AMD and ATI breaking up?

[Mako88]

If they fail we'll never see a GPU worth owning for sale from Nvidia again below $1000, nor a CPU from Intel below $500.

We all need AMD to get healthy, remain strong, and grab back some marketshare in order to foster a good competitive environment in which consumers win.

Get to work Su, and stop screwing around at nerd game shows. Leave that to the PR underlings.

Strong language is not acceptable in the technical forums
Moderator Subyman

 

[ShintaiDK]

AMDs current business model isnt substainable financially. So something have to happen. AMD got healthy parts and parts without a future. No need for the abd parts to drag down the healthy.

 

[raghu78]

AMD have gone through the most difficult past few years due to that disaster called Bulldozer. If AMD stuck it out through those dark times why would they split the company now when there is hope of a turnaround with Zen. If Zen succeeds AMD has a future otherwise not likely. By succeed I mean being competitive against Intel's big cores like Skylake. Remember AMD does not need to beat Intel but they have to be close behind in perf and perf/watt and provide competitive products at competitive prices. AMD's ace in the hole is HBM and AMD's APUs will start using HBM from 2017. AMD has confirmed a multi teraflop server APU with a transformational memory architecture in 2017. No prizes for guessing its a HBM enabled monster server APU. We can expect slightly scaled down versions for desktops and notebooks. Those APUs are potentially disruptive products. All we need is a competitive high performance Zen core and AMD is well on its way to financial health and sustainability.

 

[railven]

So we'd have a nvidia tax and an intel tax in video cards. Yay.

Yeah, but Intel is known to do some shady business deals. I'm sure my Microcenter would have amazing bundle deals for Intel CPU+GPU+MoBo Combos.

For you all without a Microcenter, my condolences.

 

[Keysplayr]

What would nVidia do with AMD? Remember no X86 license.

I'd imagine Intel would grant an x86 license if it meant that AMD CPUs (Or Nvidia CPUs) would still remain in the game with a certain level of competition to avoid a complete monopoly. In fact, the powers that be might make it mandatory.
At any rate, AMD as we've come to know it, is finally winding down. It is unsustainable in it's current state. The CEO is pretty smart to
even consider a split off option. For this to become public though, means it's been underway for a while if it bubbled to the surface like this. It's going to happen one way or another. IMHO. We all might be better off for it. Or not. One thing for sure, we are going to find out first hand.

 

[Rvenger]

My take, ATI is getting spun off but I think AMD will amend an agreement to supply the tech in their CPUs to make APUs. Unfortunately the situation could end up like Glofo with a similar supply agreement.

I believe AMD wants to be a CPU company it once was.

 

[exar333]

ATI is the only part of the company worth much these days. Besides console revenue, AMD CPUs are dead until Zen arrives. Not sure if that helps or hurts a spinoff. Not likely tho...

 

[Silverforce11]

I have a lot of hope for AMD actually.

I see great advantages with Zen + GCN 2 + HBM 2 APUs. Intel won't have a node lead come 14nm ff from Samsung/GloFo and 16nm ff from TSMC. The next node down is going to be a long ways off.

It's the perfect storm, the amalgamation of AMD's original goal of fusion, full HSA, not limited by system ram. With DX12's flexible multi-GPU application, imagine buying a Zen APU with a beefy iGPU that actually gets put to use if you add a dGPU.

Rather than now, all of us on i5/i7 with that Intel iGPU sitting there doing jack all, wasting die space & costing us $.

If Zen is good, that would make for an awesome full AMD setup. Zen APU + GCN dGPU where both will function to boost performance in DX12 games. It would be outright faster than Intel APU + NV/AMD dGPU.

A lot of AMD's future rests on Zen.

 

[jpiniero]

AMD have gone through the most difficult past few years due to that disaster called Bulldozer. If AMD stuck it out through those dark times why would they split the company now when there is hope of a turnaround with Zen.

They probably saw the Q2 numbers and they now think they can't make it all the way to Zen's release without doing something.

 

[frozentundra123456]

I have a lot of hope for AMD actually.

I see great advantages with Zen + GCN 2 + HBM 2 APUs. Intel won't have a node lead come 14nm ff from Samsung/GloFo and 16nm ff from TSMC. The next node down is going to be a long ways off.

It's the perfect storm, the amalgamation of AMD's original goal of fusion, full HSA, not limited by system ram. With DX12's flexible multi-GPU application, imagine buying a Zen APU with a beefy iGPU that actually gets put to use if you add a dGPU.

