On Thu, 2003-05-15 at 10:28, Lou Rickard wrote: > A friend of mine reccommended that I play around with > compiler optimizations a bit to see what happens. I > was a little surprised. > > I have an application that's about 10k lines of C code > written for linux. It's multi-threaded. It doesn't > do any complicated math. It compiles to about 1.8MB. > > Compiled with no optimizations its size is 1870614 > bytes. > > Compiled with -O2, it has the same size. According to > diff, it's identical to the non-optimized version. > > Compiled with -Os, it has the same size, and according > to diff, it's identical to the non-optimized version. > > Compiled with -O2 -march=pentium3 -mcpu=pentium3 > -funroll-loops, it has the same size, and according to > diff, it's identical to the non-optimized version. > > I'm running RH 7.2, with RH's 2.9.6 version of gcc, on > a pentium3. > > So, I'm a little confused, it seems like the different > optimizations would have had some impact, even if it > wasn't an improvement. Is it me? Is it the compiler? > Is it the relationship I have with the compiler? > Have I failed to nurture and provide comfort to it in > some way? Where did I go wrong? > > Also, what other optimizations would people reccommend > that wouldn't break the portability of the code itself > (ok to take out Makefile/compiler optimizations, not > ok to write processor specific code)? > > L.R.
IIRC, gcc-2.96 optimization was largely disabled. I highly recommend upgrading to Red Hat Linux 9 or Mandrake 9.1 with gcc-3.2.x, which has excellent optimization. Warren
