Announcement

Collapse
No announcement yet.

GCC 7.2 To Be Released In About Two Weeks

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

  • GCC 7.2 To Be Released In About Two Weeks

    Phoronix: GCC 7.2 To Be Released In About Two Weeks

    Richard Biener has shared plans to release GCC 7.2 in about two weeks as the newest point release to the GCC 7 compiler...

    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
    Has any distribution taken the approach of test compiling each package with different compilers and compiler versions and to choose the compiler for each package based on the best performance for that specific application?

    Seeing the performance tests on this site and the wildly differing results you get depending on the program that is being tested, I think this approach might improve the overall performance per watt of the system.

    Is this a bad idea or just too much work?

    Comment


    • #3
      Great.. That's when we start talking.. Since version 5, I usually wait for the .2 release to use GCC for production. Under the new scheme, the first "stable" release is basically a public request for bug reports. So now I consider any X.1 version more of an essential stage of thorough testing.. which don't get me wrong, I quite like. I would just put more emphasis on that, perhaps retouching the version scheme once more.
      Last edited by GdeR; 25 July 2017, 12:48 PM.

      Comment


      • #4
        i personally would prefer if central compilers would be made by others than the FSF since the FSF in my opinion haven’t been the most progressive. Even Clang ran passed GCC naturally.

        Comment


        • #5
          Originally posted by SomebodyElse View Post
          Has any distribution taken the approach of test compiling each package with different compilers and compiler versions and to choose the compiler for each package based on the best performance for that specific application?

          Seeing the performance tests on this site and the wildly differing results you get depending on the program that is being tested, I think this approach might improve the overall performance per watt of the system.

          Is this a bad idea or just too much work?
          I don't think most distros put that much thought in it besides Intel Clear Linux for compiler optimised offerings

          Comment

          Working...
          X