Help me creat a 40 bit variable for PIC18LF8722

[gravi]

Hi
This is Ravi and working on PIC18LF8722 and mcc 18 compiler, can anyone help me creating 40 bit variable.

Thanking you
Ravi

 

[alexz]

Re: 40 bit Variable

That would be an array of 5 bytes.
Can you explain why you would need such a variable?

 

[gravi]

Re: 40 bit Variable

That would be an array of 5 bytes.
Can you explain why you would need such a variable?

I have been trying to capture an 24 bit register and add the existing value to the current value, since I have taken an long(32-bit) variable by type casting to the to do the above process it's getting over flowed, to prevent over flow I would like to take an 40 bit register.


Thanking you,
Ravi kumar

Added after 16 seconds:

That would be an array of 5 bytes.
Can you explain why you would need such a variable?

I have been trying to capture an 24 bit register and add the existing value to the current value, since I have taken an long(32-bit) variable by type casting to the to do the above process it's getting over flowed, to prevent over flow I would like to take an 40 bit register.


Thanking you,
Ravi kumar

 

[alexz]

Re: 40 bit Variable

Never heard of data types longer then 32 bits

 

[CMOS]

Re: 40 bit Variable

Can't you use float variables, instaed of Long?

 

[alexz]

Re: 40 bit Variable

Can't you use float variables, instaed of Long?

Float is 32 bits too

 

[CMOS]

Re: 40 bit Variable

Floats are 32 as well as 64 bits. (Single and Double Precision)
Plus a 32 bit float can store larger value than a 32-bit long integer.

 

[alexz]

Re: 40 bit Variable

Float is 32 bits in PICs

 

[CMOS]

Re: 40 bit Variable

Float is 32 bits in PICs

Well that actually depends on the compiler and not the controller.

However even if its 32-bit float, it will store value larger than what a 32-bit long integer can store.

The limits of 32-bit unsigned long is : 0 to 4294967295
where as that for 32-bit float is : ±1.175494E-38 to ±3.402823E+38

Using float even if 32-bit long addition overflows, the float result variable will be able to handle it.

 

[alexz]

Re: 40 bit Variable

Yes CMOS you are probably right

 

[gravi]

Re: 40 bit Variable

Hi CMOS,
Thankyou, but my compiler(MCC18) doesnot support Float variables.

 

[alexz]

Re: 40 bit Variable

Hi CMOS,
Thankyou, but my compiler(MCC18) doesnot support Float variables.

It does

 

[gravi]

Re: 40 bit Variable

I have assigned "4294967295" to a unsigned long variable and an unsigned float variable and printed using %lu and %lu as %f is not supported by my compiler float variable prints some other variable other than i assaigned.

 

[alexz]

Re: 40 bit Variable

You are limited in the exponent which can be whithin the range -126 to +128

 

[C-Man]

Re: 40 bit Variable

gravi

i do not understand your problem

If you add 2 24 bit numbers it can never overflow a 32 bit value.

Example:

binary 111111111111111111111111 = hex FFFFFF
adding 111111111111111111111111 = hex FFFFFF gives hex 1FFFFFFE

But this only takes 25 bits of the 32 bit variable so there can not be an overflow.

Did I get this wrong?

best regards