Discussion: AMD's last hope for survival lies in the Zen CPU architecture

[InvalidError]

Let's just say if AMD was to go out of business, what would happen then? How would Intel react then? Would anti-trust get involved? These are questions we can speculate about.

There might be a slight disturbance in the force but aside from that, nothing would happen: AMD has almost no market share left to speak of for CPUs and Intel has plenty of mostly better chips to pick up that slack, albeit at higher price points.

On the GPU side, ATI might get sold as a whole during liquidation assuming Microsoft or someone else with a significant vested interest in x86 does not buy AMD as a whole - this should be a licensing issue since the x86 license is still attached to AMD as a wholly-owned subsidiary or something else of the sort.

Not much would happen from Intel since their chips are already priced just about as high as the market can bear. With 5+ years old CPUs being more competitive with current CPUs than ever before, Intel is in a situation where it cannot quit doing R&D on new chips just because AMD is gone because they need to give people a reason to upgrade their systems. Unless Intel increases the pace, we may end up in a situation where Sandy Bridge may still be considered viable competition against Intel's 2020 lineup, which is not going to do Intel's sales much good.

Anti-trust wise, there is absolutely nothing to do here: anti-trust is about acquiring a dominant market share through questionable business practices. Here, AMD is falling victim to falling computer sales, failing to catch up with Intel on performance, mass adoption of ARM-based mobile and portable computing devices, etc. Even Intel is suffering from falling sales because of those factors and how slowly their own chips' performance is progressing.

If you want something to keep Intel in check, cross your fingers for ARM to become a credible threat. Before that can happen though, Android and Windows Mobile will need to reclaim more of their desktop roots.

 

[turkey3_scratch]

I don't see how ARM would be a threat to Intel. x86 is a more complex instruction set, and it's not like desktops/laptops/servers are ever going to switch to ARM in my mind at least.

 

[jimmysmitty]

I don't see how ARM would be a threat to Intel. x86 is a more complex instruction set, and it's not like desktops/laptops/servers are ever going to switch to ARM in my mind at least.

True but ARM can get more complex and become more powerful. But in doing so also negates the main advantage ARM has, low power solutions although Intel is closing that gap pretty well.

The more important reason why would be the 30 years of software support.

 

[jaguarskx]

I don't see how ARM would be a threat to Intel. x86 is a more complex instruction set, and it's not like desktops/laptops/servers are ever going to switch to ARM in my mind at least.

The market for ARM devices is huge and because these types of devices are relatively inexpensive there is a lot of turnover. You can look at them as "disposable devices", they are relatively inexpensive to the point where people buy new tablets and smartphones every year or more often than that. That equates to huge annual sale volumes. How often do people replace their laptops and desktops? I am pretty sure it is not as often as people change smartphones.

ARM processors are not a threat to Intel's CPUs from a performance point of view. But it is a threat when you look at from a consumer device marketshare point of view. ARM processors are in a lot of electronics and people generally own more devices that has an ARM processor vs an AMD / Intel CPU. They are used in smartphones, tablets, Roku and similar devices, Smart TVs, HDTVs, etc.

 

[seeingeyegod]

So much exaggeration and hyperbole in this thread. Can't AMD just borrow a few bill from Intel? They have like 80 Billion cash.

 

[Supernova1138]

So much exaggeration and hyperbole in this thread. Can't AMD just borrow a few bill from Intel? They have like 80 Billion cash.

Intel is not likely to do that unless Federal regulators effectively order Intel to put AMD on life support to avoid getting broken up in an anti-trust action. That probably isn't going to happen since there technically are alternatives to Intel aside from AMD, you'd just have to give up 30 years of backwards compatibility to use those alternatives.

 

[turkey3_scratch]

I don't see how ARM would be a threat to Intel. x86 is a more complex instruction set, and it's not like desktops/laptops/servers are ever going to switch to ARM in my mind at least.

The market for ARM devices is huge and because these types of devices are relatively inexpensive there is a lot of turnover. You can look at them as "disposable devices", they are relatively inexpensive to the point where people buy new tablets and smartphones every year or more often than that. That equates to huge annual sale volumes. How often do people replace their laptops and desktops? I am pretty sure it is not as often as people change smartphones.

ARM processors are not a threat to Intel's CPUs from a performance point of view. But it is a threat when you look at from a consumer device marketshare point of view. ARM processors are in a lot of electronics and people generally own more devices that has an ARM processor vs an AMD / Intel CPU. They are used in smartphones, tablets, Roku and similar devices, Smart TVs, HDTVs, etc.

Was x86 ever actually in those devices? did televisions, flip phones, and DVD players used to run on x86?

 

[InvalidError]

I don't see how ARM would be a threat to Intel.

ARM might not replace Intel for PC enthusiasts and some server applications but it could potentially replace PCs everywhere else where people aren't shackled to legacy x86 stuff.

