Does AMD has some future?

[juanrga]

I am posting here because the AMD thread gives continuous server error 500.

Before Rory departure, my sources said me that K12/Zen was going to be made on 14FF and Keller was happy with the new node. However, AMD's SVP & CFO has announced that AMD will finally use 28nm for K12

http://seekingalpha.com/article/2739835-advanced-micro-devices-amd-2014-raymond-james-systems-semiconductors-software-and-supply-chain-conference

http://wsw.com/webcast/rj94/amd/

And if K12 is 28nm then I assume that Zen will be as well.

Adds that the AMD tablet project was silently canceled a week ago and that Carrizo D/E is in the air, and the conclusion is that AMD doesn't have a winning strategy for the future.

AMD doesn't have plans/projects for phones or tablets. The new 28nm K12/Zen will be uncompetitive on laptops and servers against 14/16nm parts. AMD has currently abandoned HEDT to Intel, and there is no short term plans for supercomputers. The promised new semicustom wins are never announced. What market remain to AMD? The desktop GPU division, which even AMD admits is not in good shape? The semicustom division, which will only provide about 3 billions on revenue on next years? Is this the end of AMD?

 

[Karsten75]

This ain't fruit salad. But then, I don't care for fruit salad anyway. In history, many great conflicts were never resolved, but superseded by different conflicts arising from changed circumstances. The Intel vs AMD war is over - long ago. The Nvidia vs AMD war is drawing to a close. AMD will have to find a different field, and hopefully something else will challenge Intel and Nvidia so we don't end up with Microsofts in those areas as well. Intel is moving to stem the challenge from Exynos or Cortex or... Ax, Nvidia is already moving into parallel processing to take on Intel Phi processors. AMD faces an uncertain future at best.

 

[jdwii]

If this is true Amd is now losing their newer markets which isn't good. Perhaps Jim keller isn't enough to save this company from coming irreverent which i always thought was the case over budget issues.

Amd NEEDS 14nm and 16nm Intel is already ahead in the fabrication process and they have a efficient design, same goes for Nvidia. Amd either needs a die shrink or a new architecture that is SO GOOD that it can compete with a higher node.

Its not like people won't just pick other companies Amd isn't really competing with Intel today anyways Intel already feels pressure from Arm base designs. Back in the day we needed Amd now its really just a choice. I want Amd to stay relevant but what i want and what the truth is are two different things.

 

[Gam3r01]

Phenom III's. Thats all I want.

 

[leeb2013]

it won't be good for the consumer if AMD continues to fail to compete in either the high end desktop CPU or GPU market. We've seen the biggest leaps in tech and drops in prices when there's been stiff competition.

 

[-Fran-]

I'll re-phrase then... The future doesn't look boring. It looks awfully depressing.

Cheers!

 

[anxiousinfusion]

Perhaps Jim keller isn't enough to save this company from coming irreverent which i always thought was the case over budget issues.

Everyone touts Keller as some kind of beacon of hope. Sure, the guy is talented.. but what about the teams of engineers working with him. Don't they, too, make or break a design?

 

[CDdude55]

It's kind of a banality at this point as this really started to occur since the Core 2 Duo era and people started to share these same sentiments since then. I think AMD might be taking the position that the effort is no longer worthwhile and is instead embracing the lower and mid end markets. And they've actually done a great job in that regard so they might feel that their place, while not the best, is still sustainable.

I do wish AMD tried to throw more resources into a new architecture, but Intel is so far ahead at this point.

 

[jdwii]

Perhaps Jim keller isn't enough to save this company from coming irreverent which i always thought was the case over budget issues.

Everyone touts Keller as some kind of beacon of hope. Sure, the guy is talented.. but what about the teams of engineers working with him. Don't they, too, make or break a design?

Of course his team deserves a lot of respect however the teams he leads typically have great success he is kinda a BA when it comes to engineering but that doesn't mean money doesn't stand in his way. CPU architecture isn't easy and it requires a lot of money and time to come up with newer designs that are worth it. If Amd doesn't have enough money for research we will know it when we review their products.

Back in 1999 Amd was going bankrupt and they designed the first athlon(K7) a winner and the first time Amd really showed they could compete with Intel however at that time Intel wasn't trying as hard and CPU's were a bit less complicated.

I own the fx 8350 i owned the phenom II x6 and the athlon II x4 and i own a Amd A8 3520m. I owned the 6850hd-6950hd-4770hd. That being said i want Amd to succeed but that doesn't mean i'm going to buy inefficient parts or weak performers just to give them my money doing so would be foolish.

