Page 2 of 3 FirstFirst 123 LastLast
Results 11 to 20 of 21

Thread: ARM Cortex-A15 Quad-Core Linux ODROID-XU Tests

  1. #11
    Join Date
    Jul 2008
    Posts
    925

    Default

    I am looking forward to the quadcore Kabini cpus from amd with 15W or 25W TDP.

    A comparsion between this 2 would be great. This thing uses a active cooler so the tdp will be also higher than 5W right? So a comparsion in TDP, Price and Speed would be great. But even if there would amd loose slightly the great opensource drivers would be the desiding factor in favor of amd I think.

    And it would be great if amd could make this boards in this century buyable. Waiting forever... maybe they produce to much garbage ps4 console socs.

    Where I see potential for such boards but for that it has to much tdp it seems (active cooler) would be tablets... because there you become from amd it seems only 800 dollar 640x480 resolution Windows 8 garbage tablets.

  2. #12
    Join Date
    Feb 2008
    Location
    Linuxland
    Posts
    5,338

    Default

    Both Mediatek and Rockchip are making equal octacores (not big.LITTLE), composed of eight A7 cores. IIRC Mediatek's one even had a Mali 400.

    These would be interesting to compare when they come out. Not as performant as A15 per core, but much lower power too.

  3. #13

    Default

    Quote Originally Posted by gururise View Post
    Would be great to see this processor compared to the new Snapdragon S4 800
    This not the Snapdragon 800's competitor, though. The 1.8 Ghz Exynos 5420 is, with a more powerful GPU. I'm not sure if the Note 3 international has it and if it's been reviewed yet.

  4. #14
    Join Date
    Aug 2013
    Posts
    91

    Default

    I'd like to see these compared to Intel's Bay Trail Atoms in terms of performance and driver stability and features. Been wanting UMPCs to make a comeback. I used to have an OQO slider and a Viliv N5 but they weren't Linux friendly.

    edit:
    Oh, this needs a fan. What's the power ratings for this board?
    Last edited by guido12; 10-22-2013 at 11:36 AM.

  5. #15

    Default

    So far Bay Trail is barely competing with last year's dual dual core Cortex A15 and ARM GPU's:

    http://images.anandtech.com/graphs/graph7428/59065.png

    http://images.anandtech.com/graphs/graph7428/59075.png

    I've stopped buying Intel's BS a long time ago. Wait until the reviews are out instead of believing Intel a year before the release about how their chips about how their new chips will totally destroy ARM chips. When they come out they tend to only be competitive with chips from a year before, in both CPU, and especially in GPU performance.

  6. #16
    Join Date
    Aug 2013
    Posts
    91

    Default

    Quote Originally Posted by Krysto View Post
    So far Bay Trail is barely competing with last year's dual dual core Cortex A15 and ARM GPU's:

    http://images.anandtech.com/graphs/graph7428/59065.png

    http://images.anandtech.com/graphs/graph7428/59075.png

    I've stopped buying Intel's BS a long time ago. Wait until the reviews are out instead of believing Intel a year before the release about how their chips about how their new chips will totally destroy ARM chips. When they come out they tend to only be competitive with chips from a year before, in both CPU, and especially in GPU performance.
    How's Bay Trail graphics support under Linux? Any real world reviews on it? I know it's based on the HD4000 but actual tests would be great. Thanks!

    Edit:
    At least their boasting against previous generation Atoms weren't too much of a lie:
    http://images.anandtech.com/graphs/graph7428/59050.png

    Of course that shows how behind the Atom was against ARM in the first place.
    Last edited by guido12; 10-22-2013 at 12:52 PM.

  7. #17
    Join Date
    Sep 2012
    Posts
    369

    Default

    Quote Originally Posted by Krysto View Post
    I've stopped buying Intel's BS a long time ago. Wait until the reviews are out instead of believing Intel a year before the release about how their chips about how their new chips will totally destroy ARM chips. When they come out they tend to only be competitive with chips from a year before, in both CPU, and especially in GPU performance.
    ARM CPUs are usually useless with Linux because of bad GPU drivers. ARM CPU would be useful only if Intel would produce ARM CPU like A57 or A53 with Intel GPU.
    Last edited by JS987; 10-22-2013 at 01:01 PM.

  8. #18
    Join Date
    Sep 2013
    Posts
    30

    Default

    Quote Originally Posted by JS987 View Post
    ARM CPUs are usually useless with Linux because of bad GPU drivers. ARM CPU would be useful only if Intel would produce ARM CPU like A57 or A53 with Intel GPU.
    I have some hope with Nvida Tegra SoCs. Because ulike the desktop GPU family, the ARM family drivers are free/open. So I hope that NVIDIA Tegra will become the de facto standart for Linux Arm.

  9. #19
    Join Date
    Sep 2012
    Posts
    369

    Default

    Quote Originally Posted by Krysto View Post
    So far Bay Trail is barely competing with last year's dual dual core Cortex A15 and ARM GPU's:
    http://images.anandtech.com/graphs/graph7428/59065.png
    http://images.anandtech.com/graphs/graph7428/59075.png
    ASUS Transformer Book T100 has slower Z3740 (1.86 GHz) instead of Z3770 (2.39 GHz):
    http://anandtech.com/show/7314/intel...z3770-tested/2
    http://anandtech.com/show/7314/intel...z3770-tested/3
    Z37x0 are quad cores, but most of these benchmarks probably don't scale with more cores

  10. #20
    Join Date
    Feb 2013
    Posts
    275

    Default

    Quote Originally Posted by schmidtbag View Post
    I still have a hard time understanding the purpose of the big.LITTLE architectures, or other similar platforms such as OMAP5.


    Anyway, I own an odroid-U2 and it's a pretty fantastic platform. I wish there was openGL acceleration (rather than OGL ES) but it being a quad core makes up for the performance hit it takes for having the CPU render everything.
    Re: The purpose of big.LITTLE: Power consumption is a concern for designers of ARM systems. If I recall correctly, big.LITTLE systems address this by using the slower low-power cores when the load is low and then switching to the high-performance cores when the load requires it. I believe this is also the reason for why the cores have to be paired - processes need to be transferred from the low-power cores to the high-performance cores seamlessly. I think that's also the only time when all of the cores are in operation; otherwise, either the A15 cores or the A7 cores are running, not both sets.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •