real functionspace giving error in parallel mode(mpirun) using docker image

ramayzoraiz commented 1 year ago

Expected Behavior

print("oh no! it worked")

Actual Behavior

Error at line: cc=df.FunctionSpace(mesh, "R", 0)

Error: Unable to re-order graph using SCOTCH. Reason: Error during re-ordering. Where: This error was encountered inside SCOTCH.cpp. *** Process: 2

DOLFIN version: 2019.2.0.dev0 Git changeset: 8060a84e8ada0df018772497534092727b06bd56

Steps to Reproduce the Problem

Create python script "xyz.py" and paste folllowing lines:

import dolfin as df
comm=df.MPI.comm_world
mesh=df.UnitCubeMesh(comm,5,5,5)
cc=df.FunctionSpace(mesh, "R", 0)
activation = df.Function(cc)
activation.assign(df.Constant(0.2))
print("oh no! it worked")

In terminal, type mpirun -n 4 python3 xyz.py

Specifications

Failed on two docker images: ghcr.io/scientificcomputing/fenics-gmsh:jupterlab-2023-04-21 ghcr.io/computationalphysiology/simcardems:latest

Platform: Docker

finsberg commented 1 year ago

Thanks for reporting this!

You could even reduce it to the following example

import dolfin as df

comm = df.MPI.comm_world
mesh = df.UnitCubeMesh(comm, 5, 5, 5)
cc = df.FunctionSpace(mesh, "R", 0)

This is not a bug in simcardems, but in dolfin, so you should open an issue there instead: https://bitbucket.org/fenics-project/dolfin/issues

jorgensd commented 1 year ago

This has been fixed. use ghcr.io/scientificcomputing/fenics:2023-08-14 or https://github.com/scientificcomputing/packages/pkgs/container/fenics-gmsh/119060949?tag=2023-08-16

ComputationalPhysiology / simcardems