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

Topic		Replies	Views
Julia messes up integer exponents? Performance integer-overflow	106	6294	February 21, 2019
Odd result with negative exponents of 10 in Julia 1.5.2 New to Julia	4	851	October 19, 2020
Integer power of large Integer returning very wrong results General Usage base	7	516	May 15, 2023
Error for 396^n General Usage bug , error	8	526	February 12, 2024
Math operation known issue? a^b General Usage question , special-issue , math , integer-overflow	11	1573	May 6, 2020