AMD mulling break, spinoff

[Fjodor2001]

In order to save at least some parts of the company.

What part(s) do you consider especially worth saving compare to the others?

And regardless if it's the CPU or GPU division that is lost, it would mean they cannot make APUs which is at the core of AMD's business. So it doesn't make sense.

 

[ShintaiDK]

And regardless if it's the CPU or GPU division that is lost, it would mean they cannot make APUs which is at the core of AMD's business.

Thats not true.

 

[Fjodor2001]

Thats not true.

What's not true and why?

 

[dark zero]

Thats not true.

Actually it is.. The only reason AMD was so good as an APU is that. They had decent CPU and excellent GPU in one chip.... Sadly they were so advanced in times where the bottleneck is now the RAM.
If they got separated, the industry would stop using AMD due their uselessness.

Also that is the reason why Nintendo, Sony and MS are making consoles. If not, they would let the PC doing everything.

And you know... A monopoly would always lead into a dissaster...

 

[ShintaiDK]

Console makers dont buy AMD CPUs. They buy AMD graphics. And nobody is really buying AMD APUs in the PC segment in the first place.

Semicustom+graphics could survive and start using ARM vanilla cores. The CPU part could grant the Semicustom+graphics part a free license. And I am sure Intel would say yes. Because that part must already be agreed with Intel when AMD signed the 2 up for the consoles. Sony and Microsoft wouldnt do it if they didnt get that garantee.

 

[Fjodor2001]

Console makers dont buy AMD CPUs. They buy AMD graphics.

Console makers buy APUs (despite that you said XBONE/PS4 would never be APU based, remember!). They chose the best combination of CPU and GPU available. And apparently AMD's GPU leadership mattered more than Intel's CPU leadership.

 

[mrmt]

What part(s) do you consider especially worth saving compare to the others?

The graphics division. Under competent management and with a small- capital injection the division could really compete against Nvidia. The same cannot be said about their CPU division against Intel and the other ARM licencees. CPUs are a game too big for AMD.

And regardless if it's the CPU or GPU division that is lost, it would mean they cannot make APUs which is at the core of AMD's business. So it doesn't make sense.

Not really. They could just license the CPU IP from someone else, for example, ARM, OpenPOWER, and add the real diferentiator which is the GPU IP.

Console makers buy APUs (despite that you said XBONE/PS4 would never be APU based, remember!). They chose the best combination of CPU and GPU available. And apparently AMD's GPU leadership mattered more than Intel's CPU leadership.

I think you are bringing a red herring to this discussion (Console makers buy APU!) instead of discuss the real question for AMD management: Is it enough to have the console customers in order to sustain our CPU business? The answer to this question is certainly no, even when we expand the scope of the analysis to the rest of the embedded market. That's why AMD is trying to get back into the server market, and not double down their bet on embedded/consoles.

The consoles were basically a business opportunity that appeared to AMD, a secondary application of their IP developed to other segments, but by no means a business opportunity with enough ROI to get more priority than others, like servers.

That said, given the anemic CPU power the consoles have, don't you think that A72 or whatever succeeds it won't pack enough punch (and be far cheaper to integrate on a SoC) than anything AMD can develop with its anemic R&D budget?

 

[dark zero]

Sorry, but Even GPU división is an absolute dissaster... I am really sorry for AMD and Matrox, but both are fated to die. NVIDIA will accompany them later. The GPU market is on their last moments. And seems that the Fury dissaster is the beginning

 

[JDG1980]

That said, given the anemic CPU power the consoles have, don't you think that A72 or whatever succeeds it won't pack enough punch (and be far cheaper to integrate on a SoC) than anything AMD can develop with its anemic R&D budget?

Unless something really strange happens, Zen will be far more powerful than anything ARM can come up with.

In fact, even Carrizo outperforms most ARM offerings. Remember, Intel's anemic Bay Trail CPUs were near the top of the charts when they showed up in tablets. ARM's only strength is its ultra low power usage; in performance, it is crap compared to x86.

Besides, the game studios like having consoles be basically a normal x86 PC.

 

[mrmt]

Unless something really strange happens, Zen will be far more powerful than anything ARM can come up with.

Not arguing that, just arguing that if the consoles can make it with 8 anemic jaguar cores, they will be able to make it with 8 A72 cores or whatever come after them, especially if it comes with a smaller price tag than current jaguar cores.

