Report: New AMD CEO readies strategy shift

In case you haven’t heard, there’s a new man at the helm of AMD: freshly minted CEO Rory Read, who joined the chipmaker late last month from Lenovo. According to Fudzilla, which quotes an “insider hint,” Read is working on a “twist in AMD’s long-term strategy.” The site says that twist may very well involve mobile devices:

We heard that Rory Read was famous for turning things around and that he helped Lenovo with some key transitions. Since AMD believes that smartphones and tablets are getting more important, and that they need even better mobile chips, its logical to assume that this new strategy can get AMD closer to these categories.

The lack of a solid game plan to tackle the ultra-mobile and handheld markets is said to have contributed to the dismissal of AMD’s previous CEO, Dirk Meyer. Unsurprisingly, during Read’s introductory conference call on August 25, many asked the new CEO about AMD’s plans in those areas. Read and other AMD execs shied away from sweeping statements, but they did express a desire to go after ARM in “lower power bands.” Perhaps we’ll hear more about that soon, then.

Comments closed
    • stmok
    • 8 years ago

    *cough cough*
    => [url<]http://img818.imageshack.us/img818/3198/lowendfuture1.jpg[/url<] [b<][u<]2011 = "Desna"[/u<][/b<] => Came into existence a few months after they got rid of Dirk Meyer. => Ultra low power variant of the C-50 APU ("Ontario"). => Marketed as Z-series. => AMD Z01 used in tablets. (See MSI WindPad 110W with Windows 7.) => Dual core 1.0 GHz with Radeon HD 6250. TDP: ~5.9W => 40nm from TSMC. [b<][u<]2012 = "Hondo"[/u<][/b<] => A re-done variant of "Ontario" with a low power (and smaller) chipset complement. => Aimed for Windows 8 release. => 40nm from TSMC. [b<][u<]2013 = "Samara"[/u<][/b<] => ??? ...They're scaling down Bobcat-based APUs to ultra-low power and tablet markets.

      • shank15217
      • 8 years ago

      You know, Dirk Meyers never said that he didn’t want the low power market, he just wanted to capture the high end as well, as he believed that AMD could go toe to toe with Intel, too bad the board didn’t think so cause they’re near-sighted fools.

        • A_Pickle
        • 8 years ago

        Yeah… I’d go so far as to say that AMD is poised to [i<]take[/i<] the low power segment from Intel much sooner than the high end market. Intel's stuff at the high end is impressive, but their stuff at the low end is absolute garbage. Intel's low end gives you general purpose processing and battery life. AMD's low end gives you general purpose processing, battery life, [i<]and[/i<] graphics performance. That last bit matters to consumers, who are the prime investors in the low end CPU market.

    • jimbo75
    • 8 years ago
      • A_Pickle
      • 8 years ago

      They’ve done pretty well with the E-350 and the C-50. Those aren’t quite the power envelopes needed for an ARM-competitive smartphone or tablet, but… give ’em time. They’ve got smart engineers working on it, and I have no doubt in my mind that they’ll figure it out. If Microsoft can reduce the Windows boot time by something like 70% in one release cycle, AMD and Intel can figure out how to incrementally reduce power consumption until they’re competitive with ARM — which appears [i<]not[/i<] to have much more downward room to go towards.

    • jcw122
    • 8 years ago

    Screw mobility…desktops forever!

      • Dashak
      • 8 years ago

      Why not both?

    • Vasilyfav
    • 8 years ago

    Good job, capt. Rory Obvious. Is your common sense tingling?

    • ronch
    • 8 years ago

    I hope all this mobile stuff AMD wants to go into will not detract them from their core business, which is to continue competing with Intel in terms of x86 performance in all market segments. The way things are going right now it’s like AMD is trying to catch Intel’s shadow, going where Intel goes, and when that segment turns out to be a fad they already spent a fortune developing products for it. Just look at Bobcat. They wanted an Atom competitor, and just when the E-350 came out the Atom/Netbook market was starting to fizz out. It’s just fortunate that E-350 is succeeding because it could fit into other market niches, such as HTPCs, but that wasn’t their original intention. And although you could argue that Bobcat will fit inside smaller form factors such as tablets, it simply isn’t being taken seriously by everyone (to be fair, even Atom couldn’t make it big into tablets) because we have better alternatives like ARM.

