Perl does its arithmetic with double-precision floating point, so you probably get 52-bit precision.
for (49..54) {
$x=2**$_;
$y = $x + 1.1;
print $_, "\t", $y-$x;
}
49 1.125
50 1
51 1
52 1
53 2
54 0
In C++, the size of integer types isn't specified, but long long is probably 64 bits.
If you want to deal with really big numbers, you need to use a package like Math::BigInt.
If you only use ++ and reverse on your "numbers", perl's increment magic will make it Just Work (TM).
$x="100000000000000000000";
$x++;
print $x, "\n";
$x += 1;
print $x, "\n"';
100000000000000000001
1e+20