piece of licensing over the head of another company in order to potentially control the outcome of a business deal? "We like you just the way you are AMD. Nice and weak."

[Torolol]

Sx86-64 family line processors.

[pupdawg21]

 

My guess is anti trust rules and such would probably prevent Samsung (foreign company) from buying AMD due to a whole slew of legal and logistical issues. If anything, maybe a joint partnership may be possible but an outright purchase/acquisition probably would be a long hard fought fight to get approved.

[+LogicalApex MVC]

I'd bet any revocation would spur some court battles and Intel would lose. Intel would effectively eliminate its competition and gain a monopoly overnight. I just don't see something like that going unchecked, legal or not. This could be good for AMD. They'd finally have some much needed financial backing. Intel pulling x86 license is just admitting they're scared of the actual competition they might face.

Meanwhile, I hear current cost cutting policies are being suspended in light of this news. Every other florescent light is now authorized to be switched on, the heat/ac is back on, hot water tanks powered up and employees are now obliged to flush toilet paper vs. save/reuse.

True, but how shady is holding the revocation of a crucial piece of licensing over the head of another company in order to potentially control the outcome of a business deal? "We like you just the way you are AMD. Nice and weak."

Why are you so sure Intel would lose?

The only reason AMD holds an x86 license is due to them originally serving as a fab for Intel. Intel severed their relationship and AMD won the right to continue making chips using Intel patents in arbitration. Intel didn't decide to grant AMD this license on its own and hasn't granted one to anyone else.

The only area of confusion is whether or not AMD can transfer this license to a new owner. Intel has said many times that they don't see this as the case. As such, buying AMD is a guaranteed court case without a certain outcome. Losing it would effectively kill all of AMD. Making it too risky a deal.

Samsung buying AMD to run around Intel would be silly. They would need to get Intel to sign off on the deal to not risk a very risky court battle. Intel definitely would not sign off on increased competition.

[Torolol]

x86 license

.. AMD are not the only one who licenses x86 from intel.

there also VIA-embedded: http://www.viaembedded.com/en/processors/

[+LogicalApex MVC]

.. AMD are not the only one who licenses x86 from intel.there also VIA-embedded: http://www.viaembedded.com/en/processors/

I forgot about VIA. They earned theirs with an acquisition in the late 90s. Sometime around when AMD was outperforming Intel.

Their agreement was far more limited I believe. Which is why they never made full on desktop chips.

[+Asmodai MVC]

No, it doesn't work that way. Companies don't get in legal trouble when their competitors get brought out and gutted, only when they actually break the law.

It DOES work that way.

Companies get in monopoly trouble when they have no significant competition no matter how they arrived at that condition.  If your competitors all fall through no fault of your own and you become the only company in a major market then you are a monopoly even if you did nothing wrong.  You don't only become a monopoly from doing something wrong.  With no competition you can charge whatever you want for your product and so you're likely to at least get a ton of government regulation to control your prices or even be split up into smaller companies to compete with each other.  The U.S. government is unlikely to allow Intel to be the ONLY company that can make x86 CPUs considering the vast majority of PCs require that for existing commercial software.  Now maybe Intel will license x86 to someone else instead of Samsung/AMD if this merger actually happens and Samsung did actually lose AMDs license.  For example if they license x86 to nVidia that would probably satisfy the U.S. Government (if nVidia is even interested in making x86 CPUs.) Again though it's unclear that Samsung/AMD would even lose the license and heck Sumsung even buying AMD is still very much in the rumor stage now (although my google-fu shows many more hits today than yesterday... I suspect they're all based on the same single source though)

[Draconian Guppy]

So far, only found "rumours" of this news.

[+Asmodai MVC]

.. AMD are not the only one who licenses x86 from intel.

there also VIA-embedded: http://www.viaembedded.com/en/processors/

I may be reading this wrong (I'm no expert on VIA) but my understanding is VIA's license has expired.

