JK> '  Extended precision is faster than double precision?  Wow!  Amazing!
 
This is probably because extended precision is the data type used by
the math coprocessor, and so PB doesn't have to do any type conversions
to your variables to store it in single/double.  Even if you put 
$float procedure in your code if you have a math coprocessor PB will
use it, unless you put a SET 87=NO in your environment table.  Your
results are interesting to know however, as that is something I've
never bothered to check on (or the constant folding either.  I am going
to have to go back and check some code of mine now that does heavy duty
calculating and make sure I take care of that.  I usually do it myself
when putting together a formula though).  Thanks for all the posted
results of your tests!     Eric
--- QM v1.00
---------------