Rather than now, all of us on i5/i7 with that Intel iGPU sitting there doing jack all, wasting die space & costing us $.

If Zen is good, that would make for an awesome full AMD setup. Zen APU + GCN dGPU where both will function to boost performance in DX12 games. It would be outright faster than Intel APU + NV/AMD dGPU.

A lot of AMD's future rests on Zen.

By the time we see Zen APUs. intel will be on or close to 10 nm. And Sansungs "14 nm" process is not necessarily equivalent to even intel's 14nm. In fact it is totally unproven for hogh power, high frequency operation.

This is assuming everthing goes perfectly for AMD and there are no delays. But again, this is the same story we continually hear about AMD. Just wait till next year. The next product or some software will transform them into a competitor again. It could finally happen, but so far, all that has happened is they have dropped to all time lows in both the cpu and dgpu market.

 

[JDG1980]

Hmm. If AMD could spin off GPU plus cat cores into a separate company, while keeping big-core CPUs in the parent AMD... it just might work. The current console APUs use cat cores, so they wouldn't be affected, and the spin-off company could work on bigger and more powerful derivatives of cat cores in the future if they wanted to. If nothing else, it would ensure that the success or failure of Zen didn't risk capsizing the graphics division at the same time. On the other hand, the licensing for this could be tricky, since this would increase the total number of x86 companies by one. Still, AMD does have some leverage; they own the x86-64 IP, and probably some graphics patents that Intel is using in their own iGPUs. Intel obviously can't drop 64-bit support or remove iGPUs, so nixing the cross-licensing agreement altogether isn't an option. AMD might just be able to pull it off...

 

[xthetenth]

A competitive architecture working against a node disadvantage is a heck of a lot better than an uncompetitive architecture working against a node disadvantage. APUs are an almost decent product, a good boost to single core IPC would make them seriously viable.

 

[JDG1980]

A competitive architecture working against a node disadvantage is a heck of a lot better than an uncompetitive architecture working against a node disadvantage. APUs are an almost decent product, a good boost to single core IPC would make them seriously viable.

Agreed. Sandy Bridge was on 32nm, and look how many people are still using that - even gamers who need lots of CPU power and single-thread performance are still on i5-2500Ks.

AMD's problem in the CPU arena isn't so much that Intel has a node advantage, but that the Bulldozer architecture is trash.

 

[ShintaiDK]

Remember IPC increase cost power consumption.

 

[JDG1980]

Remember IPC increase cost power consumption.

Then how come Bulldozer with its low IPC was less efficient than Thuban?

 

[Azix]

AMD have gone through the most difficult past few years due to that disaster called Bulldozer. If AMD stuck it out through those dark times why would they split the company now when there is hope of a turnaround with Zen. If Zen succeeds AMD has a future otherwise not likely. By succeed I mean being competitive against Intel's big cores like Skylake. Remember AMD does not need to beat Intel but they have to be close behind in perf and perf/watt and provide competitive products at competitive prices. AMD's ace in the hole is HBM and AMD's APUs will start using HBM from 2017. AMD has confirmed a multi teraflop server APU with a transformational memory architecture in 2017. No prizes for guessing its a HBM enabled monster server APU. We can expect slightly scaled down versions for desktops and notebooks. Those APUs are potentially disruptive products. All we need is a competitive high performance Zen core and AMD is well on its way to financial health and sustainability.

It wasn't just bulldozer. It was that they did those chips then seemingly completely ignored the segment. No work done, all gone to APUs. They had these strong processors and never even updated the chipsets for them. They are STILL on USB 2 on those. Based on the motherboard choices alone those chips can get avoided. The decision to start going so strongly on APUs isn't something I understand, at least it doesn't make sense till next year I think.

That business was also too far from their server business. They can use server tech in high performance CPUs and vice versa, but APUs? bad idea imo. All they had to do was make their CPUs more attractive and they would sell. most sane people would take a good 8 core at <$200 over what intel offers, but AMD simply dropped improving those and went to try competing where intel wasn't (and now is).

AMDs current business model isnt substainable financially. So something have to happen. AMD got healthy parts and parts without a future. No need for the abd parts to drag down the healthy.

Yeah yea, AMD should stop making chips and just go into the food industry.

Companies don't spin off their best aspects. They aren't as strong in CPUs currently as they are in GPUs. It would be suicide till they get back strong on CPUs anyway. Especially since they are on the verge of a new era in gaming where they have a better architecture than the competition and VR is on the verge of being the thing (Dx 12 and VR + all consoles have their tech). If zen fails, maybe all the doom and gloom will make sense.

