[ Bloomberg ] AMD Facing Bleak Future

[NTMBK]

Zen is not even taped out yet. It could for the matter of argument still be cancelled.

For the matter of argument, so could Cannonlake. Heck, if we're talking crazy hypotheticals, then for the matter of argument Intel could scrap all of their existing product line and decide to launch a 14nm shrink of Prescott. We have just as much evidence of that. :thumbsup:

 

[ShintaiDK]

For the matter of argument, so could Cannonlake. Heck, if we're talking crazy hypotheticals, then for the matter of argument Intel could scrap all of their existing product line and decide to launch a 14nm shrink of Prescott. We have just as much evidence of that. :thumbsup:

My point is AMDs focus is ARM. x86 is simply secondary.

 

[mrmt]

And where are you getting that from? No mention of that in the report.

Why would GLF give a WSA waiver to AMD if the same node was available at GLF on the same time frame as of Samsung?

It doesn't need to be a massive Haswell-killer core to be successful, necessarily. Something not much larger than the cat cores would probably be plenty for laptops- especially since the market is tending towards thin and light devices- and most consumer desktops.

Did they? We're not talking about bulky desktops for that matter, but about thin and light devices, and lots of them operating on a battery. AMD needs killer performance/watt and killer power gating, both things very expensive to develop, probably far beyond their paltry R&D budget can afford, unless you focus on a very small market bracket, just like bobcat did (at the expense of not being able to escalate up or down very easily).

I certainly agree that it will be difficult for AMD and I have my doubts about whether they can succeed, but that is a long way from completely writing it off.

I think we should have the opposite line of reasoning. What evidences we have that Zen might be something good? That's right, nothing, except AMD PR kool-aid. All the indications we have from AMD is that they lost a lot of talent, have a lot less money to develop products and are a lot less committed to the PC market than they were years ago. Why should we expect the next big thing on the PC market from them, when they are clear when they state they want to grow outside this market?

 

[NTMBK]

Why would GLF give a WSA waiver to AMD if the same node was available at GLF on the same time frame as of Samsung?

Free transfer of customers between the foundries could be stipulated as part of the GloFo licensing agreement, for all we know. Given that GloFo are now entirely dependent on Samsung for their 14nm tech, Samsung will have a not inconsiderably amount of leverage over them. There's more than one potential explanation. *shrug*

I think we should have the opposite line of reasoning. What evidences we have that Zen might be something good? That's right, nothing, except AMD PR kool-aid. All the indications we have from AMD is that they lost a lot of talent, have a lot less money to develop products and are a lot less committed to the PC market than they were years ago. Why should we expect the next big thing on the PC market from them, when they are clear when they state they want to grow outside this market?

I am not predicting that Zen will be great- I am not predicting anything, that's what I have been trying to say! I am keeping an open mind and not pre-judging. There is almost no information available, so any prediction I make at this point will be almost complete speculation (and no doubt informed by my personal biases more than anything else). I am waiting, and we will find out how good/bad Zen is when it actually arrives. The fact we don't even know what market segment it is aimed at says quite a lot!

I hope for a cheap, competent core. But I have no idea what we will actually get, and neither do you. None of us do.

 

[Gikaseixas]

IMO AMD has one last chance in 2016-17 and we will see how that project goes for them. Anything said here, either positive or negative about the outcome is pure speculation.

If SS buys AMD it means the end of the road for their x86 venture but might spark an all out war with Nvidia in graphics and another with ARM-y (provided SS would be acquiring the CPU engineering team)

 

[mrmt]

Free transfer of customers between the foundries could be stipulated as part of the GloFo licensing agreement, for all we know. Given that GloFo are now entirely dependent on Samsung for their 14nm tech, Samsung will have a not inconsiderably amount of leverage over them. There's more than one potential explanation. *shrug*

Samsung fighting for a small customer like AMD? That would be a first.

There is almost no information available, so any prediction I make at this point will be almost complete speculation (and no doubt informed by my personal biases more than anything else).

