Can you INC a BCD number?

[Maury Markowitz]

Just started playing with MAC/65, OMG I wish I had this as a kid! DDT is amazeballs! Anyway, check this out:

10 .ORG $4000
20 .FLOAT 123456789
30 SED
40 INC $4005

When you ASM this and look at $4000, you have:

44 01 23 45 67 89

Now G 4006. This will SED into BCD mode and then increment the last number. The result is:

 

44 01 23 45 67 8A

THAT'S interesting! Did I miss a step here? Or does the 6502 simply not consider BCD in INC?

 

 

[shanti77]

You must use ADC or SBC for BCD.

[thorfdbg]

12 hours ago, Maury Markowitz said:

44 01 23 45 67 8A

THAT'S interesting! Did I miss a step here? Or does the 6502 simply not consider BCD in INC?

 

 

Yup, reading the 6502 specifications for one thing. (-; The decimal flag is only honored by ADC and SBC, and thus not by INC and DEC. If you want to increment a BCD number, the only option is ADC. Or performing a half-carry check manually.

[Rybags]

The extra logic probably would have been too much.

And the usage case - it's fairly rare to do BCD operations in a direct way - it's usually a subroutine that's operating on two "variables" rather than constants.

 

[TGB1718]

Also remember FLOAT constants can only really be used by the ROM floating point routines

 

[Rybags]

Another thing - you're generally working on numbers greater than 2 figures so would want to carry over to the next significant set of digits.

 

INC works OK there since you can just do a BEQ to INC the next significant byte.

But with DEC you don't have a single flag to indicate you've underflowed from 00 to FF - so you'd probably end up doing more work than the SEC/SBC operation (for subtracting single-byte constants at least).  With SBC of a single byte value you could do a BCS to skip the next significant byte.

[Harry Potter]

Rybags: try loading the LSB and decrementing the MSB only if the LSB is 0. 

[Rybags]

You can but it gets to the point of why not just use the SBC anyway?

 

The feeling I have about it is it might have just came down to bloating out the component count just to provide seldom needed shortcut.

[Maury Markowitz]

4 hours ago, thorfdbg said:

Yup, reading the 6502 specifications for one thing. (-; The decimal flag is only honored by ADC and SBC, and thus not by INC and DEC. If you want to increment a BCD number, the only option is ADC. Or performing a half-carry check manually.

Instead I first check if the number is less than 99, at which point a carry is impossible and INC works fine.

[Wrathchild]

Are you sure? We are talking hex $99 here? If you have $09 then an INC will give you $0A and not $10 as you would want.

[Rybags]

Why the need for BCD anyway?  It can be handy for stuff like a game score.

But alternatively you could just use binary then look for quick unsigned bin->BCD conversion routines.

[TGB1718]

10 hours ago, Rybags said:

But alternatively you could just use binary then look for quick unsigned bin->BCD conversion routines.

If speed is not too much of an issue, use the internal FP routines, I wrote a couple of disk copiers bitd

and wrote the current sector to screen using FP routines which were fast enough to update the

screen between sectors.

 

The current Sector was in DAUX1 and DAUX2, copy them to $D4  and $D5

call IFP to convert to FP

call FASC to convert to ASCII

if you want to put the ASCII directly to screen subtract $20 from each character,

last character has high bit set, so AND #$7f, then subtract $20

 

[xxl]

 

BCD INC in baroque style

 

 

     lax BCD
     jsr Baroque_BCD_INC
     sax BCD

     jmp *

 

BCD .HE 89

 

Baroque_BCD_INC
     and #$F0
     cmp #$90
     php
     txa
     and #$0F
     cmp #$09
     php
     txa
     sbx #$100-$01
     plp
     txa
     bcs @+
     plp
     rts
@    sbx #$100-$06
     txa
     plp
     bne @+
     ldx #$00
@    rts
 

Edited January 30, 2022 by xxl

[+Stephen]

OMG - you do realize that by using LAX and SAX, you are in danger of having a class action lawsuit brought against you?

[TGB1718]

Don't forget he used 'sbx' too, that's 3 strikes, he's out

 

[ivop]

Why? All official Atari 8-bit computers run that code just fine. Even Bill Mensch calls them undocumented opcodes. They are not illegal or not allowed or something.

 

https://ataripodcast.libsyn.com/antic-interview-96-bill-mensch-6502-chip

 

[+David_P]

1 hour ago, ivop said:

Why? All official Atari 8-bit computers run that code just fine. Even Bill Mensch calls them undocumented opcodes. They are not illegal or not allowed or something.

 

https://ataripodcast.libsyn.com/antic-interview-96-bill-mensch-6502-chip

 

Upgraded computers may not run properly, though.

[xxl]

1 hour ago, David_P said:

Upgraded computers may not run properly, though.

so what was the point of modifying if the standard stuff doesn't work anymore?

[+DjayBee]

1 hour ago, David_P said:

Upgraded computers may not run properly, though.

Oh Kyle22, where art thou?

[ClausB]

Another thing that doesn't work as you might expect in BCD mode is the Z flag. My last BCD project was frustrated by that for a while:

 

 

[+Stephen]

5 hours ago, ivop said:

Why? All official Atari 8-bit computers run that code just fine. Even Bill Mensch calls them undocumented opcodes. They are not illegal or not allowed or something.

 

https://ataripodcast.libsyn.com/antic-interview-96-bill-mensch-6502-chip

 

 

3 hours ago, David_P said:

Upgraded computers may not run properly, though.

That's the joke!