    Going where the industry is going isn’t really all that bad, but I think AMD needs another AMD64 sort of innovation and industry vision if it wants to be taken seriously again. And oh yeah, BD has to prove AMD’s prowess again. Now I know a zillion folks will say that AMD’s products are just fine for what most of us do, and they’re right, but Intel moves forward quickly, blazing their own trail. If AMD doesn’t catch up soon they’ll end up like VIA, and later, Cyrix.

    • adisor19
    • 8 years ago

    This is beyond hilarious ! First Intel bets the farm on x86 and sells their strongARM chips, design and IP. AMD does the same because well, they had ZERO vision. Intel gets caught with their pants down when Apple dissess Atom and goes with ARM for their iPhone. AMD, well… no comments.

    The 2 x86 giants that have bet everything on the Windows-x86 mantra are now in a very very tough spot. It might not look like it for Intel yet due to them being one step ahead of the industry in terms of chip fabbing, but eventually it will catch up to them.

    Adi

      • phileasfogg
      • 8 years ago

      I gave Adi a +1 but I think he may be overstating the facts a bit. “very very tough spot”? Really? I don’t think so. But I’ll revisit this in about a year and see if I want to change my mind. Don’t minimize the importance of being an IDM (independent device manufacturer) and owning all of your fabs – those are very key weapons in the fight. OTOH, the company that bought Intel’s StrongARM portfolio (Marvell) isn’t making a whole lot of money off that group, AFAIK. Sure, they did launch a couple neat chips, but I don’t think they were/are game-changers (yet).

      Those who predict that x86 CPUs will eventually wither away in the face of the ARM onslaught do so at their own peril. This is the ultimate ‘survivor’ chip and no one, not even Ottellini and co. can predict when it will ultimately die.

      • WillBach
      • 8 years ago

      [quote<]The 2 x86 giants that have bet everything on the Windows-x86 mantra are now in a very very tough spot.[/quote<] I think that Intel's bet was more on Windows-Mac-Linux-x86-IA64 and AMD's bet was more Windows-Linux-x86. Intel and AMD have both won contests against RISC architectures before, and modern ARM processors that support Thumb2 are not even really RISC designs. Look [url=https://techreport.com/articles.x/6866/3<]here[/url<] for Scott's earlier report on Dirk Meyer talking about what x86 "overhead" really meant. You can also look [url=https://techreport.com/discussions.x/17425<]here[/url<] for another report from Scott on ARM/AMD/Intel Atom. The real issue is that Intel and AMD are both getting used to creating low-cost SoCs designs. Making an SoC requires a whole different transistor "toolkit" operating in different speed and voltage ranges, and requires on-board capacitors, resistors, and inductors. AMD figured that out with Bobcat, but they used TSMC for the fabrication, probably because TSMC had SoC experience while AMD (later Global Foundries) did not. It may be true that AMD and Intel are in tough spots - except that Intel is still making titanic ammounts of money and AMD has started shipping Bulldozer and entered a partner ship to provide GPUs to other chip design companies. That said, we'll have to see what comes out of Intel at 32 and 22 nm and what comes out of AMD at 28 and 20 nm before we know for sure how this will turn out. Edit: typo when I wrote18 nm and meant 20 nm. Extra: link to AMD licensing GPUs to ARM vendors: [url=http://semiaccurate.com/2011/06/22/amd-and-arm-join-forces-at-last/<]SemiAccurate - AMD and ARM join forces at last[/url<]

      • A_Pickle
      • 8 years ago

      No, it won’t. You might think that everyone and their dog wants a bloody iPad, but the truth of the matter is that [i<]stupid devices[/i<] need online services. And online services are made possible through servers. x86, big iron, servers -- not rackmounted iPads all Beowulfed together. AMD and Intel are fine, and their chip expertise will get them back into ARM SoC's reasonably easily. I think you're grossly overstating the performance/watt scalability of x86, as well -- I don't think it's remotely off the table that x86 could easily match ARM in terms of performance/watt. EDIT: Especially considering Intel's past experience with ARM SoC's, and AMD's current [i<]success[/i<] with x86 SoC's. Apple's got [i<]nothing[/i<] on the E-350 and the C-50, which are both pretty impressive chips -- and their follow-ups are going to be even smaller, faster, and more power efficient.

      • mnecaise
      • 8 years ago

      I think that’s a bit of an overstatement. The x86 holds the performance high ground (if you exclude the server specific corner cases where Sparc or Power reign). x86 will never match ARM in performance per watt because of complexity; although AMD and Intel have made significant progress in that reguard. They’re playing to their strenghs rather than dividing their focus on multiple architectures.

