Back to Article

  • nevertell - Monday, May 05, 2014 - link
    So they called it first. It's around 5-8 minutes. Well, maybe not precisely, but this could be the next step.
    I think that if we go so far as to have x86 and arm cores on the same SoC to do roughly the same thing, we might as well just abstract the microarchitecture even further and just develop a custom ISA and a clock synced interpreter to execute assembly from different architectures on one piece of hardware.
  • name99 - Monday, May 05, 2014 - link

    Spoken like a true (pseudo-) engineer.
    There's a certain class of people who spend their lives figuring out ways to virtualize and machines, regardless of the need --- seems to have something to do with the plasticity of the digital world and a theory that if something can be done it should be done. It's the same sort of mindset that gives us things like CORBA or OpenDoc or networked VM.

    I've no doubt that someone, somewhere, will consider it a valuable use of resources to put a high-end ARM chip in an x86 box and then do a (likely half-assed) job of trying to get Android and Windows to run simultaneously and to share a filesystem, clipboard, network, and peripherals. I've substantially more doubt that this will ever be worthwhile.

    Something like this is only worthwhile if you can articulate a convincing use case. Such a use case would be (in principle) something like:
    - OK we'll add a small screen to the outside of a laptop then
    - we'll have Android running on the ARM and it will
    - display on the screen whatever notifications, etc, come in to the laptop

    Sounds cool, right? The question is, how is this more useful than what your phone does today? Who is it easier to implement than the more general solution of having your phone send notifications to a wearable on your wrist? Working hard to solve a problem that only exists in one person's mind ("What about people without phones? What about email accounts that I have on my computer but not my phone?") is precisely the kind of pseudo-engineering I mentioned at the start of this comment.
  • Kevin G - Monday, May 05, 2014 - link

    Agreed. A full hybrid ARM + x86 chip brings all the compromises and few of the benefits of each platform.

    The idea of cramming two architectures into one chip has been done before: IBM's PowerPC 615 project. Its goal was to bridge the x86 and PowerPC markets together roughly 20 years ago. It never made it to market as the software side was too ugly. There was a bit of office politics involved (MS didn't want to support a hybrid then) but numerous bugs in BIOS and drivers. It was far from stable. The market has matured significantly in 20 years but the technical pieces that made the PowerPC 615 a challenge haven't changed.

    The only advantage to a hybrid ARM + x86 design would be in servers for virtualization purposes (where the chip's hybrid nature is not exposed) or as a mainframe-like coprocessor. In the coprocessor scenario, both architectures can address memory but the front end of an application will only deal with one particular architecture (presumably x86). Delegating specific functions or the entire backend to ARM with some abstraction isn't a big deal: it is what modern GPU's do today. Like today, the coprocessor arrangement will have some overhead dependency on the host processor. There are a few niche areas like storage and networking where embedded an ARM coprocessor would be advantageous.
  • fallaha56 - Monday, May 05, 2014 - link

    would we not be more likely to see a BIG.little approach with x86 cores? jaguar+steamroller

    or is the AMD big core so bad, so Pentium 4 that they should just take jaguar and run with it?

    or BIG.little ARM?
  • Kevin G - Monday, May 05, 2014 - link

    The problem for big.LITTLE between just x86 cores is that the dynamic power range of designs has a lot of overlap already. AMD's Jaguar cores can scale to moderate clock speeds given a bit of voltage. It'd only be under absolute load that a Steamroller derived core would awaken to handle the load. The cost-benefit of adding a steamroller module simply wouldn't be worth it for as frequently as it'd be used. The die space would be better off spent as more cache, more GPU functionality and//or more IO on the SoC.

