Announcement

Collapse
No announcement yet.

GCC vs. LLVM Clang Compiler Performance On AMD's Ryzen

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

  • GCC vs. LLVM Clang Compiler Performance On AMD's Ryzen

    Phoronix: GCC vs. LLVM Clang Compiler Performance On AMD's Ryzen

    Our latest AMD Ryzen Linux benchmarking is looking at the performance of the GCC and LLVM Clang compiler performance with a Ryzen 7 1700 on Ubuntu Linux.

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    Wins Losses is interesting... are you just counting a win as a win or are you doing something like allocating 100 pts per testing and the allocating those points based on the relative performance of each and then dividing back out by the total number of points in the end?

    Comment


    • #3
      Originally posted by cb88 View Post
      Wins Losses is interesting... are you just counting a win as a win or are you doing something like allocating 100 pts per testing and the allocating those points based on the relative performance of each and then dividing back out by the total number of points in the end?
      Just straight wins vs. losses, if anyone has any better ideas or complementary data, patches certainly welcome! The code is really easy in that part.
      Michael Larabel
      https://www.michaellarabel.com/

      Comment


      • #4
        The difference between winners and losses is quite interesting:

        GCC 6.3.0 18-4 14
        Clang 3.9.1 12-6 6
        GCC 7.0.1 5-9 -4
        Clang 4 rc2 2-18 -16
        Last edited by FireBurn; 08 March 2017, 11:14 AM.

        Comment


        • #5
          Originally posted by Michael View Post

          Just straight wins vs. losses, if anyone has any better ideas or complementary data, patches certainly welcome! The code is really easy in that part.
          would it be possible for "X Won by an [inch|mile]"? Take the % differences between win and loss, average them and if <10% then inch, else mile

          Comment


          • #6
            LLVM/Clang 4.0-RC3 has been out for at least a week now. All sorts of additions and optimizations.

            Comment


            • #7
              Stop using that EVGA 600W PSU
              Antec EarthWatts EA-650 Green's much better!
              Antec's VPF650 is a budget PSU made by Delta Electronics to achieve high-performance-per-dollar. In the US, the VPF units are sold under the EarthWatts Green series.
              Last edited by qsmcomp; 08 March 2017, 11:37 AM.

              Comment


              • #8
                Originally posted by Michael View Post

                Just straight wins vs. losses, if anyone has any better ideas or complementary data, patches certainly welcome! The code is really easy in that part.
                Should be weighted: a compiler which wins 10 times by a 0.1% margin is not better than another compiler which wins 9 times by a 70% margin!
                ## VGA ##
                AMD: X1950XTX, HD3870, HD5870
                Intel: GMA45, HD3000 (Core i5 2500K)

                Comment


                • #9
                  Originally posted by darkbasic View Post

                  Should be weighted: a compiler which wins 10 times by a 0.1% margin is not better than another compiler which wins 9 times by a 70% margin!
                  However Linux kernel can only be compiled with GCC at the moment.

                  Comment


                  • #10
                    There are so many factors, trying to put weighted averages just fools you into thinking you have better data than you do.

                    Comment

                    Working...
                    X