|
Feb 28, 2003
RE: Floating point math within Vedit
Mr. Lindgren has written about having Vedit do arbitrary precision
floating point math. Something like Unix bc or dc. I wrote a Vedit
macro which is arbitrary precision calculation of floating point
numbers. To use FPMATH10.ZIP the Vedit user places each part of
an equation into some text registers. Next the user passes the
text registers numbers to my subroutine. It builds these parts into
a single equation and solves it. It returns the results back to a
text register.
The benchmark for this FPMATH10.ZIP is about 63. msec per equation.
Arbitrary precision performance is a function of math operator
and string lenght. It is a wide range. Multiplication is slow.
Division is very slow.
Mr Ziemski's FLOATMATH.VDM, is 2.2 msec per equation.
My earlier work, FPMATH09.ZIP, is .42 msec per equation.
Use Vedit System("command") to solve the equation about is 200 msec
per equation.
One of these methods may guide the few Vedit users that want
floating point math within Vedit. I will Email FPMATH09.ZIP or
FPMATH10.ZIP when requested.
Thank you
--
kenjen@... Ken Jensen Portland, Oregon USA
---------------------------------------------------------
|
|