> When I compile the C code with -O0 (with gcc -o nbody.gcc -Wall
> --fast-math nbody.c -lm), I get a time of 1.513s which is comparable
> to OCaml (1.607s).  But as soon as I turn on -O options (as with gcc
> -o nbody.gcc -Wall -O1 --fast-math nbody.c -lm), the running time
> drops down to 0.871s (0.58%).  Can somebody tell me what is the
> optimization that has such an effect and whether it could be applied
> to OCaml ?

gcc -O0 sets out to generate the worst possible code, and mostly
succeeds.  Optimizations in gcc -O1 compared to gcc -O0 include
register allocation, dead code elimination, branch straightening,
common subexpression elimination, instruction combining, and
instruction scheduling.

ocamlopt-generated code is between -O0 and -O1 in quality, usually
much closer to -O1.  The biggest missing optimization is common
subexpression elimination.  ocamlopt puts less effort into instruction
combining than gcc.

gcc -O2 adds loop optimizations which ocamlopt never does.

A functional programming style puts different demands on the
optimizer.  ocamlopt has some optimizations that don't make
sense for C, e.g. replacing (unbox (box (value))) with value.