    This doesn't preclude the idea of big.LITTLE with any of their future x86 designs. I personally just don't see it likely given the extra power budget in laptops, desktops and servers that the x86 platform is dominate in.
  • tuxRoller - Monday, May 05, 2014 - link

    The problem is you didn't provide a convincing use case (since you so easily deconstruct it).
    Running torrents would be a great use for ARM. Basically any long running tasks that don't need huge amounts of single thread performance.
  • The_Assimilator - Monday, May 05, 2014 - link

    Because exactly what everyone needs is for Android to be even slower. Reply
  • TheSlamma - Tuesday, May 06, 2014 - link

    Exactly. Quad core 2.4 Ghz 2GB of RAM just to make it run on par and in many cases slower than iOS and Windows phone OS. Never understood the fascination with that OS, it's an embarrassment to the Linux Kernel. Reply
  • sonicmerlin - Wednesday, May 07, 2014 - link

    People vehemently deny seeing any lag or stuttering... Until you specifically point it out to them. Then they say "it doesn't affect me, I don't even notice it." Sigh... Reply
  • Anders CT - Monday, May 05, 2014 - link

    A SoC with both arm and x86 cores would be very neat. You could have an android laptop, runining legacy x86 software under WINE or simmilar compatibility layer. Reply
  • sonofliberty - Monday, May 05, 2014 - link

    just wait until all those lazy programmers update their code to ARM then is time for us to leave the aging crappy x86 to Intel alone Reply
  • MikhailT - Monday, May 05, 2014 - link

    Intel's already fab'ing ARM chips ( and they used to own StrongARM and still owns some XScale families.

    Intel has an arch license for ARM, so they can join the market with a lead over ARM quicker than we would suspect.

    They're not going to sit on their butt with x86. If the market decides to switch over to ARM for everything (unlikely), Intel will start producing ARM chips quickly and will have an advantage as they still have 2-4 years worth of lead over the process technologies than everybody else.
  • testbug00 - Monday, May 05, 2014 - link

    Intel's fab "advantage" is smaller than most people are led to believe... since 65nm they have been doing what TSMC/GloFo/Samsung are doing now. GloFo and TSMC 28nm compete with Intel's "22nm" in density quite well, while losing in other areas... but, those other areas are likely due more to finFETs.

    That leaves Intel with a product they cannot sell for margins close to what their current lineup sells for, while also having to pay the same for fabs (mostly fixed costs) with smaller chips resulting in them utilizing even less of what they have.

    Intel's biz. model does not work with ARM... Qualcomm's top end chips sell for less than Intel's lowest end chips (well, with Contra revenue that might not be true) (iirc).
  • Krysto - Monday, May 12, 2014 - link

    Intel is not going to join ARM. They failed before and they would fail now. Why do you think they sold Xscale? Because it was going so well for them?

    If they join ARM that would send a strong indicator to the market that they're "dumping" x86 (whether they admit to it later or not) - and then it would become a self-fulfilling prophecy and everyone would dump x86. Intel does not want that because their big profits are in high-end x86 chips (at least for now).

    The only way they are even convincing some clueless OEMs to adopt their Atom chips is because they are HEAVILY subsidizing them to match ARM chip prices, while they are barely competitive in CPU performance, and a full generation behind in GPU performance, even with a half-node advantage + Trigate. They're losing $1 billion every quarter with Atom right now.
  • vailr - Monday, May 05, 2014 - link

    Maybe the idea would be: run Android on the cell phone, and also have the ability to dual-boot + run full Windows (or other x86 operating system: Linux, OSX). Include a built-in mini-sized wall projector on the side of the phone, and you'd have a laptop replacement + decent adjustable sized display for portable Windows usability. Sub par "screen quality", but still usable. Carry a wallet-sized foldable full sized bluetooth keyboard + mini mouse. Reply
  • Thohean - Tuesday, May 13, 2014 - link

    Fujitsu LOOX F-07C already does this with the atom. It dual boots Window 7 and Symbian.
  • TristanSDX - Monday, May 05, 2014 - link

    Thanks to Intel, x86 wont be decreasing, but increasing. AMD speculations refres only to AMD x86 crap. If they cannot make competitive x86 with decades of experience, then they won't make competitive ARM, and their magic formula 'if we don't sell x86 then we sell ARM' will not work. There are other succesfull ARM competitors, more dangerous for AMD on ARM battlefield than Intel on x86. Reply
  • purerice - Monday, May 05, 2014 - link

    Good point. Combine that with a stretching of limited resources and they could have two losing markets. Besides, this seems more like a feature that they are trying to turn into a product. They are way behind on the CPU front and even with HSA, or perhaps all the more because of it, they need to work to catch up in their core business instead of trying to outfeature competitors with something that is out of their expertise. Reply
  • tcube - Tuesday, May 06, 2014 - link

    Nothing furter from the reality. You forget amd does NOT have the printing the chips problem. And ARM already provides the schematics for the cores with the licence. So all amd needs to do is optimise and slap in their graphics core and send it to the "printer".