 

[dark zero]

Actually AMD is so lame that is better for the companies to put that lame CPU and put 3.5 Ghz Intel Quad Cores along a Gtx 970 in their consoles on a good prices since both are getting cheaper... Since the current pricing of the consoles is not justifying AMD anymore.

To make it worse, the best games of this years are more from PC instead consoles (despite the extreme quality drop compared previous years), something that years before didn't happen.

 

[postmortemIA]

New rumor. Qualcomm is interested in buying whole or parts of AMD.

 

[eternalone]

Qualcomm May Acquire AMD: This Could Trigger Significant Upside For Both Stocks


http://seekingalpha.com/article/327...ld-trigger-significant-upside-for-both-stocks

 

[Phynaz]

Qualcomm May Acquire AMD: This Could Trigger Significant Upside For Both Stocks


http://seekingalpha.com/article/327...ld-trigger-significant-upside-for-both-stocks

Horrible conjecture, that's all that is.

 

[dark zero]

INBF Apple goes and buys AMD in one shot.

AMD is so dead that people are after it's corpse

 

[jpiniero]

Samsung makes sense. Apple makes sense to buy the GPU IP. Qualcomm, I can't see it.

 

[Ajay]

Well, this could just be hedge fund managers playing their usual games to make various plays in case Greece really tanked. This sort of stuff probably would have started last week and then filtered up. Other than that, I don't get it.

 

[Ajay]

The graphics division. Under competent management and with a small- capital injection the division could really compete against Nvidia. The same cannot be said about their CPU division against Intel and the other ARM licencees. CPUs are a game too big for AMD.

The only problem is that dGPUs are a short play right now. Unless some disruptive IP changes the landscape of the graphics market - Intel is going to continue to eat up more and more GPU market share. Intel is also going full guns on Compute and I predict that they will do the same in professional graphics.

 

[mrmt]

The only problem is that dGPUs are a short play right now. Unless some disruptive IP changes the landscape of the graphics market - Intel is going to continue to eat up more and more GPU market share. Intel is also going full guns on Compute and I predict that they will do the same in professional graphics.

The broad SoC market will need plenty of GPU IP. AMD could follow that route.

 

[cbn]

The only problem is that dGPUs are a short play right now. Unless some disruptive IP changes the landscape of the graphics market - Intel is going to continue to eat up more and more GPU market share. Intel is also going full guns on Compute and I predict that they will do the same in professional graphics.

I think Intel ( at the high end iGPU,for consumer processors) will be able to progress only so far before they are limited by power.

And as far as the lower end iGPU (which is now GT2) I am wondering if this is actually too big for non-gaming usage? How much iGPU does one need (even at low power, 5W chip) for UI acceleration (even with heavy flash) at 1440p? At 4K? With DPI scaling?

 

[MiddleOfTheRoad]

Sorry, but Even GPU división is an absolute dissaster... I am really sorry for AMD and Matrox, but both are fated to die. NVIDIA will accompany them later. The GPU market is on their last moments. And seems that the Fury dissaster is the beginning

Matrox isn't going anywhere. They've carved out a niche that's highly profitable for them. (2D Graphics for Digital Artists and multi-head video cards for video broadcasters).
They will never be mainstream again, but will probably continue with their very expensive cards a very small specialized audience.

 

[JM Popaleetus]

I still have this weird image in my head of the GPU and CPU divisions being split; for Intel or Apple to buy the GPU half, and Nvidia the CPU.

Then again, with Apple becoming increasingly independent, AMD as a whole would be quite the addition to their portfolio. MediaTek or Samsung I see as potential suitors as well.

 

[Stuka87]

I still have this weird image in my head of the GPU and CPU divisions being split; for Intel or Apple to buy the GPU half, and Nvidia the CPU.

Then again, with Apple becoming increasingly independent, AMD as a whole would be quite the addition to their portfolio. MediaTek or Samsung I see as potential suitors as well.

Would be no purpose in nVidia buying the CPU half. The x86 license becomes void if AMD is bought/sold.

 

[dark zero]

Matrox isn't going anywhere. They've carved out a niche that's highly profitable for them. (2D Graphics for Digital Artists and multi-head video cards for video broadcasters).
They will never be mainstream again, but will probably continue with their very expensive cards a very small specialized audience.

Not at all with Intel. When they has the tech to use múltiple displays without problem, that would be the end of Matrox

 

[postmortemIA]

Would be no purpose in nVidia buying the CPU half. The x86 license becomes void if AMD is bought/sold.

And that's why that half of company is not worth much. Buyer has to forfeit the product company makes.