Announcement

Collapse
No announcement yet.

New GCC 5.0 Changes, Command-Line Options That Landed So Far

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

  • New GCC 5.0 Changes, Command-Line Options That Landed So Far

    Phoronix: New GCC 5.0 Changes, Command-Line Options That Landed So Far

    While we're just a few months into the GCC 4.10 release cycle that's going to be released as GCC 5, there's already some release notes forming for this 2015 open-source compiler update...

    http://www.phoronix.com/vr.php?view=MTc3MzE

  • #2
    One small but upcoming feature that I like is that gcc can be configured to generate instructions for a default target. The configure option "--with-arch=" will set the default and so one can have the compiler produce slightly better code without having to touch any Makefiles, setting environment variables or passing options to various configure and build scripts.

    It is small, it is simple, but it is one of those things where you have to ask yourself: what took it so long?!

    Comment


    • #3
      Also in but not listed in changes yet are some warning options which will help you help the compiler devirtualize function calls
      http://hubicka.blogspot.com/2014/08/...-5-asking.html

      Comment


      • #4
        Originally posted by sdack View Post
        One small but upcoming feature that I like is that gcc can be configured to generate instructions for a default target. The configure option "--with-arch=" will set the default and so one can have the compiler produce slightly better code without having to touch any Makefiles, setting environment variables or passing options to various configure and build scripts.

        It is small, it is simple, but it is one of those things where you have to ask yourself: what took it so long?!
        Upcoming? I remember using that in 4.2 already.

        Comment


        • #5
          Originally posted by curaga View Post
          Upcoming? I remember using that in 4.2 already.
          No. You are thinking of the the "-march=" switch. I am talking about something new. A configure option to set the default.
          Last edited by sdack; 08-26-2014, 05:30 PM.

          Comment


          • #6
            C: -Wno-incompatible-pointer-types

            Hopefully it'll finally catch broken void **/void * conversions.

            - Gilboa
            DEV: Intel S2600C0, 2xE52658V2, 32GB, 4x2TB + 2x3TB, GTX780, F21/x86_64, Dell U2711.
            SRV: Intel S5520SC, 2xX5680, 36GB, 4x2TB, GTX550, F21/x86_64, Dell U2412..
            BACK: Tyan Tempest i5400XT, 2xE5335, 8GB, 3x1.5TB, 9800GTX, F21/x86-64.
            LAP: ASUS N56VJ, i7-3630QM, 16GB, 1TB, 635M, F21/x86_64.

            Comment


            • #7
              Originally posted by sdack View Post
              No. You are thinking of the the "-march=" switch. I am talking about something new. A configure option to set the default.
              Yes, I am talking about compile-time minimum cpu setting.

              Comment


              • #8
                To be extra clear: GCC's compile-time minimum cpu setting.

                Comment


                • #9
                  Exhibit A: a commit from 2003 mentioning the configure option.

                  http://repo.or.cz/w/official-gcc.git...7a5ba9027aff9c

                  Comment


                  • #10
                    Originally posted by curaga View Post
                    Exhibit A: a commit from 2003 mentioning the configure option.

                    http://repo.or.cz/w/official-gcc.git...7a5ba9027aff9c
                    Interesting. I have not seen the option before 4.10.

                    Comment

                    Working...
                    X