Maximum Decimal Values in Mat(h)

SB is all integer math.

All arrays except the precision array are 16 bit, which means
Decimal 0 to 65535
Hex     0 to FFFF
The precision array is 32 bit (signed), which means

Decimal - 2,147,483,648 to + 2,147,483,647
Hex       8000 0000     to   7FFF FFF
But as you observed, you can do floating point work if you keep track of the decimal place.

Addition is straight forward, but remember if you multiply you add the number of decimals in each multiplicand.

Example
    - 10.03 x 3 if working with two decimal places is 1003 x 300 = 300900.  Then divide by 10000 to get final answer = 30.09
    - or you can do 1003 x 3, then you only need to divide by 100 to get answer

Ok finally found in the manual for precision array to have the letter P before the regular array number… but why wouldn’t you want all arrays to have the capabilities of precision arrays? memory? speed of processing?

Ah I see... so overkill for most of what happens through MidiFire, ok fair enough.

I ran into another limitation it seems... can't use precisions arrays in calculations sometimes?


_____
#calc 1 does not work

ASS PL1 = $4000000
Log PL1 PL1 +D
ASS PL2 = $2000
Log PL2 PL2 +D

mat PL3 = PL1 / PL2
Log PL3 PL3 +D


#calc 2 works

mat PL4 = $4000000 / $2000
Log PL4 PL4 +D




even with small values this does not work...