Actually right now i'm building a 400$ Rig with a A10 7850K and DDR3 2400 ram

 

[iron8orn]

http://www.tweaktown.com/news/40339/amd-to-release-its-radeon-r9-390x-early-next-year-20nm-with-hbm-tech/index.html

sure they have a future and why not some 20nm cpu's to match

 

[leeb2013]

Perhaps Jim keller isn't enough to save this company from coming irreverent which i always thought was the case over budget issues.

Everyone touts Keller as some kind of beacon of hope. Sure, the guy is talented.. but what about the teams of engineers working with him. Don't they, too, make or break a design?

It's very rare that engineers make or break a design, or company. The output is mainly dependant on financing, headcount, paying enough to inspire good engineers, good corporate policy, good management with stretched but realistic targets that support and understand their teams well.

Unfortunately, as is usually the case, companies want to cut cost, reduce headcount, cut pay, have inexperienced management which pull the teams every which way with ever changing unrealistic targets, who think they are above listening to the engineers, poor corporate structure, reduced benefits, constant threat of redundancy or offshoring, creating demoralized, ineffective teams. That's what screws tech companies, not the limitation of engineering knowledge and skill.

 

[juanrga]

The old thread continues broken. This is a snapshot of the server error that gives when trying to post

http://postimg.org/image/631z2yu9v/

Credit: jdwii

 

[juanrga]

It seems that AMD has confirmed to the above tech writter that AMD vicepresident mispoke during the conference and that K12 still targets a FinFET node

http://seekingalpha.com/article/2739835-advanced-micro-devices-amd-2014-raymond-james-systems-semiconductors-software-and-supply-chain-conference#comment-44600805

I e-mailed AMD, and it looks like User 12115671 was correct and Devinder did indeed misspeak: K12 is targeted for a FinFET node.
--AE

Although I would like to see an official claim about this. In any case if this is official confirmed, I still find weird that the senior vicepresident of AMD makes this kind of mistake about flagship products of the company.

 

[gamerk316]

Perhaps Jim keller isn't enough to save this company from coming irreverent which i always thought was the case over budget issues.

Everyone touts Keller as some kind of beacon of hope. Sure, the guy is talented.. but what about the teams of engineers working with him. Don't they, too, make or break a design?

It's very rare that engineers make or break a design, or company. The output is mainly dependant on financing, headcount, paying enough to inspire good engineers, good corporate policy, good management with stretched but realistic targets that support and understand their teams well.

Unfortunately, as is usually the case, companies want to cut cost, reduce headcount, cut pay, have inexperienced management which pull the teams every which way with ever changing unrealistic targets, who think they are above listening to the engineers, poor corporate structure, reduced benefits, constant threat of redundancy or offshoring, creating demoralized, ineffective teams. That's what screws tech companies, not the limitation of engineering knowledge and skill.

Q
F
T

Seriously, you said it perfectly.

 

[juanrga]

Through a sell-them-cheap strategy because, as stated in the article, "AMD isn’t going head to head with Intel on vPro as AMD doesn’t have the investment power to go toe to toe with Intel’s enterprise client investments."

My question is: How many more years can AMD sustain that strategy? Selling cheap implies less money available for R&D of future products, which means future poor products than competence which implies that will have to compete again on prize reducing more the R&D budget and the spiral continues out.

 

[juanrga]

Antitrust lawyers acted then, and rightfully, when Intel abused of its position. The current situation is different and AMD complete irrelevance on many strategic markets is a consequence of a large sequence of mistakes made by AMD: strategic, technological, and execution mistakes.

Intel no longer seems to care about maintaining AMD’s market share and has pounded it into the ground in most markets, notably server, where Intel’s unit share hit 98.3% in 3Q14, according to Mercury Research’s PC Processor Report, and 98.5%, according to IDC. In notebooks, Mercury’s 3Q14 figure is 92.9%, while IDC pegs it at 90.3%. Mercury says Intel’s desktop share was 82.7%, while IDC puts it at 81.8%, but those desktop figures are small comfort to AMD, since desktop is the least profitable of the three segments. At this point, Intel’s revenue is an order of magnitude larger than AMD’s, and its market cap is nearly two orders of magnitude greater.

http://www.forbes.com/sites/rogerkay/2014/11/25/intel-and-amd-the-juggernaut-vs-the-squid/

AMD is afloat thanks to the sell-them-cheap strategy. This week they dropped the price of the Firepro to compete against Nvidia on professional markets because otherwise AMD cannot compete. This strategy will work today, and tomorrow, and next month, and next year,... but in the end AMD will be cornering itself up to transforming into the new VIA by reasons stated above.

