The AMD Execution Thread [2007 - 2017]

Status
Not open for further replies.
Yeah, I really can't think of any way to split AMD into two halves that wouldn't destroy both halves.

If the CPU division can't recover with Zen and the ARM solutions in 2016/2017, then the graphics division should be salvageable for discrete graphics cards and licensable IP for APUs.

Truth be told, AMD isn't in the unique position of being able make high-performance HSA APUs/SoCs anymore.
Intel's graphics division is catching up in performance/power and even absolute performance, and the iGPUs in ARM SoCs are catching up really fast.
Given access to Wide I/O2 and a generous TDP, a bunch of SoC manufacturers could come up with a chip matching the PS4 in performance using Cortex A72 and PowerVR 7XT.
 
Bristol Ridge APUs are Carrizos with DDR4

Bristol Ridge is the desktop variant equivalent to Carrizo (only mobile) which is said to support both DDR3 and DDR4 (here I would like to add that on hardware level the memory controller might be able to work with DDR4 but that feature might be locked in this version).
Maybe Bristol Ridge would ultimately drop DDR3 all together.
 
Last edited:
Bristol Ridge is the desktop variant equivalent to Carrizo (only mobile) which is said to support both DDR3 and DDR4 (here I would like to add that on hardware level the memory controller might be able to work with DDR4 but that feature might be locked in this version).
Maybe Bristol Ridge would ultimately drop DDR3 all together.
It will, it's socket compatible with Zen so has to go on same mobos
 
Could AMD clone parts of itself, bundle them into a new division, and then sell that? It might cannibalize some sales in the medium term, once the buyer started selling product based on the purchase, but it might raise a sizable piece of change and also raise AMD's profile in markets it's struggling in. So if, say, Samsung was the buyer, then they could get oodles of benefits from the technology, and AMD would get cash and would be seen as having sexy/desirable technology by younger buyers of gadgets. Companies like Apple would come out as losers if that plan worked. Or, on the other hand, maybe AMD could get a deal with Apple.
On the gripping hand, AMD could seed the rumor of such a spinoff, and then wait to see what deeply concerned tech companies come knocking at their door. :) Mostly kidding, I know AMD is a big league outfit and has standards. Still ...
 
Last edited:
If the CPU division can't recover with Zen and the ARM solutions in 2016/2017, then the graphics division should be salvageable for discrete graphics cards and licensable IP for APUs.
Is there a viable space to support the kind of development needed for large high-power GPUs in the discrete form for a graphics provider with no CPU or presence in other markets? The graphics group was middling before and losing money now.
Nvidia is the big player that is severely hurting the discrete space, and Intel is both hurting AMD's APU efforts even as it hurts the discrete mobile space.
ARM, Qualcomm, Nvidia, Intel, Via, and now IMG are all capable producing graphics IP while also possessing CPU capabilities.

Maybe sell it off while there is still some value? AMD could keep the rights to the extant GPU IP it has been leveraging for years.
For cosmic irony, Qualcomm could buy it, since AMD sold its mobile graphics and then lost the CTO that headed the creation of GCN to them. Not that GCN is successful at all in the power range Qualcomm would be interested in.
Could Samsung find use in buying it to counter Apple's increasing assertiveness in graphics?
How about one of the few graphics IP licensors without notable CPU presence like Vivante? It was part of HSA so it has some dealings with AMD. Or at least it was a member until this May.

Maybe if AMD can get some hundreds of millions from such a sale, then keep its cash balance, and then make some hundreds of millions more from a year of at least non-disastrous Zen sales, it might be able to retire enough of its debt that it is worth buying as an ARM design wing and a potentially modest licensing fee from Intel once the x86 operations are wound down.
 
Is there a viable space to support the kind of development needed for large high-power GPUs in the discrete form for a graphics provider with no CPU or presence in other markets? The graphics group was middling before and losing money now.
Nvidia is the big player that is severely hurting the discrete space, and Intel is both hurting AMD's APU efforts even as it hurts the discrete mobile space.
ARM, Qualcomm, Nvidia, Intel, Via, and now IMG are all capable producing graphics IP while also possessing CPU capabilities.

Maybe sell it off while there is still some value? AMD could keep the rights to the extant GPU IP it has been leveraging for years.
For cosmic irony, Qualcomm could buy it, since AMD sold its mobile graphics and then lost the CTO that headed the creation of GCN to them. Not that GCN is successful at all in the power range Qualcomm would be interested in.
Could Samsung find use in buying it to counter Apple's increasing assertiveness in graphics?
How about one of the few graphics IP licensors without notable CPU presence like Vivante? It was part of HSA so it has some dealings with AMD. Or at least it was a member until this May.

