Loading ...
Sorry, an error occurred while loading the content.
 

A Blast from the Past: Trying out gcc-2.95.3 with Freecell Solver

Expand Messages
  • Shlomi Fish
    Randall Parker reports here: http://tech.groups.yahoo.com/group/fc-solve-discuss/message/614 that: {{{ Your solver [= Freecell Solver ] takes about 70% longer
    Message 1 of 1 , Apr 23, 2009
      Randall Parker reports here:

      http://tech.groups.yahoo.com/group/fc-solve-discuss/message/614

      that:

      {{{
      Your solver [= "Freecell Solver"] takes about 70% longer to run with the MS
      Visual C++ Toolkit 2003 as compared to MSVC v6.
      }}}

      Moreover, Ted Ts'o mentions in http://lwn.net/Articles/313798/ that:

      {{{{{{{{{{{
      (And BTW, I consider the pre-2.6.28 compile times only barely
      acceptable; most of this isn't kbuild's fault, I know, it includes the
      fact that link steps are still slow, and gcc is constantly getting
      increasingly slower over time. Sigh... I will just note wryly that
      it used to be that I could compile 0.9x kernels on a 40 MHz 386
      machine in 10 minutes. Some 15 years later, it still takes roughly
      the same amount of time to compile a kernel, even though computers
      have gotten vastly faster since then. Something seems wrong with
      that.... :-)
      }}}}}}}}}}}

      The Linux kernel has grown bigger and more complex since then, but it's also
      in large part to gcc becoming slower.

      Finally, I recently re-compiled a mini-graphical-demo (see:
      http://en.wikipedia.org/wiki/Demoscene ) I wrote in ANSI C using gcc (= the
      open-source compiler collection which is popular on Linux and other UNIX
      systems) again and the result seemed slower than before.

      As a result, I decided to try out gcc-2.95.3, which I still remember using
      back in the time.

      I downloaded the archive from the GCC site, and tried to build it. ./configure
      followed by "make" was not successful, and I realised I needed to do "make
      bootstrap". This was successful after one or two manual tweaks, and I was able
      to install it under my home directory.

      Here are the results:

      Using gcc-2.95.3:
      -----------------

      Compilation:

      {{{{
      5.74user 1.25system 0:08.40elapsed 83%CPU (0avgtext+0avgdata 0maxresident)k
      784inputs+8544outputs (0major+190805minor)pagefaults 0swaps
      }}}}

      Running the MS 32K: 285.193099021912

      Using gcc-4.3.2-5mnb2:
      ----------------------

      Compilation:

      {{{{{
      12.35user 1.51system 0:15.07elapsed 91%CPU (0avgtext+0avgdata 0maxresident)k
      304inputs+2912outputs (1major+229206minor)pagefaults 0swaps

      12.45user 1.45system 0:15.05elapsed 92%CPU (0avgtext+0avgdata 0maxresident)k
      16inputs+2912outputs (0major+229205minor)pagefaults 0swaps
      }}}}}

      Running the MS 32K: 235.637364864349

      -------------------------

      So gcc-2.95 compiles faster, but the resultant machine code is slower. A note
      is that the gcc-4.3.x run was ran in sub-optimal conditions and I've already
      measured benchmarks of 224.068351984024 seconds.

      My demo also doesn't seem faster when compiled with gcc-2.95.3.

      Regards,

      Shlomi Fish

      --
      -----------------------------------------------------------------
      Shlomi Fish http://www.shlomifish.org/
      What Makes Software Apps High Quality - http://xrl.us/bkeuk

      God gave us two eyes and ten fingers so we will type five times as much as we
      read.
    Your message has been successfully submitted and would be delivered to recipients shortly.