Announcement

Collapse
No announcement yet.

AMD Kaveri: Open-Source Radeon Gallium3D vs. Catalyst 14.12 Omega Driver

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

  • #21
    These APUs aren't really gamer chips to begin with, so what is with the focus on game performance? It would be much more useful to see BUSINESS SOFTWARE type workloads. You know, like scrolling through a million lines of spreadsheet with 15 sheets open on 2-4 monitors with a dozen firefox windows with a dozen tabs each, several terminals and file browsers, eclipse.... and all open at the same time. Lets see how much shit can be open before scrolling gets chunky. STUFF THAT YOU WOULD ACTUALLY DO WITH THIS TYPE OF HARDWARE!!!!

    Comment


    • #22
      Originally posted by droidhacker View Post
      These APUs aren't really gamer chips to begin with, so what is with the focus on game performance? It would be much more useful to see BUSINESS SOFTWARE type workloads. You know, like scrolling through a million lines of spreadsheet with 15 sheets open on 2-4 monitors with a dozen firefox windows with a dozen tabs each, several terminals and file browsers, eclipse.... and all open at the same time. Lets see how much shit can be open before scrolling gets chunky. STUFF THAT YOU WOULD ACTUALLY DO WITH THIS TYPE OF HARDWARE!!!!
      Maybe you think so, but Catalyst on Windows install raptr Gaming Evolved app by default on all supported chips even on little 2 core 1 GHz semprons So all can be used for gaming, casual lightweight or demanding often... well that depends
      Last edited by dungeon; 16 December 2014, 09:52 AM.

      Comment


      • #23
        dual channel RAM

        Originally posted by przemoli View Post
        Michael did clarify that previous test where done on dual channel RAM. ...)
        Michael should clarify that THIS test was done on dual channel RAM!

        Comment


        • #24
          RANK

          Originally posted by tomtomme View Post
          ... does this also tell about single/dual RANK?
          RANK has no impact on performance.
          Maybe you wanted to write single/dual channel?

          Comment


          • #25
            The test was done on dual-channel DDR3-1600MHz (all of my 2133MHz memory is currently in use in the 32 system test farm), which is fine given that for all driver configurations tested in this article, the same exact hardware and software settings were used.

            FYI, when running as root, the Phoronix Test Suite does attempt to report the memory configuration, frequency, and even the RAM model numbers, etc. However, it's only done as root as dmidecode, et al will only work as root. Additionally, for a lot of motherboards/chipsets, the decoding fails and there's various other issues. I've yet to see a proper universal solution for finding a nice way to obtain all of the memory information of a system -- especially as non-root.
            Michael Larabel
            https://www.michaellarabel.com/

            Comment


            • #26
              Originally posted by Michael View Post
              The test was done on dual-channel DDR3-1600MHz ...
              Thank you.

              ... which is fine given that for all driver configurations tested in this article, the same exact hardware and software settings were used. ...
              But it is not fine, if you want to compare to discrete cards.
              If I spend my dollars for an A10-7850K APU I will not run it with 1600 MHz RAM.

              Comment


              • #27
                Originally posted by drSeehas View Post
                But it is not fine, if you want to compare to discrete cards.
                If I spend my dollars for an A10-7850K APU I will not run it with 1600 MHz RAM.
                You are free to add +50%+ more performance to this benchmark, if you use 2xDDR3@2400MHz modules with A10-7850K APU slightly overclocked and of course cpufreq performance

                Comment


                • #28
                  Originally posted by dungeon View Post
                  Maybe you think so, but Catalyst on Windows install raptr Gaming Evolved app by default on all supported chips even on little 2 core 1 GHz semprons So all can be used for gaming, casual lightweight or demanding often... well that depends
                  "by default"... "on all". Kinda implies that they just do it all the time and it is therefore entirely unrelated to what the chip is likely to actually be USED for.

                  And there is a HUGE difference between "can", and "likely". The point is that any consideration of hardware has to be based on how it performs on the workloads for which it is LIKELY to be used for.

                  Comment


                  • #29
                    Originally posted by droidhacker View Post
                    These APUs aren't really gamer chips to begin with, so what is with the focus on game performance? It would be much more useful to see BUSINESS SOFTWARE type workloads. You know, like scrolling through a million lines of spreadsheet with 15 sheets open on 2-4 monitors with a dozen firefox windows with a dozen tabs each, several terminals and file browsers, eclipse.... and all open at the same time. Lets see how much shit can be open before scrolling gets chunky. STUFF THAT YOU WOULD ACTUALLY DO WITH THIS TYPE OF HARDWARE!!!!
                    That is just one opinion. Other people might argue that APUs are great for budget gaming PCs, since they do have a great integrated GPU. What do you do with that GPU in a business environment? Etc..

                    Comment


                    • #30
                      Originally posted by Nietzsche View Post
                      Hi. While testing APUs, you should include more information about the RAM used (single or dual channel, frequency and CAS latency) since it's an important performance factor.
                      He's using the same ram for both tests. So as a comparison it is valid.
                      Also, I believe he uses the fastest ddr3 available.

                      Comment

                      Working...
                      X