 

[juanrga]

It is stated in the same article why Intel doesn't care anymore about AMD:

the main reason that Intel no longer obsesses on AMD is that both companies are finding themselves competing in a much larger arena: the total market for endpoints, which includes phones, tablets, and a burgeoning array of other nodes that already or soon will include Internet-of-Things devices (like thermostats and drawbridges), automobiles, and houses. In the mobile market, vendors shipping ARM architecture devices move billions of units in contrast to the few hundreds of millions in the x86 market.

Intel real competitors are others (e.g., Qualcomm).

 

[bmacsys]

I am posting here because the AMD thread gives continuous server error 500.

Before Rory departure, my sources said me that K12/Zen was going to be made on 14FF and Keller was happy with the new node. However, AMD's SVP & CFO has announced that AMD will finally use 28nm for K12

http://seekingalpha.com/article/2739835-advanced-micro-devices-amd-2014-raymond-james-systems-semiconductors-software-and-supply-chain-conference

http://wsw.com/webcast/rj94/amd/

And if K12 is 28nm then I assume that Zen will be as well.

Adds that the AMD tablet project was silently canceled a week ago and that Carrizo D/E is in the air, and the conclusion is that AMD doesn't have a winning strategy for the future.

AMD doesn't have plans/projects for phones or tablets. The new 28nm K12/Zen will be uncompetitive on laptops and servers against 14/16nm parts. AMD has currently abandoned HEDT to Intel, and there is no short term plans for supercomputers. The promised new semicustom wins are never announced. What market remain to AMD? The desktop GPU division, which even AMD admits is not in good shape? The semicustom division, which will only provide about 3 billions on revenue on next years? Is this the end of AMD?

There are tens of thousands of these "AMD death" threads. The same stuff is rehashed over and over and over. If you have real sources within the company juanrga why do you need to start these useless threads with people who obviously don't know the answers either? No offense to any of the posters. Nobody outside of AMD can answer any of this.

 

[juanrga]

The article may be right but how long can intel afford to payout incentives to keep market share.

http://www.bloomberg.com/news/2014-12-04/intel-led-industry-doubles-debt-as-growth-options-dwindle-tech.html

"The flood of debt to fund shareholder returns is a major change for companies that historically hoarded cash and avoided borrowing. While payouts help burnish stock prices, some investors say that money would be better used on other methods of boosting future growth, such as acquisitions and investment in research and development -- spending that can help companies keep up with the brisk pace of technological innovation."

I do not subscribe to the intel not caring theory, Intel knows that one good product stake can change the tables. APU's have changed the whole landscape and Intel knows this. If not we would still have hd 2000 graphics,
with the flood of ddr4 memory hitting the market apu's will be even a better option.

Intel is the underdog on mobile. They tried to get tablet market share by incentivizing OEMs because x86 cannot compete against ARM. But you are right on that Intel mobile division cannot lose 1 billion per quarter trying to maintain artificial tablet market share and that is why Intel is apparently abandoning mobile pretensions.

Of course, that are not good news for AMD because this means that Intel will increase pressure on the traditional x86 market to get more revenue to compensate for the loses. This, added to AMD management myopia regarding ARM market evolution, will probably means AMD will be crushed by both sides: ARM and x86.

About APUs: It is Intel who first announced their intention to develop integrated graphics which would kill discrete graphics over time. Then AMD saw the wolf ears and decided to purchase ATI starting the FSA initiative (Fusion), latter renamed to HSA: "In many ways, this merger is a lot about AMD trying to compete with Intel"

http://www.anandtech.com/show/2055/2
http://www.anandtech.com/show/2055/3
http://www.anandtech.com/show/2055/5

Of course, without AMD APUs we would have poor Intel graphics, but it also goes in the other way: without Intel graphics we would have poor AMD APUs.

 

[bmacsys]

The article may be right but how long can intel afford to payout incentives to keep market share.

http://www.bloomberg.com/news/2014-12-04/intel-led-industry-doubles-debt-as-growth-options-dwindle-tech.html




Intel is the underdog on mobile. They tried to get tablet market share by incentivizing OEMs because x86 cannot compete against ARM. But you are right on that Intel mobile division cannot lose 1 billion per quarter trying to maintain artificial tablet market share and that is why Intel is apparently abandoning mobile pretensions.