It would be like sony throwing away the playstation business. Crazy talk

 

[Nhirlathothep]

i think they re gonna bankrupt (together or separately)

i don t like this, but i m pretty sure

 

[Azix]

i think they re gonna bankrupt (together or separately)

i don t like this, but i m pretty sure

This is what the article said

The deliberations are preliminary and no decision has been made, the people said. The review highlights Chief Executive Lisa Su's determination to consider every possible option to turn the company around.

AMD had explored such a move in the past and decided against it, the people said. Su, however, who took over as CEO last October, judged that there is merit for the company to at least consider such a possibility again, the people added. There is no certainty that a split or spin-off will occur, the people cautioned.

This is just more speculation. Suppose AMD does really well with the new cards?

 

[xthetenth]

Remember IPC increase cost power consumption.

Luckily bulldozer is so bad that you spouting generalities doesn't necessarily mean that's the case in this instance.

 

[krumme]

The wsa with mubadala and the x86 agreement with intel makes everything completely dark for us and everyone that dont know the details in those contracts and agreement - and amd business.
And the opportunities.
All at the same time. In one person with inside knowledge.
And can handle that complexity

How many is that?
Hmm
Seems like the perfect job for a consultancy firm that knows nothing about either and therefore can come to conclusions in no time.

 

[IEC]

From Ars Technica's "article" on it:

Updated June 21, 1720 EST (2220 BST): AMD Spokesperson Sarah Youngbauer issued a statement over the weekend denying Reuters' report. She wrote, "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."

Source:
http://arstechnica.com/business/2015/06/amd-weighing-a-business-break-up-or-spin-off-reuters-says/

Consider this rumor started by panicked short sellers of AMD stock to be officially debunked.

 

[ShintaiDK]

From Ars Technica's "article" on it:



Source:
http://arstechnica.com/business/2015/06/amd-weighing-a-business-break-up-or-spin-off-reuters-says/

Consider this rumor started by panicked short sellers of AMD stock to be officially debunked.

http://www.eweek.com/c/a/Desktops-and-Notebooks/AMD-Denies-Fab-Sell-Off

Sounds familiar?

AMD has to deny it, nomatter if they do it or not.

 

[IEC]

That's a very different situation, and a very different impetus for the move they made then.

But feel free to compare apples and oranges.

 

[frozentundra123456]

It wasn't just bulldozer. It was that they did those chips then seemingly completely ignored the segment. No work done, all gone to APUs. They had these strong processors and never even updated the chipsets for them. They are STILL on USB 2 on those. Based on the motherboard choices alone those chips can get avoided. The decision to start going so strongly on APUs isn't something I understand, at least it doesn't make sense till next year I think.

That business was also too far from their server business. They can use server tech in high performance CPUs and vice versa, but APUs? bad idea imo. All they had to do was make their CPUs more attractive and they would sell. most sane people would take a good 8 core at <$200 over what intel offers, but AMD simply dropped improving those and went to try competing where intel wasn't (and now is).



Yeah yea, AMD should stop making chips and just go into the food industry.

Companies don't spin off their best aspects. They aren't as strong in CPUs currently as they are in GPUs. It would be suicide till they get back strong on CPUs anyway. Especially since they are on the verge of a new era in gaming where they have a better architecture than the competition and VR is on the verge of being the thing (Dx 12 and VR + all consoles have their tech). If zen fails, maybe all the doom and gloom will make sense.

It would be like sony throwing away the playstation business. Crazy talk

What "new era in gaming"? You mean when their new flagship is card is finally available that will probably trade blows with the GTX980Ti that is already on the market?

I also dont know what "better architecture" you are talking about unless it is HBM. But that is for a few Halo products, and they limited themselves to 4gb. What they need is a better more efficient architecture through the whole product stack, not just what I will generously call "refreshes" of rebrands, while nVidia has had a whole new, more efficient architecture on the market for over a year. And nVidia will have HBM as well at some point.

 

[Elixer]

It would be like sony throwing away the playstation business. Crazy talk

Pretty much this, it is crazy talk.

Right now, AMD is sitting on some good tech, that seems to be a winner.
That fury line is just their first step in the revamp of AMD.
Then, AMD has Zen in the pipeline, which has some solid tech people involved with it.
Then you will have Fury cards with 8/16/32GB of HBM 2.

That could be a good one-two punch to bring in the much needed $$$ they have been missing out.

The only thing we don't know is yields, if they can solve that issue, AMD might be back in the double digits sometime early '17, or, if sales are stellar for Fury, mid to late '16.