    Intel however has to deal with the entire process not being a multiple design business model like the rest of the fab world it would be a huge organisational change... amd doesn't know about such problems... one fabber can't build it no problem go to another. This is overly simplified but it still holds true.

    As to expertise... you really compare arm to amd?? You really think that amd is not capable of building a much better arm then arm itself when designers with less experience like qualcomm and apple did it?? Are you on the moon?? Amd has proabably the biggest chip design pattent portfolio in the world probably toppes only by ibm's research labs. To put it into perspective intel LEASES all the gpu patents from nvidia for a LOT of money. Amd doesn't need to. Amd has also loads of memory patents and many many others. So... what was your concern again?? That they can't build a 10x simpler chip then any modern x86 chip... you are really ... really funny.

    Both intel and amd can build arm chips while half asleep drinking coffeeand shagging at the same time... the problem for intel is NOT design after given specs... but addapting the business model and theyr fab processes to support large variations in designs. If you will it's the transition from enormously long conveyourbelt assembly lines to small robotic assembly stations... they are simply not eqipped to do it the way tsmc, glofo and now samsung are doing it.

    As to your claim that they are behind...

    Beema/mullins chips perform similarly x86 wise to intels offerings (in regard to TDP) and at the same time being almost a factor better graphics... still standing in the same power envelope... being on 28nm vs intels 22nm... so... if 22nm process advantage was nullified by amd with this chip... what do you think they'll do on samsungs 14nm process that glofo will use and amd will profit from?? Talk about expertise... what a clown...
  • testbug00 - Monday, May 05, 2014 - link

    AMD has more experience fully designing CPUs than all companies in the ARM space to my knowledge.

    Well, perhaps less than ARM holdings. It really depends on if AMD takes a NVidia "just throw a GPU onto an ARM core" or the Qualcomm route "fully custom, aimed at low power" and, as we can clearly see, they choose the Qualcomm route.

    I don't expect AMD to be the best ARM manufacturer, but, they should be able to compete with every large company in the space, except perhaps Qualcomm.
  • rocketbuddha - Tuesday, May 06, 2014 - link

    There is no way that AMD would face the "ARM"ageddon of facing their ARM competition with purely ARM designed Cortex cores. It has to have its own custom architecture to differentiate from the numerous Cortex ARM cores that will be in the market from Allwinner to POP all the way up to QCOM.
    So AMD had to move to custom ARM ISA like QCOM/Apple. Else it cannot afford to compete aginst competition that would have the same technological or in some case better tech nodes than it. It can have Cortex cores as a back-up in case their custom design takes longer to materialize. See how (QCOM is using Cortex A5/A7 & in the future A53/A57 cores).

    If you look basically in the x86/x64 world AMD is already developing its own CPU cores based on the x86/x64 ISA. So it has been doing that since the 80s and now it is just adding ARM to the repertoire.
  • gruffi - Tuesday, May 06, 2014 - link

    Or Intel is losing in both markets. They are already losing in the ARM dominated markets. There is actually no dangerous ARM competitor for AMD in the targeted markets yet. Or can you show me a successful company for ARM server chips or ARM clients beyond tablets and smartphones? Reply
  • KenLuskin - Monday, May 05, 2014 - link

    In 2015 MSFT will unveil their new OS code named "Threshold".

    A threshold is the end of something, and the beginning of something else.

    Why would MSFT continue to only make Windows for X86 ONLY, when there will be a number of ARM 64 bit chips being released in 2015?

    One of the main reasons that PCs are so expensive is because of Intel.

    Intel has caused people to believe they need a more expensive chip than necessary to do what 90% of them require.

    The other 10% do need a more powerful chip, but most of the additional power is graphics based.

    AMD is working closely with MSFT!

    Because of the Xbox1 custom APU from AMD, MSFT probably has a closer relationship with AMD than with Intel at this time.

    In fact, the main advantage of a PC over Tablet is the substantial increase in GRAPHICS power, and not in the speed of the CPU.