Maybe if AMD can get some hundreds of millions from such a sale, then keep its cash balance, and then make some hundreds of millions more from a year of at least non-disastrous Zen sales, it might be able to retire enough of its debt that it is worth buying as an ARM design wing and a potentially modest licensing fee from Intel once the x86 operations are wound down.

I don't like scenario you're painting here but part of me agrees with your assessment. AMD needs miracle now to recover from the loses they suffered (engineering and financial). It's not impossible, see Apple, but highly unlikely we will see AMD as x86 and GPU company by 2020.
Even if Samsing or Microsoft decided to pick them up and push for new x86 designs it would take years before anything meaningful landed in stores. Zen is AMD's last chance to remain relevant in processor game and due to lack of resources GPU side is suffering as well.
 
Reuters backing up the claim:
http://www.reuters.com/article/2015/06/19/us-amd-split-idUSKBN0OZ2KP20150619

It seems that the consulting companies have been approached very recently. Perhaps AMD is just looking for fallback measures in case Carrizo and Zen fail to get any major clients.
I for one don't believe AMD will split from their graphics division before getting a powerful Zen APU with HBM out there.

Is there a viable space to support the kind of development needed for large high-power GPUs in the discrete form for a graphics provider with no CPU or presence in other markets?
Isn't that what nVidia is right now and ATi was, before the acquisition?
The PC gaming market has been growing. Isn't there a growing demand for PC graphics cards to keep up with it?
 
One scenario, speculated by our site staff, could be AMD simply splitting to CPU/APU, GPU and Server entities, which would be clients to each other. This could enable them to be more flexible on hiring/firing people while maintaining the competitive edges
 
Isn't that the same story linked earlier, or is there something added to it that adds corroboration?

Isn't that what nVidia is right now and ATi was, before the acquisition?
Nvidia has a CPU presence, as limited as it is currently. Those on the leading edge of graphics seem to agree that they have a better chance at growth, so pure graphics licensing does not seem to be a safe place to run to now.
I suspect that a spinoff company would not be better-staffed or funded than the already desperate and emaciated AMD. There's also likely no way creditors are going to let AMD saddle a sacrificial division with its massive debt.

It would take mobile or embedded to even want to license graphics IP at this point, and what a graphics group would offer is no CPU, less design capability, a GPU that is not good for the market, and more debt than its competitors.
I think it would have to be some kind of long-term play by a company with existing large scale that is willing to eat losses to set up its own graphics wing, based on a set of teams and IP with graphics experience. This assumes there's more value there than a fire sale picking it clean.


The PC gaming market has been growing. Isn't there a growing demand for PC graphics cards to keep up with it?
The low and low middle end are losing ground to integrated, and the high end's volumes are not that great. The console market whose games get ported to the PC is not really that great relative to the realities of today.
Then there's the fact that AMD is being dominated there so that market's growth is not going to them.
 
Another big issue is that AMD has little presence in professional graphics, where the really fat margins are. The way I see it, next year with Zen and 14nm GCN will be the "make it or break it" moment for AMD.
 
Apple of old was a sinking ship. Mac's didn't sell well enough, company made couple of bad decisions, their OS was behind the curve as well.

This is first search result I came up with more info on this topic:

http://www.roughlydrafted.com/RD/Q4.06/9FD12E37-8DC7-4AD1-872F-2021BEDE6D96.html
Unless Bill Gates (or a similar benefactor) is going to give Lisa Su some (considerably sized) stacks of coins because she delivers great presentations and gets it done, the comparison might not be completely adequate. Although it is nice to only consider the romantic version in which lord Jobs came back from Avalon to save his company and lead it to countless billions, it is likely that Apple would be a memory by now had there not been any external intervention at key points during its dark days.
 
Isn't that the same story linked earlier, or is there something added to it that adds corroboration?

It was practically the same story, the only difference being that Reuters are claiming they were tipped by their own sources.


Regardless, AMD is officially refuting the claim:
AMD provided official confirmation that we have not hired an outside agency to explore spinning-off/splitting the company. We remain committed to the long-term strategy we laid out for the company in May at our Financial Analyst Day, which encompasses all parts of the business.
 
Status
Not open for further replies.
Back
Top