There's almost no information available about Zen itself, but there's plenty of information available showing a deteriorated project environment at the company (R&D cuts, personnel cuts, marketing cuts, management in disarray, etc) that ought to impact negatively the project.

If SS buys AMD it means the end of the road for their x86 venture but might spark an all out war with Nvidia in graphics and another with ARM-y (provided SS would be acquiring the CPU engineering team)

Why would Samsung buy AMD and all their liabilities when they could just scoop the engineering teams for free, just like they did two years ago with a cat core team?

 

[Arachnotronic]

Why would Samsung buy AMD and all their liabilities when they could just scoop the engineering teams for free, just like they did two years ago with a cat core team?

This. Apple also got a lot of their CPU engineers as well as reportedly a bunch of GPU engineers from AMD.

http://www.macrumors.com/2013/05/29...es-for-an-orlando-gpu-team-still-hiring-more/

 

[skipsneeky2]

Amazing how fast such a thread that isn't 12 hours old picked up 3 pages.:awe: If it was some random game performance thread it would have been a few more days i am sure.

These doom and gloom threads really pick up momentum that is unrivaled.

On topic this whole Amd and the buh bye situation has been played going back to faildozer.

 

[Arachnotronic]

On topic this whole Amd and the buh bye situation has been played going back to faildozer.

Since the launch of Bulldozer, AMD's revenue, profit, and balance sheet have deteriorated significantly.

Company deaths, should they happen, aren't usually sudden. They're painful and slow because the management teams will fight tooth and nail, using every trick in the book to try to make it to the next quarter.

AMD isn't just going to "roll over and die" -- it's going to keep fighting until it either succeeds in transforming the company or until they simply can no longer afford to keep the lights on.

Whether AMD succeeds or not is a big unknown. I think if the company plays its cards right and tries to tackle projects of appropriate scope for its R&D budget, then there's probably a decent future ahead of it.

 

[Dribble]

We all need Samsung to buy AMD, it's a win win win scenario!

Think it would be much cheaper and easier for Samsung to open offices near to AMD dev centres and suck all the talent out of them. Then they get the important stuff (the good people) without all the baggage.

 

[Rezist]

Well it's not like Intel is going to increase performance much over the next 5 years over the 2500k, we could see probably 15% speed increase and 35% power reduction over 5ish years.

It's not that bad though when you consider you can keep your rig for 8+ years and just upgrade GPU's.

 

[john5220]

https://www.google.com/webhp?sourceid=chrome-instant&ion=1&espv=2&ie=UTF-8#q=INTC
Market cap
182.13B

Qualcomm Market Cap
Market cap
121.98B
I won't comment on where I see the two companies in the future because they aren't exactly direct competitors. I don't think one's success means anothers downfall, Intel still holds the x86 desktop/laptop/miniPC/business sector while ARM can get servers/Smartphones/other products that need a small chip. While Intel can get into all of Qualcomm's markets, Qualcomm can't get into Desktops/Laptops.

I'm not familiar with Excavator but if it's going to be the same tech in the APU that will be using HBM, I think that will be the huge think AMD needs. AMD needs a viable APU and it's so close to getting there. Then couple that with a marketing strategy that toots it's own horn about PS4/XboxOne/WiiU(doesn't matter that WiiU isn't an APU from AMD, no one freaking cares, APple would have already told the world twice over that it's in every major console every single day yet AMD, a company that also has a huge gaming customer base that is naive as anything doesn't use this? Why won't they market... Why?!)
---------------------------------------------------


I think that AMD's market position is bad but that doesn't mean their future is bad. That article feels exceptionally weak and doesn't touch on any company's products. It's like the writer simply glanced at some financial highlights of each company.

If I'm not mistaken, AMD's partnership with Samsung brings 14nm to the table soon. AMD doesn't have more performance than Intel's CPUs but it's close. A die shrink and some architecture change could put them neck and neck. I'd say their future looks bleak if 14nm doesn't pan out.
Not to mention intel's "too fast" move to 14 nm has delayed some products which gives AMD some more time to get their 14nm on the table.
Then they're also moving into ARM processors with their GPU knowledge, they may have come competitive processors.

