go to the make your prog, click options -> compile -> advanced optimizations

now here u have several options, and u can click help if u need it

pretty safe bets are
--assume no aliasing (if u dont know what this is youre 100% not using it)
--remove safe pentium FDIV checks - this applies to some floating point bug in pentium 120mhz and lower!

others
--------
the array, integer, and floating point checks mean that where an overflow error would occur u simply get a value with no error, and the value may not be correct

Allow Unrounded Floating Point Operations for all projects, heres what msdn says
When selected, allows the compiler to:

Use floating point registers more efficiently


Avoid loads and stores from memory


Do floating point comparisons more efficiently

Note Using this option may result in calculations being maintained to a higher precision than expected and may cause the comparison of two floating point values to show them unequal when you expect them to be equal.