Will my three years old niece or two years old nephew ever use a PC? By the time they enter grade schools, iPads will be mandatory in most public schools (it already is in half of public schools where I live), so they aren't going to learn what a conventional PC is there. Neither of my sisters nor their boyfriends use PCs either, so they aren't going to learn about PCs at home unless that changes. By the time they enter high-school, ARM-based netbooks will be far more powerful than any high-school homework might ever require and likely more than powerful to run substantial games, so there is a very high probability they won't encounter PCs there either. By the time they reach college, I would be extremely surprised if ARM did not grow a fully self-supporting ecosystem with full-blown OSes, SDKs and all the traditional desktop productivity stuff by then.

When I say "a threat to Intel," I do not mean for next year, or even 3-5 years from now. I mean ~10 years from now when the new generation of software developers will have little to no specific interest in x86. Chances they won't even know or care what CPU architecture is in their computer because most software will be in MSIL or other run-anywhere bytecode format.

 

[jaguarskx]

Was x86 ever actually in those devices? did televisions, flip phones, and DVD players used to run on x86?

You are missing the point.

As I have stated, the threat that ARM posses to Intel is not it's performance, but it's ubiquitousness. People are surrounded by more devices (and vehicles) that have ARM processors than AMD or Intel CPUs. And in many cases those devices are disposable electronics. The threat is marketshare and potential lost sales to electronic devices that do not require something as complex and expensive as AMD or Intel CPUs.

 

[Rodrigodrt]

So the real worry here is not about AMD going bye. But x86 x ARM / PC x Mobile?

 

[Hellowalkman]

All hope is not lost guys .. I found this in one of the new articles about AMD ..


Suzanne Plummer, the veteran Austin chip engineer who heads the Zen team, exudes confidence about the project.

“It is the first time in a very long time that we engineers have been given the total freedom to build a processor from scratch and do the best we can do,” Plummer said. “It is a multi-year project with a really large team. It’s like a marathon effort with some sprints in the middle. The team is working very hard, but they can see the finish line. I guarantee that it will deliver a huge improvement in performance and (low) power consumption over the previous generation.”

 

[jaguarskx]

So the real worry here is not about AMD going bye. But x86 x ARM / PC x Mobile?

That is just a side argument. The mobile device market is important to both AMD and Intel because there is a lot of potential revenue in the future. In 2014 Intel lost $4 billion in their mobile CPU division simply to expand their marketshare in the mobile device market. Intel will probably continue to lose money in that market for a few more years in order to expand, but losses should decline as they sell more Atom CPUs and lower their production costs.

AMD's primary focus it their Zen CPU and Radeon GPUs. I am sure they are still working on a low power mobile APU for the mobile device market to some degree. But they are in no position to take massive losses like Intel just to grab marketshare. At best AMD will only have an extremely tiny fraction of that market or they maybe even non-existent until they are financially strong enough take on an uphill battle.

 

[InvalidError]

AMD's primary focus it their Zen CPU and Radeon GPUs. I am sure they are still working on a low power mobile APU for the mobile device market to some degree. But they are in no position to take massive losses like Intel just to grab marketshare. At best AMD will only have an extremely tiny fraction of that market or they maybe even non-existent until they are financially strong enough take on an uphill battle.

Intel is fighting an uphill battle against ARM which already has an established dominance in mobile, in large part thanks to device manufacturers being able to source chips from a dozen chip designers so they do not need to worry much about conflicts with single-vendor components like Intel and arbitrary price increases.

If Intel actually manages to break significant ground with mobile x86, it might make the job easier for AMD since it would spare AMD from having to spend as much effort to convince device developers that x86 is relevant. It also makes AMD the default second source for any manufacturer running into issues while dealing with Intel or wanting to maintain a product line based on AMD chips just to keep Intel in check (nothing keeps vendors in line like knowing their clients can swap their parts out on relatively short notice) as long as AMD manages to produce comparable products.

 

[Hellowalkman]

Some interesting development going on at GlobalFoundries . Check this out .. http://www.kitguru.net/components/anton-shilov/globalfoundries-we-started-to-tape-out-products-using-second-gen-14nm-process-technology/

Apparently AMD is a customer but just a few days back rumor was going around that they have contracted TSMC to produce Zen on 16nm FF ...

 

[jaguarskx]

If Intel actually manages to break significant ground with mobile x86, it might make the job easier for AMD since it would spare AMD from having to spend as much effort to convince device developers that x86 is relevant. It also makes AMD the default second source for any manufacturer running into issues while dealing with Intel or wanting to maintain a product line based on AMD chips just to keep Intel in check (nothing keeps vendors in line like knowing their clients can swap their parts out on relatively short notice) as long as AMD manages to produce comparable products.

