The size of ac_fixed changes between signed and unsigned

[agustinhurquiza]

I am observing that the ac_fixed sizes change when I use unsigned 32 bits.

Here is an example:

ac_fixed<32, 16, true> x = 3.14;
std::cout << "Sizeof: " << sizeof(x) << " bytes\n";
[Out] Sizeof: 4 bytes

ac_fixed<32, 16, false> x = 3.14;
std::cout << "Sizeof: " << sizeof(x) << " bytes\n";
[Out] Sizeof: 8 bytes

This happens in version AC Datatypes v4.4.1

I'm not sure this is a bug, but I did find this effect strange, forcing me to change my code to accommodate this.

Compilation Info:

* Oracle Linux Server release 7.9
* g++ (GCC) 4.8.5 20150623 (Red Hat 4.8.5-44.0.3)
* g++ example.cpp -o run -I ../include/ -O3 -Werror -Wall -g -std=c++11

[jmartinl]

I don't think it's a bug, because of the following reason:
ac_fixed template is defined as
class ac_fixed : private ac_private::iv<(W+31+!S)/32>
That is, the number of elements in iv is the number of 32 bit elements required to hold the value. But also whether it's signed or not, influences the sum.
If S is true (signed) and W is 32, you get indeed (32+31)/32 (=1), but if S is false (unsigned), you get (32+32)/32 (=2), as one bit more is needed for the sign.