How to apply a continuous-changing parameter in a Poisson problem using PartionedArrays and Gridap

[ZiwenH]

Hey guys,

I'd like to solve a problem with the bilinear form like following:
a(u,v) = ∫((vec)⋅∇(u)⋅∇(v))dΩ
where vec is of Vector{Float64} type. But in parallel mode, it will lead to the error:
ERROR: Scalar indexing on PVector is not allowed for performance reasons.
There is no similar case in the tutorial, does anybody have any idea or experience how to treat the parameter?

Best regards

[fverdugo]

Hi @ZiwenH,

This question is more related with GridapDistributed than with PartitionedArrays.

In any case, you first need to wrap your vector into a CellField object

f = CellField(partition(vec),Ω)
a(u,v) = ∫(f⋅∇(u)⋅∇(v))dΩ

If this does not solve the issue, then ask in https://github.com/gridap/GridapDistributed.jl