The largest issue AMD faces is convincing anyone to use their product. No matter how great a product AMD makes on the CPU or GPU side, they fail to market that product successfully to a wide consumer base. That's the largest issue they're facing right now.

So ya, I agree with the other users, nothing new posted here. Fool did some great articles on Skylake and intel's position in the market moving forward. This seems like a joke in comparison.
http://www.nasdaq.com/article/intel...s-a-bigger-deal-than-you-might-think-cm420203

I would have to say my thoughts about you have been wrong.

You are indeed a very knowledgeable man. I taught you were some sort of AMD hater and intel lover but I was wrong, your posts are very very informational and I enjoy reading them.

I hope you will accept my apology on how I responded rude to you earlier on another thread.

 

[tential]

I would have to say my thoughts about you have been wrong.

You are indeed a very knowledgeable man. I taught you were some sort of AMD hater and intel lover but I was wrong, your posts are very very informational and I enjoy reading them.

I hope you will accept my apology on how I responded rude to you earlier on another thread.

Read the thread. People have already elaborated on the mistakes I posted in that post you quoted.

Don't want you to think everything is 100% accurate in that post, AMD's roadmaps/future aren't my forte.

 

[exar333]

But that's not what Samsung needs. Samsung needs *competitive* x86 parts, something that AMD isn't really able to deliver by themselves since a few years ago. Building a custom core could solve this problem, if adequately funded, but Samsung market share isn't big enough to make this a viable venture and, and we would be 4 years down the road, because whatever AMD is developing now is the result of the resource-choked pipeline of the Rory era, probably not enough for Samsung needs.



The WSA does prohibit AMD from manufacturing chips in other foundries, that's why AMD must get a waiver from GLF every time they want to use TSMC. With GLF stance getting tougher (they forced AMD to backport their 28nm designs to its process and are bringing even GPUs there), I really doubt they would allow AMD to use Samsung as a foundry.

All this makes it clear why AMD went fabless in the first place. Debt. The idea of more 'flexibility' and 'agility' to cut costs and select the fab was all smoke and mirrors...

 

[john5220]

Well it's not like Intel is going to increase performance much over the next 5 years over the 2500k, we could see probably 15% speed increase and 35% power reduction over 5ish years.

It's not that bad though when you consider you can keep your rig for 8+ years and just upgrade GPU's.

could be even longer.

Say you bought a i3 today like me, and lets say this lasts me 5 years because at this rate it seems that it will. By then instead of swapping motherboards one could buy a used Devils Canyon on ebay for cheaps? And then you would be good to go for even longer maybe another 5 to 7 years.

I imagine it will take a very very long time for games to utilize 16 threads on a Devils Canyon. For all we know we could be looking at 15 year lifespan on a motherboard these days. Assuming the capacitors and other stuff hold up that long?

 

[jdubs03]

Well it's not like Intel is going to increase performance much over the next 5 years over the 2500k, we could see probably 15% speed increase and 35% power reduction over 5ish years.

It's not that bad though when you consider you can keep your rig for 8+ years and just upgrade GPU's.

I think we should see more than that increase over 15% over 5 years, that's 3% a year (single-thread). That would be quite disappointing. Plus the power reduction should follow along with the reduction in transistor size, so I think that may be a pessimistic perspective on the next 5 years, keep in mind this may cover all the way to the beginnings of 7nm. We'll see higher drive currents, which should enable faster switching so we should see improvements above the numbers you're saying; for all SKU segments. Otherwise, Intel's comments that Moore's Law can still be attained for the foreseeable future is complete bs.

 

[john5220]

Read the thread. People have already elaborated on the mistakes I posted in that post you quoted.

Don't want you to think everything is 100% accurate in that post, AMD's roadmaps/future aren't my forte.

ok yes thanks I am reading it now

 

[witeken]

Think it would be much cheaper and easier for Samsung to open offices near to AMD dev centres and suck all the talent out of them. Then they get the important stuff (the good people) without all the baggage.