 

From the Wikipedia:

On the basis of the IDT Centaur acquisition, VIA appears to have come into possession of at least three patents, which cover key aspects of processor technology used by Intel. On the basis of the negotiating leverage these patents offered, in 2003 VIA arrived at an agreement with Intel that allowed for a ten year patent cross license, enabling VIA to continue to design and manufacture x86 compatible CPUs. VIA was also granted a three year grace period in which it could continue to use Intel socket infrastructure.

According to this the VIA/Intel cross licensing agreement ended in 2013.

I don't know how old those designs on the VIA url you posted are but since they're all 40nm process or larger they may all be from 2013 or earlier. (The most recent date I can find for a VIA Nano CPU is 2011)

[Sikh]

If this is true I hope they don't kill AMD by trying to copy nvidia every ######ing chance they get. I actually respect AMD, but my respect for samsung  going down, besides their TV division. 

[torrentthief]

I'm pretty sure the EU would stop Intel from revoking the x86 licence is samsung bought out AMD due to a lack of competition.

[HeartsOfWar]

You answered you own question.  ARM, in a few years, is going to lose ground to Intel (and maybe AMD).  If AMD can pull off low wattage while using x86 then ARM will become more irrelevant.  Intel is already showing this with Bay Trail; why buy a tablet with an ARM SoC when you can get a full-blown Windows tablet that runs x86 programs (and even Android apps) for the same price?

 

If I were you, I wouldn't bet that Intel or AMD will kill or make ARM less irrelevant. ARM has been underestimated for the last 10 years and is a major factor as to why they have dominated for the last 8 years. ARM's major strength is their willingness to license and a focused market on embedded. Over the last few years Intel has definitely been competitive with their Atom line, but they likely won't be able to match the single unified front of ARM.

 

If you need more proof that ARM isn't going to just die - Look at Apple; they make their own ARMv8-A compatible SoC and have been licensing from ARM since 2010(Apple A4-A8). Apple sells 5 *major* products based on ARM licensed technology compared to the 1 *major* product that uses Intel.

 

If you had predicted prior to 2005 that RISC would YoY dominate CISC for 10 years, you'd have been the laughing stock of the world, but would have been a rich SoB if you chose to invest in ARM. As computing continues to advance and push the boundaries of packing the strongest punch into the smallest package, I think ARM will maintain just fine.

[+Asmodai MVC]

You answered you own question.  ARM, in a few years, is going to lose ground to Intel (and maybe AMD).  If AMD can pull off low wattage while using x86 then ARM will become more irrelevant.  Intel is already showing this with Bay Trail; why buy a tablet with an ARM SoC when you can get a full-blown Windows tablet that runs x86 programs (and even Android apps) for the same price?

 

I really don't see ARM losing ground to Intel.  First it's a confusing comparison because ARM doesn't make chips and Intel does.  So really the comparison is architectural and thus ARM vs x86 not ARM vs. Intel.

 

x86 is an inefficient design compared to ARM.  Intel dominates with x86 for two major reasons: 1) backwards compatibility and 2) they have the best fabs on the planet that make up for the inefficiencies with superior fab processes.

 

Now backwards compatibility doesn't mean much unless Windows is going to come along for the ride and it doesn't appear it will on phones or smaller tablets (<8").  Despite the fact Intel may make processor that can run in phones Microsoft has decided that on small tablets and phones users will NOT get the desktop or compatibility with existing x86 (Win32/64) software.  Instead these devices will be limited to "Universal" (UAP) apps.  That eliminates the usefulness of backwards compatibility.  Now on > 8" tablets MAYBE Intel/Windows will take off with Win10 but that remains to be seen.  I really think it's a bad idea for MS to segment Windows based on screen size instead of architecture but they didn't ask me   It would make more sense to me to have every x86 Windows box support the desktop and every ARM Windows box not no matter what the screen sizes are.