    Therefore, I expect MSFT to tweek their next OS to make better usage of AMD's superior graphics, to further DIFFERENTIATE their products from iOS and Android products.
  • wintermute000 - Monday, May 05, 2014 - link

    "In fact, the main advantage of a PC over Tablet is the substantial increase in GRAPHICS power, and not in the speed of the CPU."

  • MikhailT - Monday, May 05, 2014 - link

    Threshold is just a codename for the Windows 9 series of updates for Windows Phone, Windows RT, and Windows OS.

    It's special because it's going to integrate them all further. We already see the signs of it with the recent universal Windows app stores and universal Windows app binaries. Threshold will go further than that.

    Other than that, Windows will live on x86 for a long time due to backward compatibility. It's not going to disappear any time soon.

    ARM chips is not at that level where it can emulate x86 quickly enough to not see the difference.
  • Krysto - Monday, May 12, 2014 - link

    I bet Threshold will be another idea from Microsoft that's as stupid as Metro was. Probably something like a subscription-based cloud OS that will alienate 90 percent of their consumer market (I imagine the enterprise market would be fine with it, since they're already kind of paying a "subscription" for Windows/Office support anyway - but it would mean the death of Windows in the consumer market). Reply
  • Penti - Monday, May 05, 2014 - link

    So now they (officially) enter spaces they left years ago. They left the mobile gpu business when they sold of the Xilleon and Imageon tech to Broadcom and Qualcomm (Adreno) and they have to compete against Intel, Qualcomm, Broadcom, Imagination, Nvidia and ARM for driver quality, performance and so on now. That will be hard for them. Embedded makes sense, but getting into Android and have to compete against competent gpu-solutions including some of their former technologies of their own with teams that has written drivers for that environment for years will be though. It's quite crowded compared to the PC side with three vendors. Not sure it makes sense to enter GCN into that space with Android. Reply
  • Mathos - Monday, May 05, 2014 - link

    Or people could use their brains, and actually read the contents of the article. Realizing they aren't putting ARM and x86-64 on one chip. They're making chips and chipset/mainboard designs that are architecture independent, but fully socket compatible.. Meaning if someone needs a x86 solution they can use an AMD x86 chip. But if they need a low power solution, they can use an AMD ARM based chip on the same base system by simply changing out the CPU. Which would effectively allow fully functional ARM based desktops, with full APU capability, etc.

    Unfortunately, reading other comments on here I can tell that we have a lot of armchair experts at nothing. Which fail to actually read and comprehend what was written in the article.

    People also don't seem to realize that AMD's original chip designs were all RISC based, with x86 decode through firmware.
  • extide - Tuesday, May 06, 2014 - link

    Heh, heh, yeah I don't think people were actually thinking they were making a chip with both ARM and x86 on it (wait they already make one of those*), but they were speculating the possibility.

    *Cortex A5 security core in Beema/Mullins
  • Silma - Tuesday, May 06, 2014 - link

    AMD hasn't produced a single interesting processor in years.

    While the ARM road would in fact be an interesting one to explore by Intel, because it also leads the production side of things, I fail to see what added value AMD can bring, since it doesn't own a single fab.
  • Silma - Tuesday, May 06, 2014 - link

    The big guys like Apple and Samsung do their ARM themselves, the small guys buy from Qualcomm and Nvidia or can even work directly with TSMC. Reply
  • gruffi - Tuesday, May 06, 2014 - link

    "AMD hasn't produced a single interesting processor in years."

    That's funny. I always think the same about Intel. I actually wanted to buy Haswell. But even this one is uninteresting for me.
  • Gizmosis350k - Wednesday, May 07, 2014 - link

    What does this mean for enthusiast desktop processors? Will we have to wait until 2017 before the FX refresh is out? Reply
  • Cpuhog - Monday, May 12, 2014 - link

    Can someone explain what the whole point is? They are not doing a dual architecture on the same chip. They are coming out with a low end ARM processor which will be pin-compatible with a low end x86 processor. At the high end they already have x86 offerings and now they will also have an ARM K12 based design. It does not appear that they have a good handle on how ARM is going to help them. They know they can't compete with x86 so how will they compete with ARM? Reply

Log in

Don't have an account? Sign up now