Slashdot Mirror

← Back to Stories (view on slashdot.org)

Intel Announces Clover Trail+ Atom Platform For Smartphones and Tablets

Posted by ryuzaki0 on from the check-it-out dept.

MojoKid writes "Today, Intel announced the follow-on to their Medfield Atom platform for smartphones, code-named Clover Trail+. Clover Trail is powering a few Windows 8 Pro tablets currently. However, Clover Trail+, Intel's new performance and feature-optimized version of Clover Trail for smartphones and tablets, has a long row to hoe versus incumbents like Qualcomm, Samsung and NVIDIA, at least in the highly competitive handset arena. What's interesting this time around is that Clover Trail+ seems to really have the chops (at least on paper) to keep pace with the performance of current, best-of-class ARM-based architectures that have been so dominant in smartphones. Clover Trail+ is another 32nm design and Intel has beefed up almost every major functional block on the platform. From its now dual-core, 4-thread capable Atom CPU, to its new PowerVR SGX 544MP2 graphics engine, 2GB of LPDDR2 1066 DRAM, up to 256GB of NAND storage, a higher resolution 16MP camera and Intel's XMM 6360 HSPA+ 42Mbps modem, with LTE support from their XMM 7160 radio moving forward; Intel's Clover Trail+ smartphone reference design brings a lot more to the table than Medfield ever did."

7 of 56 comments (clear)

  1. Finally by binarylarry · · Score: 4, Funny

    I can use my phone in the winter *and* keep my hands warm.

    Oh Intel, what would I do without you!

    --
    Mod me down, my New Earth Global Warmingist friends!
  2. Mythbusting time! by CajunArson · · Score: 3, Insightful

    Myth 1: Intel can't make x86 power/performance competitive with ARM: Being busted as we speak.

    Myth 2: ARM can't scale up performance: Beginning to be busted with the A15, more to come with the 64-bit chips.

    Myth 3: ARM can just press a button and get Intel level performance without using any extra power: Busted wide open by ARM themselves with the whole "littleBIG/BIG/little/etc" approach and by the conspicuous lack of high-end A15 chips in smartphones (note Tablet != Smartphone, and look at the Cortex-A9 based Tegra4i for the latest example of manufacturers not putting high-clocked A15s in smartphones).

    --
    AntiFA: An abbreviation for Anti First Amendment.
    1. Re:Mythbusting time! by CajunArson · · Score: 5, Insightful

      Fact 2: Compilers exist.
      Fact 3: x86 compilers exist.
      Fact 4: NDK for x86 exists.
      Fact 5: ARM fanboys who say that ARM will win simply be recompiling Linux while claiming that it is physically impossible to run Linux on an x86 architecture are almost as amusing as they are ignorant.

      --
      AntiFA: An abbreviation for Anti First Amendment.
    2. Re:Mythbusting time! by tlhIngan · · Score: 2

      Myth 1: Intel can't make x86 power/performance competitive with ARM: Being busted as we speak.

      I read the summary and the article twice. I have not found anything that said what the power consumption of this SoC is. In fact, the article gushes all about specs EXCEPT power consumption.

      Are they at the 1mW/MHz level of ARMs yet or aren't they? The closest thing I cound is "Hurry Up and Get Idle" thing which doesn't tell me a thing about power.

      As for A15 power guzzler, yes, it's a huge drain, but that doesn't mean you have to use where power is a concern (big.LITTLE is a way to mitigate it, but there are problems, though you can run all cores at once if you wanted). It's also a microarchitecture issue - enough such that 64-bit actually scraps a lot of things that the 32-bit ARM had purely because it doesn't work cleanly to a modern superscalar processor (e.g., conditional execution).

    3. Re:Mythbusting time! by Artraze · · Score: 2

      > conspicuous lack of high-end A15 chips in smartphones

      Sort of? Qualcomm's Snapdragon S4 chips have a custom A15-styled "Krait" core which has been in phones for a while. It's a little different than a stock A15 design, but not massively. It's just a few design tweaks (Added a L0 cache, reduced L1, reduced pipeline) that make a fair bit of sense. Performance is only slightly reduced (~5%).

      With that said, I don't know if that "ARM can just press a button" myth was ever really present. There may have been some discussion that ARM would be able to do so with Intel-quality manufacturing processes, and to a degree that's true: the A9 - A15 (or Qualcomm's A8/9-ish Scorpion to Krait at least) transition is a pretty good example of that being true.

      > Myth 1: Intel can't make x86 power/performance competitive with ARM: Being busted as we speak.

      First, Intel is still nowhere near Context-M (embedded; controls your toaster) levels of power/performance, and that's a target they are highly unlikely to ever hit. Even though it is not as high-profile/large(?) a market the Cortex-A, it is still a very important one, and enough to cement ARM as a very important architecture. At those power levels, ISA matters a lot as there is no cache, branch prediction, register renaming, out of order execution, etc. You are just running the instructions as written, and ARM has solid advantages over x86 there.

      As for the Cortex-A (phones) levels, well, it very much remains to be seen. Intel is competitive but not really surprisingly. Manufacturing has advanced to a point where you can put enough transistors in a chip to use a lot of the aforementioned optimizations. That means the ISA is starting to be abstracted enough that things like x86's limited registers aren't going to hurt and much, and the CISC instructions show some benefit. However, the conspicuous lack of power specification indicates that they probably haven't gotten there quite yet.

      So it's still a battle, but I think that myth is pretty fact-y yet. As manufacturing improves and designed matures, we will probably see them meet in the middle, but Intel/x86 flat out winning? Pretty unlikely. (Keep in mind that it's not just x86 vs ARM, but also SSE vs NEON and MMU vs MMU. ARM is a little less burdened with history as was more designed with low power and concrete use cases in mind.)

      As you allude to, though, 64bit ARM is coming, and it's going to change a lot. X86-64 addressed many of x86's shortcomings, so the battle will begin anew there.

  3. Can't help myself. by mythosaz · · Score: 3, Funny

    It's an ARMs race!

  4. The graphics still sucks by tlambert · · Score: 3, Interesting

    Intel is unwilling to pair a higher end GPU with a lower end CPU, since given that much of the CPU is pent on eye candy and CODECs these days, doing so would cannibalize higher end CPU sales.

    If they could guarantee that this would only every be used in a hone or a tablet without a keyboard dock like the transformer, they'd likely be willing to go for it, but just as the recent Samsung ARM ChromeBook demonstrated, phone/tablet chips can and will be used in laptops, and likely eventually desktops. The thing which has stopped this so far is the need for Intel software compatibility, which the ChromeBook side-steps by not running (non-NaCl'ed) native code, and being mostly a browser.

    If Intel came out with a CPU that was not a compute giant, but had a good GPU which could be used for higher powered math calculations, thus obviating the need for a high powered CPU, then there would quickly be a lot of machines in the laptop space grabbing them up. This wouldn't be terrible for Intel, as long as they charged higher prices for the things based on the GPU power rather than the CPU power --- but doing that would be disastrous for their ability to compete in the tablet/phone market, so they are somewhat pilloried by having one monolithic instruction set across their product line. Ironically, capping the instruction set to make it inappropriate for desktop would throw the CPU out as yet-another-Intel-incompatible-ARM-competitor, so Catch-22.