Integer math library for sxb 2.0

Peter Verkaik

Attached is a library for integer math.
This library contains wrapper functions that implement
the *, /, //, */ and ** operators for 16bits signed and unsigned variables,
plus 32bits addition and subtraction. As these operators generate quite
a bit of code, calling a subroutine might save some codespace.
I didn't bother including wrapper functions for 8bit variables as the calling
overhead would prabably cancel out any space saving.

regards peter

Peter Verkaik

I had forgotten to enclose all sx type conditional compile blocks
inside a seperate $ifused block. Sorry about that.

regards peter

Zoot

Peter -- this libs are fairly similar to my own math libs, although I handle banking for tmp variables a little differently. In any case, one thing I noticed is that you are not accounting for rounding, something I do in my own functions, e.g. (simplified example, my actual libs have more ASM)....

FUNC MATH_DIV
   IF __PARAMCNT < 4 THEN
      __PARAM4 = 0
      IF __PARAMCNT < 3 THEN
         __PARAM3 = __PARAM2
         __PARAM2 = 0
      ENDIF
   ENDIF
   tmpW1 = __WPARAM12
   tmpW2 = __WPARAM34
   \CLC                              ' add 1/2 of divisor for cleaner rounding
   \RR __PARAM4
   \RR __PARAM3
   tmpW1 = tmpW1 + __WPARAM34
   tmpW1 = tmpW1 / tmpW2
   RETURN tmpW1
   ENDFUNC

For my mid-multiply (*/) function I do something similar, but obviously #128 is added to the first word since the divisor is always 256.

I've found that this gives me much closer integer results, esp. when numbers may be manipulated a bit (a good example would be encoder odometry and trig).

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
When the going gets weird, the weird turn pro. -- HST

1uffakind.com/robots/povBitMapBuilder.php
1uffakind.com/robots/resistorLadder.php

Peter Verkaik

Library update.
I added wrapper functions for roundoff versions (thanks Zoot for mentioning that)
and some functions that deal with 16bit fractions, eg. converting between 65536 based fractions
and 10000 based·fractions which are·great for displaying decimals of a fractional number.
Also included a function (signed and unsigned) to calculate the 65536 based fraction for a given
numerator and denominator.


regards peter

·

QuattroRS4

Peter,
Thanks for that ..
Regards,
John Twomey

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
'Necessity is the mother of invention'

Those who can, do.Those who can’t, teach.