Announcement

Collapse
No announcement yet.

AMD Athlon 5350 APU On Linux

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #71
    Originally posted by tuke81 View Post
    True, like this(but it's like a build in pico psu):
    http://www.asrock.com/mb/AMD/AM1H-ITX/
    And what are that adapter's limitations? What's it's efficiency? How do I power drives off it? Will I only be able to run a Samsung MZ-MTE1T0BW mPCIe drive? What if I add a PCIe card?

    Comment


    • #72
      Originally posted by Kivada View Post
      And what are that adapter's limitations? What's it's efficiency? How do I power drives off it? Will I only be able to run a Samsung MZ-MTE1T0BW mPCIe drive? What if I add a PCIe card?
      FYI, the board has a SATA power connector on it to power peripherals. This is perfect for the typical use case where everything you need is on the board, except the hard drive and possibly an optical drive. Anything that requires more power will need an internal power supply, but most people won't need this anyway.

      Comment


      • #73
        Those results, at least compilation, are so wrong. Don't know what went wrong during your benchmark, but I just did a test run(s) and the results are a bit different from those mentioned in the article: http://openbenchmarking.org/result/1...KH-ATHLON53567

        Comment


        • #74
          Originally posted by szarpaj View Post
          Those results, at least compilation, are so wrong. Don't know what went wrong during your benchmark, but I just did a test run(s) and the results are a bit different from those mentioned in the article: http://openbenchmarking.org/result/1...KH-ATHLON53567
          Yup something went wrong at that day 3 months ago , your numbers are correct . You got perfectly that 990 vs 330 seconds for kernel compilation benchmark

          it must be 3 times faster then e-350...
          http://www.phoronix.com/forums/showt...578#post410578
          Last edited by dungeon; 07-03-2014, 06:06 PM.

          Comment


          • #75
            New results (much closer to reality)

            Yes,

            I can't blame anyone for "wrong" results. All I can say is that my results with the cheapo MSI motherboard and 8G of good ram paint an entirely different picture. From these results you are pretty much seeing a doubling of performance over E-350, and I might add that the system (only the motherboard and cpu are different from my previous e-350) idles at 3 Watts lower, and peeks pretty much the same 30 Watts.

            No one in their right mind would run only 4G of ram on a multi-purpose machine. If it was a dedicated single-app server sure, but not a desktop.

            http://openbenchmarking.org/result/1...KH-1404098KH91

            the AMD Kabini 5350 is a very good low power cpu!

            Comment


            • #76
              Small amounts of RAM more common than you think

              Originally posted by cbxbiker61 View Post
              Yes,

              I can't blame anyone for "wrong" results. All I can say is that my results with the cheapo MSI motherboard and 8G of good ram paint an entirely different picture. From these results you are pretty much seeing a doubling of performance over E-350, and I might add that the system (only the motherboard and cpu are different from my previous e-350) idles at 3 Watts lower, and peeks pretty much the same 30 Watts.

              No one in their right mind would run only 4G of ram on a multi-purpose machine. If it was a dedicated single-app server sure, but not a desktop.

              http://openbenchmarking.org/result/1...KH-1404098KH91

              the AMD Kabini 5350 is a very good low power cpu!
              There is one obvious reason someone would build a desktop with 4G of RAM: if they are building or rebuilding from existing parts and that's what they've got. I built two machines from older Phenom II x4 setups, had 8G of ram on hand. The one that got the most video editing work got 6G and the other got only two and works fine, though Kdenlive must be run by itself on the latter and projects periodically closed and reopened. In all other work it's ever done you would not notive. Both run copies of my main 64bit OS, if the smaller one got much hard use I would redistribute the RAM back to how it was in 2011 when each board had 4G and both worked fine for everything they were used for.

              Comment


              • #77
                Originally posted by Luke View Post
                There is one obvious reason someone would build a desktop with 4G of RAM: if they are building or rebuilding from existing parts and that's what they've got. I built two machines from older Phenom II x4 setups, had 8G of ram on hand. The one that got the most video editing work got 6G and the other got only two and works fine, though Kdenlive must be run by itself on the latter and projects periodically closed and reopened. In all other work it's ever done you would not notive. Both run copies of my main 64bit OS, if the smaller one got much hard use I would redistribute the RAM back to how it was in 2011 when each board had 4G and both worked fine for everything they were used for.
                Yes, if you're throwing something together for cheap, by all means reuse parts if you can get away with it.

                But really the main point is if you've got $300-$600 invested in a machine, $30-$40 for 4G only make sense. The old adage "penny wise and pound foolish" applies here.

                Comment


                • #78
                  @cbxbiker61

                  Maybe you don't being aware of, but there are few tweaks to raise graphic performance in games of course . Keep in mind that kernel 3.16 is generaly faster for SI hardware due to VM implemented/enabled.

                  So as i see you have 56 fps in Xonotic low for example, i am seeing 70 fps with kernel 3.16.

                  Additionaly with hyperz enabled i have 80 fps there.

                  With governor set as 'performance' 84 fps.

                  New parameters to play with in kernel 3.16 are vm_size and vm_block_size can also raise perf by few fps - so currently with the current stack what i can get is 86 fps .

                  All in all that is 30 fps difference or +35% in graphic performance in Xonotic 1080p low case, raised with the few tweaks .
                  Last edited by dungeon; 07-29-2014, 07:26 PM.

                  Comment


                  • #79
                    Of course maybe someone on Gentoo can get even 90 fps, but that is it currently .

                    Comment


                    • #80
                      Originally posted by cbxbiker61 View Post
                      Yes,

                      I can't blame anyone for "wrong" results. All I can say is that my results with the cheapo MSI motherboard and 8G of good ram paint an entirely different picture. From these results you are pretty much seeing a doubling of performance over E-350, and I might add that the system (only the motherboard and cpu are different from my previous e-350) idles at 3 Watts lower, and peeks pretty much the same 30 Watts.

                      No one in their right mind would run only 4G of ram on a multi-purpose machine. If it was a dedicated single-app server sure, but not a desktop.

                      http://openbenchmarking.org/result/1...KH-1404098KH91

                      the AMD Kabini 5350 is a very good low power cpu!
                      I don't know of many cases, multipurpose or otherwise, where you would need more than 4GB RAM. I'd like to know what kind of software you are using to consume so much RAM.

                      Comment

                      Working...
                      X