I'm not so sure of that. For example, there's a reason why Intel bought Infineon instead of doing some version of what you say.

 

[Arachnotronic]

I'm not so sure of that. For example, there's a reason why Intel bought Infineon instead of doing some version of what you say.

Time to market. Intel needed to build a leadership comms group ASAP. It didn't have the luxury of building an organic effort from scratch.

With ImgTec/ARM licensing perfectly good CPU/GPU, there is no urgency to acquire ready-made tech.

 

[turtile]

Zen definitely won't be as powerful as Intel cores. They are replacing the entire lineup with Zen. Currently, they are using Puma+ and Excavator to fill the market. Puma+ performs pretty well so a much faster replacement with a higher clock speed should have no problem getting into servers.

Now that they just have Zen, the engineers only have to concentrate on two architectures instead of three.

 

[oobydoobydoo]

AMD hasn't been competing for high end destktops since 2011, I am not sure how this is news apart from obvious head nod to Samsung as Intel's new "big competitor". People have been saying AMD is doomed for what seems like centuries... I have my doubts.


Intel has little to worry about from Samaung, but a lot to worry about from Qualcomm and Apple. Qualcomm is pretty much ubiquitous in any high end smartphone not made by Apple, and amazingly Apple is nearly matching Broadwell -U performance with A8X in a much tighter thermal and power envelope. It already beats Core M in multicore.


Here is Broadwell- U
http://browser.primatelabs.com/geekbench3/search?dir=desc&q=ID1604+&sort=multicore_score

Here is Core M
http://browser.primatelabs.com/geekbench3/search?utf8=✓&q=5Y10

And here is A8X, keep in mind Apple is drawing 1/3-1/4 the wattage of a U part while throttling less
http://browser.primatelabs.com/geekbench3/1061742


Even Samaung is beating Core M in multicore with their latest Exynos 5433. These numbers do not tell the whole story though, because Intel's superior process already gives it an advantage that they appear to have squandered with poor design decisions like keeping that 32nm PCH. Intel also is working with a much larger (17W?) TDP.


With samaung's profits in free fall and its Galaxy series and tablet tanking, I would look to Qualcomm and Apple for future innovation. Intel is clearly a server margin-sustained dinosaur that will one day need to be culled, but for now we will keep in a Jurrasic Park of sorts: x86.


Don't count AMD out tho, they do great work with Video Cards. IMO superior to nvidia in hardware.

 

[SPBHM]

I still see some "enthusiasts" buying the FX line, but outside of that I almost exclusively see Intel, and when I see AMD it's in the worst possible way, like Bobcat/Kabini based thing with the lowest possible clock and poor usage (like a regular desktop, 15" 768p laptop); it's totally different from how it was 10 years ago, even during the k6 days AMD was more relevant,

and for a company with so many problems it can't be a good sign, but that's just my small observation...
I just hope they can have another great GPU release next year and keep competing well with Nvidia, as for the CPUs, I don't have much hope left,

 

[Headfoot]

Why would Samsung buy AMD and all their liabilities when they could just scoop the engineering teams for free, just like they did two years ago with a cat core team?

Absolutely this. It's the asset sale of asset sales. As strings-free as it gets, the only potential strings being noncompetes and trade secret misappropriation which are much, much murkier + weaker than the strings attached via a straight acquisition or merger.

 

[oobydoobydoo]

Absolutely this. It's the asset sale of asset sales.

Why would you think this would be good for anybody? You don't merge two failing companies and get a success. Add that to the completely different ideologies of the two companies, along with the knowledge that nobody who works at AMD lives anywhere near the ROK and you get what?


A bad idea, thats what. Horrible, even.

 

[frozentundra123456]

But isnt there a non-compete agreement with a lot of the engineers, at least higher level ones, that prohibit working for another company for a period of time after leaving a company, or are those illegal now?