inline floating point

[bachrach@boromir.stanford.edu (Jonathan Bachrach)]

Is there anyway to get the compiler to generate inline floating point
instructions for machines with floating point coprocessors instead of
requiring a subroutine call for each floating point routine?

Also is there anyway to achieve zero consing floating point computations
without resorting to the dreaded C language.

Ultimately, i would like to have inline floating point within functions and
permit the return of a floating point result without consing.

Finally is there anyway to turn off range checking on arrays?  I tried
(optimize (speed 3) (safety 0)) without luck.

-- jonathan (bachrach@psych.stanford.edu)