 

As for better fabs that will likely remain an advantage of Intel, I don't think anyone is going to pass them but the problem is it's ONLY an advantage for them and their set designs.  AMD has to use the same fabs as the ARM companies and so they lack this advantage.  More importantly companies can't custom design chips to their needs for Intel fabs (at least not at a reasonable price).  The beauty of ARM is that in addition to their specific product designs they licenses (like their Cortex CPU cores and Mali GPU cores) they license the architecture (such as ARMv8) for others to design compatible CPUs.  Supposedly Samsung is working on their own custom CPU design (historically they've just used the Cortex design), AMD is working on their own ARMv8 design, nVidia is working on their "Project Denver" ARMv8 design, Apple has their Cyclone ARMv8 cores, Qualcomm is working on their Kyro ARMv8 design, etc.  With all these different designs (let alone different combinations of features on SoC's beyond just core design differences) customers have a much greater amount of choice to get exactly what they need then the few options Intel may decide to offer.

 

Furthermore as time goes on people are using their PCs less and less and their iOS/Android devices more and more.  For many it's going to be less of an issue to lose their x86 apps than it is to lose all their existing iOS/Android apps.  I don't see Apple putting Intel cpus on iOS devices since they design their own ARM cores now and while Android can technically run on x86 already I really don't see it unseating ARM since "backwards compatibility" on the Android platform is ARM not x86.

[Open Minded]

 

You never answered my question.  Do you think ARM can compete with Intel as far as R&D goes?

 

Intel is going 10nm this year with Skylake.  Samsung, whom is the closest in terms of lithography, just went to 14nm this year and that's only on the front-end of the SoC.  TSMC and Global Foundries are at what, 20nm?  If Intel stopped die shrinks now, it would take everyone else a few years to catch them.  I doubt Intel is going to stop with the die shrinks, so unless a miracle happens (change from silicon to whatever), it's going to stay this way. 

 

Do you think Apple will continue to design their own ARM SoCs when Intel brings the power usage down to be good enough for use in phones?  Why waste the money if you're Apple?  Just let Intel do it.  Cherry Trail (14nm) is on the way and who knows how well that's going to preform.  I would bet in less than 5 years Apple will be using Intel SoCs in their phones and tablets.  There, I said it and you can quote me.  :-)

 

Looking at it from a different direction, do you think Apple will scale-up an ARM SoC for use in desktops?  Dropping x86 compatibility on the desktop and in their Macbook line would kill sales.  Every professional would have to go PC to do any real work.  Does ARM or Apple have anything that's in the ballpark of Broadwell Core M?  Why does the new Macbook use a Core M CPU and not one from ARM?

[BinaryData]

This doesn't effect me in the slightest. I refuse to run AMD GFX Cards, but love Samsung phones. I'm getting the Galaxy 5 soon as 6 drops!

[spaceelf]

Samsung, whom is the closest in terms of lithography, just went to 14nm this year and that's only on the front-end of the SoC.  TSMC and Global Foundries are at what, 20nm?

Global Foundries has a deal with Samsung to use their 14nm process last I checked.

[Open Minded]

Global Foundries has a deal with Samsung to use their 14nm process last I checked.

 

More proof GF is slipping.  How much is it going to cost them to catch up to Intel or even Samsung?

[+Asmodai MVC]

You never answered my question.  Do you think ARM can compete with Intel as far as R&D goes?

They don't have to. ARM just needs license the ARCHITECTURE (ARMv8) and make a reference platform (Cortex CPUs).

I don't think ARM designed Cortex cores are likely to beat Intel cores but I seriously doubt the Cortex cores will be the best ARMv8 cores.

Instead the best ARMv8 cores will be designed by others who have licensed the ARCHITECTURE and not the specific core design. Companies like Apple with their Cyclone core, nVidia with their "Project Denver" core, Qualcomm with their Kyro core, etc. The list goes on and on.

