Integer power of large Integer returning very wrong results

[dmetivie]

I am with Julia 1.9 on a fresh session (windows). (Maybe it happens on previous version already)
Is it normal to get stuff like

julia> 5^2000
-7794321243379011391

julia> 2^200
0

With floating power 5^2000., it is Inf.

[gbaraldi]

That is just integer overflow. Normal julia Ints aren’t BigInts.

[dmetivie]

Why not Inf everywhere?

[gbaraldi]

Inf is a floating point thing. One could use checked integer arithmetic that would throw an exception on overflow like GitHub - JeffreySarnoff/SaferIntegers.jl: These integer types use checked arithmetic, otherwise they are as system types.. But they are slow, so julia doesn’t do that by default

[nilshg]

For the second time this week I can dig out this classic thread:

Discussion about integer overflow

Yes, it is a limitation but I would not call it big since we could use BigInt if Int64 is not enough and an overflow is something we can catch. I find more worrying this: > julia> 1 / 10^20000 > Inf I’d rather have an overflow here than Inf, or at least an Infinitesimal value… Maybe something we could ask for future versions of Julia.

[BioTurboNick]

And I’ll plug OverflowContexts.jl (GitHub - BioTurboNick/OverflowContexts.jl: Fine control of integer overflow checking.) if you’d like control over it.

[StevenSiew]

Why not Inf everywhere?

Because Inf is NOT a part of Int64

[dmetivie]

Thanks for all your answers, I missed these link in my pre-question search.
Plus I missed Julia docs on that.

I missed the “overflow” keyword.