> I still like this the best:
>
>         LD      HL,srcpntr
>         LD      DE,destpntr
>         LD      BC,bytestomove
>         LDIR

Much too slow.  Better is (if memory serves) a routine like this, which moves
8 bytes per loop faster than an LDIR.

        DI      ; Disable interrupts
        mov     (OLD_SP),SP     ; save stack pointer
        mov     SP,(source)
        mov     HL,(dest)
        mov     a,(octets)
        clc
        subb    HL,SP
        exx
        mov     HL,(source)
        mov     DE,(dest)
        clc
        sbb     HL,DE
        inc     HL
        clc
Loop:
        pop     BC
        pop     DE
        exx
        pop     BC
        pop     DE
        adc     SP,HL
        push    DE
        push    BC
        exx
        push    DE
        push    BC
        adc     SP,HL
        dec     A
        jp      nz,Loop
        ei

The pushes, pops, and exx's take 28 cycles per loop.  I think the ADC's are
13 cycles each; the dec and jp nz total 14 more.  So 28+26+14 = 68 cycles
to move 8 bytes, an average of 8.5 cycles/byte.  If memory serves, that's
about twice as fast as LDIR.