REPNE / REPNZ

REPNE and REPNZ name the F2h prefix. In front of SCAS or CMPS it repeats while CX is not 0 and the compare finds the elements different. Each round: stop if CX = 0, compare once, step SI / DI, decrease CX, stop if ZF = 1. Its main use is REPNE SCASB: search ES:DI for the byte in AL.

Test ZF (JE found / JNE not found) and not CX: a match on the last element leaves CX = 0 as well. Going forward over bytes the match is at DI - 1, its 0-based index is DI - start - 1, and N - CX elements were looked at.

CLD
MOV     DI, OFFSET TXT1
MOV     CX, 8                ; length of the text
MOV     AL, 'w'
REPNE   SCASB                ; stop at the first "w"
JNE     MISS1
MOV     AX, DI
SUB     AX, OFFSET TXT1 + 1  ; 0-based index: 3
ADD     AL, '0'              ; digit -> character
MOV     DL, AL
MOV     AH, 02h              ; DOS: print DL
INT     21h
JMP     DONE1
MISS1:
MOV     DL, '-'
MOV     AH, 02h
INT     21h
DONE1:
 
TXT1    DB      'hi world'

With CX = 0FFFFh and AL = 0 the same search measures a 0-terminated string (see SCASB / SCASW). REPNE CMPSB finds the first position where two strings hold the same byte, which is less common.

Pitfalls: forgetting CLD, forgetting that the search is in ES (set it in an .EXE), and reading DI as the match itself instead of one past it. In front of MOVS, STOS or LODS use REP, not REPNE.