The following code, which gives an order of degrees of freedom in the case of periodic boundary conditions, gives different results for dolfin 1.4 and dolfin 1.5, depending on the different value of parameters.reorder_serial_dofs=True/False.
import dolfin as df
df.parameters.reorder_dofs_serial = False
mesh = df.IntervalMesh(4, 1, 5)
class PeriodicBoundary(df.SubDomain):
def inside(self, x, on_boundary):
return bool(x[0] < 1+df.DOLFIN_EPS and x[0] > 1-df.DOLFIN_EPS and on_boundary)
def map(self, x, y):
y[0] = x[0] - 4.0
# Create periodic boundary condition
pbc = PeriodicBoundary()
fspace = df.VectorFunctionSpace(mesh, 'CG', 1, 3, constrained_domain=pbc)
expression = df.Expression(['x[0]+0.1', 'x[0]+0.2', 'x[0]+0.3'])
f = df.interpolate(expression, fspace)
print f.vector().array()
On dolfin 1.4 with parameters.reorder_serial_dofs=True the order of degrees of freedom is:
[ 5.1 5.2 5.3 2.1 2.2 2.3 3.1 3.2 3.3 4.1 4.2 4.3]
where all vector components at a single node are grouped together, e.g. [mx1, my1, mz1, mx2, my2, mz2, mx3, my3, mz3, ….]. In this example, the first number is the mesh node and the second is the component of a three-dimensional vector field (for instance, 3.2 is the y components at mesh node 3 - my3). On the other hand, with parameters.reorder_serial_dofs=False, all values of a single vector component at all mesh nodes are grouped together like [mx1, mx2, mx3, … my1, my2, my3, … mz1, mz2, mz3, …]:
[ 5.1 2.1 3.1 4.1 5.2 2.2 3.2 4.2 5.3 2.3 3.3 4.3]
However, with dolfin 1.5, this order with parameters.reorder_serial_dofs=False is:
[5.1, 2.1, 5.2, 2.2, 5.3, 2.3, 3.1, 3.2, 3.3, 4.1, 4.2, 4.3]
which is a bit surprising. Is this the dolfin bug or within expected behaviour of dolfin 1.5?