      There will continue to be demand for processors of the x86 architecture with performance characteristics that ARM can’t meet, for the forseable future.

      • Krogoth
      • 8 years ago

      You are smoking some fine stuff.

      x86 isn’t going anymore, nor is Intel/AMD.

      ARM didn’t catch anybody’s pants down. They simply made them aware of smartphone and tablet markets. It is ARM who should be worrying. Intel has the fiscal resources, fab capacity and engineering talent to make a viable and compentive smartphone/tablet platform.

      • obarthelemy
      • 8 years ago

      i’d love to be in the same tough spot as intel.

    • derFunkenstein
    • 8 years ago

    If they get going now, they can probably beat Intel in this call to ARMs.

      • NeelyCam
      • 8 years ago

      Why would they go ARM to be merely one of bazillion ARM companies, when they have the x86 license almost nobody else has?

        • sweatshopking
        • 8 years ago

        cause it seems to suck at low power.

          • esterhasz
          • 8 years ago

          I think that NeelyCam is spot on. It may make technological sense (your point) to go ARM but, really, they’re struggling against Intel, why would they choose to compete in a market with a much larger company than Intel (Samsung) and a bunch of other guys that are not exactly incompetent competing around a common chip design. They have problems getting enough chips from TSMC as is, why enter a market with higher volumes and lower margins?

          With Windows 8 going tablet and Android going x86, I would say that the prospect of going ARM looks a lot less attractive than a week ago.

          If they can get Brazos to a 6-8W TDP for CPU+GPU+NB, I’d say they can probably get into the tablet game OK.

            • flip-mode
            • 8 years ago

            Not saying that I disagree with you, but two things come to mind:

            1. If AMD can compete against Intel, they can probably compete against anyone. There’s a huge asterisk next to that because Intel could easily price-cut AMD out of the market but they choose not to. So maybe this point has no validity at all.

            2. AMD seems to be intimately familiar with working with low margins. This point is more of a fun-poking at AMD.

            As I said, I’m not claiming to disagree with you.

            • BobbinThreadbare
            • 8 years ago

            [quote<]There's a huge asterisk next to that because Intel could easily price-cut AMD out of the market but they choose not to.[/quote<] No they can't, not legally anyways.

      • ronch
      • 8 years ago

      I think AMD needs to focus on their core business. Spreading themselves thinly now (on ARM) when their x86 lineup is two generations behind Intel just doesn’t make sense when their x86 business has been their proven bread and butter for decades.

        • just brew it!
        • 8 years ago

        Here’s the thing though… with mobile devices (and browser-based apps in general) increasingly becoming the centerpiece of computing, x86 simply isn’t as important as it once was. You can run a web browser on practically anything; the importance of x86 desktops (and MS Windows) is going to decline going forward.

        Servers will still matter; *something* has to power the data centers behind all those browser-based apps.

        Bulldozer appears to be a reasonably competent server CPU (based on the limited information available so far). So now AMD really needs to figure out what to do about the client side.

        [url<]http://xkcd.com/934/[/url<]

      • paulWTAMU
      • 8 years ago

      some people just can’t appreciate a good pun. +1 from me though.

    • Tristan
    • 8 years ago

    Strategy shift for AMD:
    – Drop Zambezi. No one really needs eight slow, expensive, power hungry cores
    – Increase production of Llano and Brazos
    – Better prepare Trinity to avoid delays and lost of the face (the face is already lost, they cannot lost it futher)
    – Heavy investment in notebook, netbook, tablet and phone ready chips. The future of PC is mobility
    – Improve software, marketing and partner support
    – less slides and more products

      • Elsoze
      • 8 years ago

      [quote<]- Drop Zambezi. No one really needs eight slow, expensive, power hungry cores[/quote<] Slow? Expensive? Compared to what? Chips Intel doesn't have out yet?. The price/performance value appears to be there. But that's just an assumption, just like yours, since the parts aren't out yet. [quote<]- less slides and more products[/quote<] Agreed.

        • bittermann
        • 8 years ago

        Denial can lead to very bad things…

          • OneArmedScissor
          • 8 years ago

          “Drop Zambezi. No one really needs eight slow, expensive, power hungry cores”

          You imply that selling them as desktop chips eats into their production. It doesn’t. They’re like the Pentium/Celeron equivalent of the Opteron chips that didn’t come out “right” and they can either throw them away or sell them to a different market.