While inroads made by Intel can open the market to AMD in the mobile device arena, AMD does not really have a compelling APU for that market segment as far as I am aware of when it comes to performance and power consumption. Additionally, AMD is not in a financial position to sell mobile APUs at a loss like Intel is doing to become a contender in the mobile device arena; at least not on a large scale. As I stated in a previous post back in 2014 Intel took a $4 billion loss in the mobile CPU division primarily due to selling Atom CPUs at below cost.

 

[InvalidError]

Additionally, AMD is not in a financial position to sell mobile APUs at a loss like Intel is doing to become a contender in the mobile device arena; at least not on a large scale. As I stated in a previous post back in 2014 Intel took a $4 billion loss in the mobile CPU division primarily due to selling Atom CPUs at below cost.

AMD won't need to spend billions to convince mobile device developers to adopt their chips if Intel manages to break significant ground. Right now, most mobile developers are using ARM-based chips for all of their Android, Windows and other mobile devices while x86/Intel is little more than a novelty they are doing on-the-side on select few models in large part thanks to Intel's marketing and R&D subsidies because they would not bother with setting up a workflow for x86-based products otherwise.

If Intel manages to get into mobile, 90-95% of the required support effort for AMD x86 SoCs will already have been fielded by Intel.

 

[jaguarskx]

The point is AMD would need to sell their APUs for mobile devices at a loss just like Intel until the cost of manufacturing the APUs will at most be equal to the price AMD is selling them for. One way to lower the cost of production per unit is to manufacture in large quantities. The problem with manufacturing in large quantities is that it is not a viable business strategy unless AMD has proven their APUs can compete on a price / performance / power consumption ratios against ARM processors and Intel.

Back in May 2014 AMD released the Beema APU for low end laptops (like the A8-6410 APU), they also released the Mullins APU for mobile devices. To the best of my knowledge as of December 2014 there have been no design wins for the Mullins APU. Meaning there have been no announcements of any mobile device that will be using the Mullins APU. Mullins failed to impress on both performance and power consumption characteristics.

 

[InvalidError]

The point is AMD would need to sell their APUs for mobile devices at a loss just like Intel until the cost of manufacturing the APUs will at most be equal to the price AMD is selling them for.

Or AMD can back out of mobile x86 for the time being, let Intel loosen up the market and jump back in if demand for mobile x86 picks up enough to make it a potentially self-sustainable project instead of sinking tons of cash to generate demand from scratch in a market that is for the most part not interested (yet) as was the case for Mullins and still is with Intel's Atom.

AMD still being stuck at 28nm in a world of 14-22nm designs is not exactly ideal either.

 

[jaguarskx]

If I were AMD I would continue to devote limited resources to APU R&D for mobile devices simply to progress the technology. I would not bother with serious R&D until AMD is financially sound to do, but unfortunately that could mean waiting until after 2019 when (and if) AMD can reduce it's debt load which can allow them borrow some money to devote mobile device APU R&D assuming their balance sheet is in decent shape.

The problem is it will be at least 4 years before they can realistically issue more bonds to raise cash. A lot can happen to the mobile device industry in 4 years. At that point in time I would expect Intel to be close to breaking even with their Atom CPUs (cost per unit = selling price per unit).

 

[InvalidError]

At that point in time I would expect Intel to be close to breaking even with their Atom CPUs (cost per unit = selling price per unit).

At the rate Intel is putting all the stuff they ripped out of of desktop CPUs to initially shoehorn x86 into mobile-friendly power budgets back into Atom to keep up the pace with ARM CPUs and the rate at which ultra-low power desktop/laptop chips are extending into ultra-mobile power budgets (Broadwell goes down to 3-5W), I would not be too surprised if Intel went through another cycle of consolidating mobile and desktop architectures, much like what they did with laptop and desktop with Core 2 after Netburst's brute force approach turned into a dead-end.

If AMD plays it smart, they should have designed Zen with the ability to scale power down in mind to avoid having to design a whole new core specifically for mobile. At the rate ARM is picking up performance optimization tricks that Intel had to leave behind when they started the Atom product line to meet the power budget, significantly cut down x86 cores won't be able to compete against ARM a few years from now.

 

[gamerk316]

Additionally, AMD is not in a financial position to sell mobile APUs at a loss like Intel is doing to become a contender in the mobile device arena; at least not on a large scale. As I stated in a previous post back in 2014 Intel took a $4 billion loss in the mobile CPU division primarily due to selling Atom CPUs at below cost.

AMD won't need to spend billions to convince mobile device developers to adopt their chips if Intel manages to break significant ground. Right now, most mobile developers are using ARM-based chips for all of their Android, Windows and other mobile devices while x86/Intel is little more than a novelty they are doing on-the-side on select few models in large part thanks to Intel's marketing and R&D subsidies because they would not bother with setting up a workflow for x86-based products otherwise.