I have no idea which ONE of them will have the fastest core but I do believe their combined R&D far outpaces Intel.

Intel is going 10nm this year with Skylake.  Samsung, whom is the closest in terms of lithography, just went to 14nm this year and that's only on the front-end of the SoC.  TSMC and Global Foundries are at what, 20nm?  If Intel stopped die shrinks now, it would take everyone else a few years to catch them.  I doubt Intel is going to stop with the die shrinks, so unless a miracle happens (change from silicon to whatever), it's going to stay this way.

None of this has anything to do with ARM. ARM doesn't have any fabs nor do they even use them. They don't actually make any physical product at all, they just sell IP.

I agree that Intel will stay ahead of others in fabs but their inflexibility of designs and (power) inefficiencies in the x86 architecture (as well as MS not putting the desktop on < 8" devices) will ensure they don't overtake ARM on mobile.

Do you think Apple will continue to design their own ARM SoCs when Intel brings the power usage down to be good enough for use in phones?

Absolutely.

Why waste the money if you're Apple?  Just let Intel do it.

Because they don't want to be at the mercy of an external company. It would be a huge problem to convert all ARM based iOS devices to x86 and they can custom design their chips however they want and not be at the mercy of what Intel decides they want to do.

If they make some cool new breakthrough they can keep if for themselves. If Intel makes it they're going to want to sell it to everyone who buys their chips not let Apple alone have it.

I would bet in less than 5 years Apple will be using Intel SoCs in their phones and tablets.  There, I said it and you can quote me.  :-)

I'd take that bet. In fact I'd go so far as to say that it's MORE likely that Apple laptops and computers will be running Apple designed ARMv8 CPUs before iPhones run Intel CPUs (though I don't think that will happen in under 5 years)

Looking at it from a different direction, do you think Apple will scale-up an ARM SoC for use in desktops?  Dropping x86 compatibility on the desktop and in their Macbook line would kill sales.  Every professional would have to go PC to do any real work.  Does ARM or Apple have anything that's in the ballpark of Broadwell Core M?  Why does the new Macbook use a Core M CPU and not one from ARM?

I think iOS will always be ARM based. Mac OS X will remain x86 (espeically on the desktop) in the near term (<5 years) but who knows farther out than that.

The question here is will iOS grow up to eventually run on laptops/desktops to replace OS X? I honestly have no idea but wherever iOS goes ARM goes.

Again though I think it's more likely that a version of OS X will run on ARM then a version of iOS will run on x86. Not in the near term but IF it does happen I suspect it will start with servers and work it's way down to desktops.

[HeartsOfWar]

You never answered my question.  Do you think ARM can compete with Intel as far as R&D goes?

 

If you're going to quote me, actually quote what I wrote instead of inserting your own Gif...

 

In answer to your question, I gave you the shorthand answer: licensing, and Asmodai essentially gave you the longhand version.

 

  I would bet in less than 5 years Apple will be using Intel SoCs in their phones and tablets.  There, I said it and you can quote me.  :-)

 

First off, 5 years is a long time for technology, and I wouldn't doubt that x86 will invade the embedded market as time marches on (we are seeing this); however, I believe there will be a fixed limit / rate to the invasiveness.

 

The major benefit of ARM and why its so pervasive in the embedded market is due to the fact that the RISC architecture limits a single clock cycle per instruction and therefore needs less hardware and less demanding hardware. In other words, RISC architecture by design lends itself to smaller form factors and will likely stay the 'go-to architecture' for the embedded market on devices that can progressively shrink without worry.

 

Will we see x86 phones and tablets at some point? Likely, but I think it's more realistic to see ARM and x86 co-existing in a single device in the next 5 years than total x86 domination into the embedded market.

[spaceelf]

More proof GF is slipping.  How much is it going to cost them to catch up to Intel or even Samsung?

It seems if they're using Samsungs process they already caught up with Samsung.