- Conversion and display of BCD (Binary Coded Decimal) values
- 07 Nov 2014 12:36:49 am
- Last edited by Unknownloner on 07 Nov 2014 12:42:58 am; edited 1 time in total
Binary coded decimal
Original here
No idea if this sort of thing has been posted before, but I wrote this for displaying scores in 2048
and figured it might be useful to somebody.
One method for display numbers that take more than 16 bits is to convert it to Binary Coded Decimal
first, and display the result. BCD works by using four bits to store each decimal (base 10) digit of
a number. The following code can convert a number to BCD, and display it. It's currently written to
convert a 24 bit number to a 10 digit BCD number, but can be modified to support anything really. It
is memory ineffecient, because it uses one byte for each digit rather than storing two digits per
byte. This is useful though because it makes the display routine simpler.
Conversion to BCD
This routine converts a little endian value to a little endian BCD value
Code:
Displaying BCD
Displaying BCD is quite simple, since each digit is stored within its own byte. You can simply add
the char code for '0' to the value to get the text char you want. This routine displays an unsigned
BCD value, without leading zeroes
Code:
In case anyone was wondering, no, this is not a heavily optimized routine. I tried to make it at least somewhat clear what the code was doing rather than hyper-optimize it.
Original here
No idea if this sort of thing has been posted before, but I wrote this for displaying scores in 2048
and figured it might be useful to somebody.
One method for display numbers that take more than 16 bits is to convert it to Binary Coded Decimal
first, and display the result. BCD works by using four bits to store each decimal (base 10) digit of
a number. The following code can convert a number to BCD, and display it. It's currently written to
convert a 24 bit number to a 10 digit BCD number, but can be modified to support anything really. It
is memory ineffecient, because it uses one byte for each digit rather than storing two digits per
byte. This is useful though because it makes the display routine simpler.
Conversion to BCD
This routine converts a little endian value to a little endian BCD value
Code:
;Converts a 24 bit unsigned int pointed to by HL to BCD
;Convert to BCD with double dabble method, see http://en.wikipedia.org/wiki/Double_dabble
;Handles up to a 10 digit number
;Input: HL - pointer to number
;Output: bcdScratch - stores BCD number.
NUM_DIGITS = 10
NUM_SRC_BYTES = 3
.var NUM_DIGITS, bcdScratch
.var NUM_SRC_BYTES, bcdSource
ConvertToBCD:
ld de,bcdSource
ld bc,NUM_SRC_BYTES
ldir
ld ix,bcdSource
xor a
ld hl,bcdScratch
ld b,NUM_DIGITS
_zeroScratch:
ld (hl),a
inc hl
djnz _zeroScratch
ld b,NUM_SRC_BYTES * 8
_bcdConvLp:
;Do increment
ld c,NUM_DIGITS
ld hl,bcdScratch
;Iterate through each BCD digit.
;If digit > 4, add 3
_bcdIncLp:
ld a,(hl)
cp 5
jr c,{@}
add a,3
@:
ld (hl),a
inc hl
dec c
jr nz,_bcdIncLp
;Shift SRC bits
sla (ix)
.for _off, 1, NUM_SRC_BYTES - 1
rl (ix + _off)
.loop
ld c,NUM_DIGITS
ld hl,bcdScratch
_bcdShiftLp:
ld a,(hl)
rla
bit 4,a
jr z,{@}
and %1111 ;Mask out high bits, since we only want the lower 4 bits for the digit
scf ;Set carry if bit 4 set
@:
ld (hl),a
inc hl
dec c
jr nz,_bcdShiftLp
djnz _bcdConvLp
ret
Displaying BCD
Displaying BCD is quite simple, since each digit is stored within its own byte. You can simply add
the char code for '0' to the value to get the text char you want. This routine displays an unsigned
BCD value, without leading zeroes
Code:
;Displays the BCD value at HL
;1 byte per digit
NUM_DIGITS = 10
DispBCD:
ld de,NUM_DIGITS - 1
add hl,de ;Go to end
;Skip leading zeroes, except if the value IS zero
ld b,NUM_DIGITS - 1
_skipLeadingZeroes:
ld a,(hl)
or a
jr nz,{@}
dec hl
djnz _skipLeadingZeroes
@:
inc b ;B = num digits to display
_dispBCDDigits:
ld a,(hl)
add a,'0'
push hl
push bc
;Replace this with anything that takes a char code in 'A' and displays it
b_call(_VPutMap)
pop bc
pop hl
dec hl
djnz _dispBCDDigits
ret
In case anyone was wondering, no, this is not a heavily optimized routine. I tried to make it at least somewhat clear what the code was doing rather than hyper-optimize it.
Great resource! 