If Intel manages to get into mobile, 90-95% of the required support effort for AMD x86 SoCs will already have been fielded by Intel.

Call me when someone comes out with a viable OS that runs on X86 on mobile. Because outside of Windows Mobile, there aren't any.

See the chicken and egg problem here? When it comes to mobile, Wintel lives and dies together.

 

[gamerk316]

If I were AMD I would continue to devote limited resources to APU R&D for mobile devices simply to progress the technology. I would not bother with serious R&D until AMD is financially sound to do, but unfortunately that could mean waiting until after 2019 when (and if) AMD can reduce it's debt load which can allow them borrow some money to devote mobile device APU R&D assuming their balance sheet is in decent shape.

The problem is it will be at least 4 years before they can realistically issue more bonds to raise cash. A lot can happen to the mobile device industry in 4 years. At that point in time I would expect Intel to be close to breaking even with their Atom CPUs (cost per unit = selling price per unit).

Here's the problem: 2019 is when AMD magically needs $600 Million on hand to meet its debt obligations. AMD's been taking out debt to pay it's current debt for years, and they simply can't do it anymore. AMD needs CASH, and the last new product design before then is Zen.

Why do you think they keep cutting people? Who do you think they are still using the GCN architecture on their GPUs? Because they're running out of money. EVERYTHING comes down to making a couple hundred million in profits per year on Zen. To say "AMD can borrow money in four years and everything is OK" ignores the fact AMD is teetering very, very close to bankruptcy.

 

[InvalidError]

Call me when someone comes out with a viable OS that runs on X86 on mobile. Because outside of Windows Mobile, there aren't any.

Android runs fine on x86 and most applications work too, just like how most Windows Mobile applications work just as well on both x86 and ARM devices.

The joys of Dalvik and MSIL bytecode.

 

[jimmysmitty]

At that point in time I would expect Intel to be close to breaking even with their Atom CPUs (cost per unit = selling price per unit).

At the rate Intel is putting all the stuff they ripped out of of desktop CPUs to initially shoehorn x86 into mobile-friendly power budgets back into Atom to keep up the pace with ARM CPUs and the rate at which ultra-low power desktop/laptop chips are extending into ultra-mobile power budgets (Broadwell goes down to 3-5W), I would not be too surprised if Intel went through another cycle of consolidating mobile and desktop architectures, much like what they did with laptop and desktop with Core 2 after Netburst's brute force approach turned into a dead-end.

If AMD plays it smart, they should have designed Zen with the ability to scale power down in mind to avoid having to design a whole new core specifically for mobile. At the rate ARM is picking up performance optimization tricks that Intel had to leave behind when they started the Atom product line to meet the power budget, significantly cut down x86 cores won't be able to compete against ARM a few years from now.

I thought that was the plan? To slowly rebuild their low power chips back to desktop level while having the power sipping properties of Atom.

I mean the Haswell mobile top end quad core CPU was almost as powerful as the i7 4770 but could also clock down to ultra low power levels.

Of course this is also what AMD should do. It would benefit them in the long run.

 

[gamerk316]

At that point in time I would expect Intel to be close to breaking even with their Atom CPUs (cost per unit = selling price per unit).

At the rate Intel is putting all the stuff they ripped out of of desktop CPUs to initially shoehorn x86 into mobile-friendly power budgets back into Atom to keep up the pace with ARM CPUs and the rate at which ultra-low power desktop/laptop chips are extending into ultra-mobile power budgets (Broadwell goes down to 3-5W), I would not be too surprised if Intel went through another cycle of consolidating mobile and desktop architectures, much like what they did with laptop and desktop with Core 2 after Netburst's brute force approach turned into a dead-end.

If AMD plays it smart, they should have designed Zen with the ability to scale power down in mind to avoid having to design a whole new core specifically for mobile. At the rate ARM is picking up performance optimization tricks that Intel had to leave behind when they started the Atom product line to meet the power budget, significantly cut down x86 cores won't be able to compete against ARM a few years from now.

I thought that was the plan? To slowly rebuild their low power chips back to desktop level while having the power sipping properties of Atom.

I mean the Haswell mobile top end quad core CPU was almost as powerful as the i7 4770 but could also clock down to ultra low power levels.

Of course this is also what AMD should do. It would benefit them in the long run.

It's what they should have done the second they purchased ATI. They were uniquely positioned to dominate the mobile market (first with laptops, back when they were things), but took far too long to get a decent APU together, and at the end of the day, Intel's superior low power performance won the day. Now Intel basically caught up on GPU performance, and ahead in both CPU performance and power draw.

Zen looks to be a big chip, and AMD has said on multiple occasions they aren't focusing on mobile anymore. So I don't think AMD is going to invest significant resources in that market segment anymore.