There is something call fixed point numbers (as opposed to floating point), and it sounds like that's what you need. In this case, I fixed the point (B-scaling) to 0 (unsigned) integer bits (allowing numbers from 0 to +1, not including +1).
foreach (
0.750
0.500,
0.400,
0.250,
0.125,
) {
my $num = $_;
my $num_B0 = $num * 4294967296; # 0..+1 not incl +1.
#my $num_B1 = $num * 2147483648; # 0 to just over +1.
printf("%f = %s B0$/",
$num,
unpack('B32', pack('N', $num_B0)),
);
}
__END__
output
======
0.750000 = 11000000000000000000000000000000 B0
0.500000 = 10000000000000000000000000000000 B0
0.400000 = 01100110011001100110011001100110 B0
0.250000 = 01000000000000000000000000000000 B0
0.125000 = 00100000000000000000000000000000 B0
.5, .25 and .125 are round numbers cause they're 2^(-1), 2^(-2) and 2^(-3) respectively.
The machine for which I write software doesn't support floating point numbers, so we use fixed point arithmetic (which any integer arithmetic machine can do). It can be a pain at times, since we deal with an extreme number of decimal numbers like valve positions, temperature measurements, etc. (It's the control system of a nuclear power plant.)
Let me know if you want to know how to do math with fixed point numbers.