3 ms·
What helped me better understand floats, was playing around with them in their bit representation. For example, if you go to https://www.h-schmidt.net/FloatCon
by danybittel 4y ago
What helped me better understand floats, was playing around with them in their bit representation.
For example, if you go to https://www.h-schmidt.net/FloatConverter/IEEE754.html https://www.h-schmidt.net/FloatConverter/IEEE754.html and enter 1.0. Then use +1 and -1 to go through ULPs.
- stabbles 4y agoAnother nice tool is the Julia repl: julia> 1.4 + eps(1.4) === nextfloat(1.4) true julia> bitstring(1.4) "0011111111110110011001100110011001100110011001100110011001100110" julia> bitstring(nextfloat(1.4)) "0011111111110110011001100110011001100110011001100110011001100111"
- OscarCunningham 4y agoThe site I really like for exploring the binary representation of floats is https://float.exposed https://float.exposed.