div and mul of BCD or Hex?

[JoeJoe]

Hello coder,

I am currently working on the display and calculation of scores. I have already read the thread "BCD Arithmetic - howto^". The simple adding and subtracting is no problem. But I have to divide the score by 50. Also the points can be increased by graduation 3*90 or 5*240 etc.
Now to my problem. For dividing and multiplying with BCD there are no 1:1 commands. Multiplying can be avoided by using LookUp tables and adding. But for x:50 there is no alternative.
Of course I can calculate everything in hex and convert the hex values to decimal for the display. But this eats a lot of cycles.
How can I solve the problem without overloading the computer.

Thanks in advance
JoeJoe

@Admins: please move to right section, sorry

[meynaf]

You can convert the divide into a multiply.
First, add the number to itself to do *2.
Then, remove the two last digits to do /100.
You will have achieved /50.

[alkis]

/50 can be achieved by /100 (dropping last byte) and add the number to itself (so /100*2)

[JoeJoe]

thanks for your replies

[8bitbubsy]

Originally Posted by meynaf

Then, remove the two last digits to do /100.

How is this achieved in a fast way when working with binary?

[meynaf]

Originally Posted by 8bitbubsy

How is this achieved in a fast way when working with binary?

Who cares ? This thread is about BCD, not binary.

[JoeJoe]

That's right, but we don't want to be like that

I do this with the reciprocal method. x/50 = x*(1/50)

[Thomas Richter]

Originally Posted by JoeJoe

I am currently working on the display and calculation of scores. I have already read the thread "BCD Arithmetic - howto^". The simple adding and subtracting is no problem. But I have to divide the score by 50. Also the points can be increased by graduation 3*90 or 5*240 etc.
Now to my problem. For dividing and multiplying with BCD there are no 1:1 commands. Multiplying can be avoided by using LookUp tables and adding. But for x:50 there is no alternative.

:50 is simple enough (already answered by others), but general multiplication or division in BCD is slow, in general. To multply, use the pen-and-paper algorithm of multiplying digit by digit, then add the partial products by a BCD add. Similarly, a BCD division follows the pen-and-paper algorithm of continuous subtraction until an overflow, and keeping the number of times a number was subtracted, then carry over the remainder, and continue with the next digit.


All these algorithms are slow...

[Thomas Richter]

Originally Posted by 8bitbubsy

How is this achieved in a fast way when working with binary?

If you have no division instruction, the same way as in BCD, except that each digit can be either 0 or 1, which simplifies continuous subtraction to a single compare. If you do, well, use the division instruction.