Expect to underflow to 0 and overflow to biggest value, but results below seem opposite? I train ML and had nan from dividing by big # in gradients julia> 1f0/2^64 Inf32 julia> 1f0/(2^64) Inf32 julia> 1f0*(2^64) 0.0f0 julia> 1/(2^64) Inf julia> 1.0*(2^64) 0.0

Weird floating point under/overflow behavior

mbauman March 31, 2024, 1:26am 2

This isn’t about floating point — it’s the integers! Check the value of 2^64.

4 Likes