          Oops, meant as a reply to Tristan.

            • chuckula
            • 8 years ago

            [quote<]You imply that selling them as desktop chips eats into their production. It doesn't. They're like the Pentium/Celeron equivalent of the Opteron chips that didn't come out "right" and they can either throw them away or sell them to a different market.[/quote<] That's not accurate. Each Zambezi die is basically identical to one-half of an Interlagos chip (that includes two dice) or is exactly the same as a single-die Valencia server chip, just repurposed for the desktop. What's more, each and every 4-core, 6-core, and 8-core Zambezi chip uses the exact same 315 mm^2 die. Lots of people dismissed the die size as immaterial because the 6 core Thubans were 345 mm^2, but they are missing 2 incredibly crucial points: 1. Thuban didn't come out until well over a year after AMD went to 45nm... and that was an *AMD* 45nm process, not a 3rd party Glofo process. 2. Even after Thuban came out, AMD could selectively produce 4 core and 2 core parts with smaller die sizes (~250 mm^2) and produce the large Thuban chips *only as needed to fill that niche*. AMD has no such luxury with Bulldozer... that 4 core desktop chip that's going up against a $120 Intel chip is just as expensive to produce as the 8 core parts that will still sell for less than a 2600K.

      • ronch
      • 8 years ago

      [quote<]Drop Zambezi. No one really needs eight slow, expensive, power hungry cores[/quote<] Nobody really knows for sure how Zambezi will work out. They also need to catch up to Intel and this is their best bet. [quote<]Increase production of Llano and Brazos[/quote<] Why? Is there market data to support this? Does everyone really want a Llano or a Bobcat variant? [quote<]Better prepare Trinity to avoid delays and lost of the face (the face is already lost, they cannot lost it futher)[/quote<] No need for you to say that. I bet that no chip maker ever wants its products delayed or turn out performing worse than they expected it to, and AMD is no different. They're surely doing everything they could right this minute. [quote<]Heavy investment in notebook, netbook, tablet and phone ready chips. The future of PC is mobility[/quote<] This is, as this article reports/implies, what Rory Read is planning to do. [quote<]Improve software, marketing and partner support[/quote<] Software? You mean drivers? Don't they roll out new drivers every month or so? They are improving their stuff all the time. Marketing requires more money. Partner support all depends on their products, and AMD is already as good as they could be with their OEM relationships. [quote<]less slides and more products[/quote<] They have a licensed copy of Microsoft Office. Why waste it?

    • Silus
    • 8 years ago

    Took their time to realize something as obvious.

    • OneArmedScissor
    • 8 years ago

    Now that Windows will run on ARM, they could jump in pretty quickly by just slapping a run of the mill Cortex A15 onto something like the upcoming Bobcat SoC, and then work their way into phones from there. Of course, that would also put them right back in the position of doing exactly what Nvidia is doing, in the exact same target market lol…

      • sweatshopking
      • 8 years ago

      it would be a tricky proposal. I’m not sure it would really be in their benefit in the long run. amd has the gpu’s and they’ve got the low power x86 market firmly in their grasp, despite what neely says, and god knows i love her, but atom might be ok for cells in a generation or two, but for pc’s i’d never recommend anything but brazos in that range.

        • OneArmedScissor
        • 8 years ago

        The problem is that today’s low power is tomorrow’s high power. We’re only a year away from laptops moving to the equivalent of ultra low voltage SoCs as the standard, and they’ll be the majority of the PC market at that point. An ARM SoC would be in AMD’s interest if they want to follow the shift down and go from desktops and laptops to laptops and phones.

        Everything is already in place. They have organized their entire business to increasingly revolve around upscaling GPU power, while retaining similar controller functions and just swapping out whatever CPU core is best suited for the job. Just add a pre-baked ARM core into that strategy and they have phones covered, as well.

          • sweatshopking
          • 8 years ago

          that’s a good point. they have already gone pretty modular with bobcat and bulldozer, it shouldn’t be THAT tricky to just swap one core for another, if the risk is low, why not?

      • phileasfogg
      • 8 years ago

      I don’t see how and why Scissor called Cortex A15 “run of the mill”. Since when has a CPU core become a commodity before it has even begun to ship?

    • maxxcool
    • 8 years ago

    Considering one of the major points in the previous beheading was a complete lack of “mobile” strategy, this is not so surprising.

Pin It on Pinterest

Share This