PIC Microcontoller Subtraction Math Method

16-bit subtract routine for PIC16 instruction set. valid C and Z

if you can do better (smaller), tell me! this is the best I could do.


;------------------------------
;19/6/2006
;code by Christian Blondin
;]b[e]l[t]o[i]d[9]7[@s@p@a@m@]e[m]a[i]l.c]o[m -> delete the [][][[] @spam@
;rsc-mirror3.netfirms.com/pic.html -> coming up.
;
;code is for PIC16 instruction set (PIC10F, 12F, 16F...)
;remove the last two lines if you don't need a valid zero flag.
;performs [f01|f00] = [f01|f00] - [f03|f02] while keeping f03|f02 intact.
;(f00 and f02 being the low-bytes, f01 and f03 the hi-bytes)
;briefly tested in MPSIM ("special cases" only)

	movf	f02,W
	subwf	f00,F		;f00-f02 (STATUS,C=!BORROW)
	movf	f03,W
	btfss	STATUS,C	;skip if borrow=0 (C=1)
	addlw	01
	subwf	f01,F		;f01-f03	STATUS,C=!BRW
	btfsc	STATUS,Z	;skip if MSbyte<>zero
	movf	f00,F		;else, check LSbyte for zero.
;------------------------------

Comments:

• doesn't work for 0x0000 - 0xFFFF. The result should be 0x0001 and it's 0xFF01. The problem is that carry from LSB is added to f03 instead to result (f01). But of course that's done because carry flag will be lost at the end. So the only way is to store carry flag somewhere until the end.