hi all. i was looking at the optimization options that gcc provides and am confused by a couple (for now!) -ffast-math a> -fno-math-errno tried some programs using sqrt function. should it not be using the FSQRT instruction in the pentium? it does not seem to be doing so. so no difference on x86. should this be a glibc query instead? b>-funsafe-math-optimizations turned this option on with some functions like pow but there was no difference c>fno-trapping-math no idea what this does! anyone here with ideas on how to test this option? cheerio, ravi P.S: also if anyone has info about -fexpensive-optimizations, do let me know.