Of course, that are not good news for AMD because this means that Intel will increase pressure on the traditional x86 market to get more revenue to compensate for the loses. This, added to AMD management myopia regarding ARM market evolution, will probably means AMD will be crushed by both sides: ARM and x86.

Again this makes no sense. AMD's sales are just a rounding error for Intel. A blip on the radar compared to Intel. Intel isn't going to do anything different in the x86 market to "compensate for losses". If Intel grabbed every single sale AMD makes it wouldn't change the ledger sheet one iota for them. How isn't AMD already being "crushed" by Intel?

 

[juanrga]

But who is intel fighting in the mobile market, ARM ? there is no arm running full windows so it must be amd. spending a bil dollars a qtr just to hold down amd tells me they are concerned about amd. Intel is losing apple and other big oem have signaled a desire to look at other options
"Intel fights to keep HP, Dell and other customers from defecting"
http://www.mercurynews.com/business/ci_21052802/intel-fights-keep-hewlett-packard-dell-from-defecting

Intel is fighting ARM since several years ago. AMD is irrelevant and not in the radar as also mentioned in the article. Intel is spending billions to get market share against ARM. For your information what Intel spends during a single quarter on OEM incentives for its mobile chips equals what AMD will earn with two semicustom business in a three year period.

Intel is not as safe as people think, only because of its backroom deals has it been able to maintain market share over amd.

Nobody said Intel is safe, but has the resources that AMD doesn't.

 

[juanrga]

Again this makes no sense. AMD's sales are just a rounding error for Intel. A blip on the radar compared to Intel. Intel isn't going to do anything different in the x86 market to "compensate for losses". If Intel grabbed every single sale AMD makes it wouldn't change the ledger sheet one iota for them. How isn't AMD already being "crushed" by Intel?

You didn't get it. For instance Intel is increasing lately the pressure on x86 servers because the ARMy is coming to servers not not because AMD is in the radar. AMD is irrelevant for Intel, it is Qualcomm and others who worry Intel. This is why I said "AMD will be crushed by both sides: ARM and x86". In fact AMD Seattle is already being crushed by Intel Atom and APM X-Gene.

 

[juanrga]

Posting here again because the AMD Speculation thread was broken again days ago and nobody fixes it...

I have words that Excavator introduces no change in FP unit and continues being a weak 8 FLOPS/core (confirming my prediction and disproving my critics), and that the integer IPC gains will be compensated by further reduction in base frequencies (probably around 2.0--2.5GHz).

AMD has reduced the pipeline to reduce power consumption (to try to compete against Broadwell) and increase IPC by eliminating part of the branch misprediction penalty associated to a longer pipeline. But this comes at the cost of further reductions in achievable frequencies.

As a consequence, the originally planned 65W Carrizo for FM2+ platform and the server part were canceled. Carrizo doesn't scale well above 35W and will be used only in a subset of laptops and low power desktops (AIO, SFF).

This is ridiculous, even the Denver core for tablets achieves 16 FLOPS/core. The dual core K1 @2.5GHz for tablets will produce more FP throughput than a quad-core FX Carrizo @2.1GHz for high-performance laptops...

7W Dual core K1: 80 FLOP/s
35W Quad-core Carrizo FX: 67.2 GFLOP/s

What I wonder is how many x86 laptops wins AMD expect to win using 28nm planar when competence is on 14nm FinFET. My guess is that we will see less Carrizo laptops than Kaveri laptops.

 

[anxiousinfusion]

When it became apparent that their AM3 line was retired, okay I get that.. the world is going integrated and they need to consolidate their socketed platforms. But now AMD are retiring their last desktop/socketed APUs? What are they thinking?

 

[juanrga]

When it became apparent that their AM3 line was retired, okay I get that.. the world is going integrated and they need to consolidate their socketed platforms. But now AMD are retiring their last desktop/socketed APUs? What are they thinking?

First abandoned servers/HPC, then abandoned HEDT, and now abandon the >35W APUs market. This looks as a long-term planned exit from the PC market towards semicustom only.

What I find amazing is it was not only two years ago that Intel had plans to ditch their socketed LGA platforms in favor of BGA platforms and at that time, AMD promised us that they were committed to prepare socketed CPUs for consumers and enthusiasts in the long run. Another broken promise.

Intel will release Broadwell and Skylake CPUs for the consumer socketed platforms soon, whereas AMD are canceling the promised socketed Excavator-based desktop platforms. We still see some (<35W) Carrizo desktop but in soldered BGA variants.