Intel assembler optimization

Question

I am currently trying to optimize the code emitted by a makeshift compiler for home text.

I tried Intel VTune to find out where the bottlenecks are: http://www.imada.sdu.dk/~sorenh07/misc/vtune-assembly-optimization.png

I find it very impressive that the word "trimmed" is responsible for more than 38% of the hours in a program that runs for 30-90 seconds! Can someone give an explanation why?

VTune's "optimization report" function does not seem to exist for programs not compiled with icc. Is there a program that offers optimizations for assembler code? (i.e. not code coming from a high-level language).

Answer 1

I assume this is the idivl command, which actually takes 38% ... a longer division makes a little more sense than subtraction is not?