Announcement

Collapse
No announcement yet.

See How Your Linux System Performs Against The Latest Intel/AMD CPUs

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

  • See How Your Linux System Performs Against The Latest Intel/AMD CPUs

    Phoronix: See How Your Linux System Performs Against The Latest Intel/AMD CPUs

    This holiday weekend (in the US) can be a great time to test your Linux system to see how it's performing against the latest AMD and Intel processors to see if it's time for a good upgrade...

    http://www.phoronix.com/vr.php?view=MTc3NzE

  • #2
    Erm why do you test the intel cpus on performance and the amd on ondemand?!?

    ondemand is a bit broken on amd,
    at least on my phenomII performance govenor always gets a few percent more than ondemand
    Last edited by rudl; 08-31-2014, 12:15 PM.

    Comment


    • #3
      Originally posted by rudl View Post
      Erm why do you test the intel cpus on performance and the amd on ondemand?!?

      ondemand is a bit broken on amd,
      at least on my phenomII performance govenor always gets a few percent more than ondemand
      It's the kernel defaults... Modern kernels use the Intel P-State driver by default and opt to performance where as AMD CPUs use CPUfreq ondemand.
      Michael Larabel
      http://www.michaellarabel.com/

      Comment


      • #4
        I would like to see game benchmarks with vsync on, because that is default

        Really Michael, make governor performance for all benchmarks

        Comment


        • #5
          Originally posted by dungeon View Post
          I would like to see game benchmarks with vsync on, because that is default

          Really Michael, make governor performance for all benchmarks
          Intel P-state performance governor is not same as the which would run with AMD processor.

          P-state performance doesn't keep the highest frequency all the time.

          Comment


          • #6
            Originally posted by reCAPTCHA View Post
            Intel P-state performance governor is not same as the which would run with AMD processor.

            P-state performance doesn't keep the highest frequency all the time.
            Word is about governor ondemand, which is missfeature when it comes to the performance benchmarks .

            Comment


            • #7
              Lintian complains about the deb package, a lot of errors like this:

              E: phoronix-test-suite: wrong-file-owner-uid-or-gid etc/ 1000/1000

              Comment


              • #8
                Originally posted by reCAPTCHA View Post
                Intel P-state performance governor is not same as the which would run with AMD processor.

                P-state performance doesn't keep the highest frequency all the time.
                Just a note ondemand is known to behave like this and everybody knows that It is not only on AMD processors, on Intel earlier than SNB too or if you wish on PPC64 processors too

                http://forum.hyperion-entertainment....&t=2569#p29072

                And now please do STK benchmark on Intel and AMD default i will laugh
                Last edited by dungeon; 08-31-2014, 03:13 PM.

                Comment


                • #9
                  Originally posted by dungeon View Post
                  I would like to see game benchmarks with vsync on, because that is default

                  Really Michael, make governor performance for all benchmarks
                  Few games enable V-Sync by default.

                  Comment


                  • #10
                    You know I tried to run agains this, buy:

                    Code:
                    HPC Challenge 1.4.3:
                        pts/hpcc-1.1.4 [Test / Class: G-HPL]
                        Test 1 of 10
                        Estimated Trial Run Count:    3
                        Estimated Test Run-Time:      1 Hour, 39 Minutes
                        Estimated Time To Completion: 3 Hours, 58 Minutes
                    Really? I just aborted the whole thing.

                    This is my biggest issue with many of the tests in the phoronix test suite, they take forever. It would be nice to limit the runtime of these tests to something reasonable, a few minutes tops. This is what I love of tests like geekbench. Sure, they are not real life, but most of the time i need a quick test, not an overnight one.

                    Comment


                    • #11
                      Originally posted by Calinou View Post
                      Few games enable V-Sync by default.
                      When someone do GPU benchmarks he disable it for sure in the driver. The same is with ondemand, when someone do gaming benchmarks he disable it, because of known (miss)behavior

                      Both vblank and ondemand are good default, but for performance benchmarks you surely don't want it

                      Comment


                      • #12
                        Originally posted by mendieta View Post
                        You know I tried to run agains this, buy:

                        Code:
                        HPC Challenge 1.4.3:
                            pts/hpcc-1.1.4 [Test / Class: G-HPL]
                            Test 1 of 10
                            Estimated Trial Run Count:    3
                            Estimated Test Run-Time:      1 Hour, 39 Minutes
                            Estimated Time To Completion: 3 Hours, 58 Minutes
                        Really? I just aborted the whole thing.

                        This is my biggest issue with many of the tests in the phoronix test suite, they take forever. It would be nice to limit the runtime of these tests to something reasonable, a few minutes tops. This is what I love of tests like geekbench. Sure, they are not real life, but most of the time i need a quick test, not an overnight one.
                        HPCC is really an exception... Most PTS tests don't take longer than a couple minutes with modern hardware, but HPC Challenge is notoriously big and depending upon your core count and RAM capacity can be quite lengthy, but many people love the test.

                        You could also set SKIP_TESTS=hpcc environment variable if you want to skip over it.
                        Michael Larabel
                        http://www.michaellarabel.com/

                        Comment


                        • #13
                          CPU on max, governor performance, blah, blah

                          Comment


                          • #14
                            Originally posted by Calinou View Post
                            Few games enable V-Sync by default.
                            If by few you mean almost all games. Not having v-sync would be stupid. Disabling v-sync looks bad and lowers performance by forcing the machine to do pointless busy work for no benefit. You only disable it for benchmark and if you are a pro-gamer who favors a few more milliseconds of reaction time over graphics.
                            Last edited by carewolf; 08-31-2014, 05:15 PM.

                            Comment


                            • #15
                              I have my 4930k going now. Cant wait to see how she does!

                              Comment

                              Working...
                              X