ApproxFun.jl
ApproxFun.jl copied to clipboard
Support pre-allocated Fun construction
The idea is to support, for example:
f = Fun(Chebyshev(), Array{Float64}(1_000_000))
x = Fun()
f .= cos.(1_000_000.*x) # no allocation
this is useful for time-stepping, for example.
To make this work, we need:
[ ] Non-allocating points #531
[ ] In-place transforms (plan_transform!
)
[ ] A good syntax for doing this from an anonymous function. Something like
Fun!(f, x-> cos(1_000_000x))