ApproxFun causes Jupyter to hang


#1

Hey everyone, I’m currently trying to use ApproxFun in a Jupyter notebook, but the cell containing Using ApproxFun never terminates. I can load ApproxFun fine (albeit a bit slowly) in the terminal

Is this a known issue? Is there a known workaround?

Thanks in advance!


#2

Here is what I get with julia 1.0 on windows 10.

julia> @time using ApproxFun
30.295569 seconds (61.13 M allocations: 4.873 GiB, 3.71% gc time)

In Jupyter notebook

@time using ApproxFun
29.376509 seconds (61.04 M allocations: 4.867 GiB, 3.57% gc time)

I think the issue here is why so slow to load and huge allocations.


#3

Are those 4.867 GiB allocated all at once? My computer has only 4…


#4

No, it is the total allocation during the execution of the command.


#5

There are compiler regressions in Julia v1.0 that cause this large compile time, I believe due to large union types. The issue starts with the pack LowRankApprox.jl, which has many StridedArray overrides.

Hopefully this issue will be resolved in 1.x.


#6

Sure, but they could still be allocated almost all at once or one GiB at a time. They could be retained or freed for all @time tells us