This page was generated from unit-2.1.4-bddc/bddc.ipynb.

2.1.4 Element-wise BDDC Preconditioner

The element-wise BDDC (Balancing Domain Decomposition preconditioner with Constraints) preconditioner in NGSolve is a good general purpose preconditioner that works well both in the shared memory parallel mode as well as in distributed memory mode. In this tutorial, we discuss this preconditioner, related built-in options, and customization from python.

Let us start with a simple description of the element-wise BDDC in the context of Lagrange finite element space \(V\). Here the BDDC preconditioner is constructed on an auxiliary space \(\widetilde{V}\) obtained by connecting only element vertices (leaving edge and face shape functions discontinuous). Although larger, the auxiliary space allows local elimination of edge and face variables. Hence an analogue of the original matrix \(A\) on this space, named \(\widetilde A\), is less expensive to invert. This inverse is used to construct a preconditioner for \(A\) as follows:

\[C_{BDDC}^{-1} = R {\,\widetilde{A}\,}^{-1}\, R^t\]

Here, \(R\) is the averaging operator for the discontinuous edge and face variables.

To construct a general purpose BDDC preconditioner, NGSolve generalizes this idea to all its finite element spaces by a classification of degrees of freedom. Dofs are classified into (condensable) LOCAL_DOFs that we saw in 1.4 and a remainder that includes these types:

WIREBASKET_DOF
INTERFACE_DOF

The original finite element space \(V\) is obtained by requiring conformity of both the above types of dofs, while the auxiliary space \(\widetilde{V}\) is obtained by requiring conformity of WIREBASKET_DOFs only.

Here is a figure of a typical function in the default \(\widetilde{V}\) (and the code to generate this is at the end of this tutorial) when \(V\) is the Lagrange space:

title

[1]:

from ngsolve import *
from ngsolve.webgui import Draw
from ngsolve.la import EigenValues_Preconditioner
SetHeapSize(100*1000*1000)

[2]:

mesh = Mesh(unit_cube.GenerateMesh(maxh=0.5))
# mesh = Mesh(unit_square.GenerateMesh(maxh=0.5))

Built-in options

Let us define a simple function to study how the spectrum of the preconditioned matrix changes with various options.

Effect of condensation

[3]:

def TestPreconditioner (p, condense=False, **args):
    fes = H1(mesh, order=p, **args)
    u,v = fes.TnT()
    a = BilinearForm(fes, eliminate_internal=condense)
    a += grad(u)*grad(v)*dx + u*v*dx
    c = Preconditioner(a, "bddc")
    a.Assemble()
    return EigenValues_Preconditioner(a.mat, c.mat)

[4]:

lams = TestPreconditioner(5)
print (lams[0:3], "...\n", lams[-3:])

 1.00054
 1.03405
 1.09891
 ...
  4.48943
 4.60085
 4.79194

Here is the effect of static condensation on the BDDC preconditioner.

[5]:

lams = TestPreconditioner(5, condense=True)
print (lams[0:3], "...\n", lams[-3:])

 1.00023
 1.02753
 1.08363
 ...
  4.06546
 4.12581
 4.22957

Tuning the auxiliary space

Next, let us study the effect of a few built-in flags for finite element spaces that are useful for tweaking the behavior of the BDDC preconditioner. The effect of these flags varies in two (2D) and three dimensions (3D), e.g.,

  • wb_fulledges=True: This option keeps all edge-dofs conforming (i.e. they are marked WIREBASKET_DOFs). This option is only meaningful in 3D. If used in 2D, the preconditioner becomes a direct solver.

  • wb_withedges=True: This option keeps only the first edge-dof conforming (i.e., the first edge-dof is marked WIREBASKET_DOF and the remaining edge-dofs are marked INTERFACE_DOFs).

The complete situation is a bit more complex due to the fact these options can take the three values True, False, or Undefined, the two options can be combined, and the space dimension can be 2 or 3. The default value of these flags in NGSolve is Undefined. Here is a table with the summary of the effect of these options:

wb_fulledges

wb_withedges

2D

3D

True

any value

all

all

False/Undefined

Undefined

none

first

False/Undefined

False

none

none

False/Undefined

True

first

first

An entry \(X \in\) {all, none, first} of the last two columns is to be read as follows: \(X\) of the edge-dofs is(are) WIREBASKET_DOF(s).

Here is an example of the effect of one of these flag values.

[6]:

lams = TestPreconditioner(5, condense=True,
                          wb_withedges=False)
print (lams[0:3], "...\n", lams[-3:])

 1.00106
 1.07034
 1.22031
 ...
  25.4058
  25.406
 26.9095

Clearly, when moving from the default case (where the first of the edge dofs are wire basket dofs) to the case (where none of the edge dofs are wire basket dofs), the conditioning became less favorable.

Customize

From within python, we can change the types of degrees of freedom of finite element spaces, thus affecting the behavior of the BDDC preconditioner.

To depart from the default and commit the first two edge dofs to wire basket, we perform the next steps:

[7]:

fes = H1(mesh, order=10)
u,v = fes.TnT()

for ed in mesh.edges:
    dofs = fes.GetDofNrs(ed)
    for d in dofs:
        fes.SetCouplingType(d, COUPLING_TYPE.INTERFACE_DOF)

    # Set the first two edge dofs to be conforming
    fes.SetCouplingType(dofs[0], COUPLING_TYPE.WIREBASKET_DOF)
    fes.SetCouplingType(dofs[1], COUPLING_TYPE.WIREBASKET_DOF)

a = BilinearForm(fes, eliminate_internal=True)
a += grad(u)*grad(v)*dx + u*v*dx
c = Preconditioner(a, "bddc")
a.Assemble()

lams=EigenValues_Preconditioner(a.mat, c.mat)
max(lams)/min(lams)

[7]:

9.989385163106014

This is a slight improvement from the default.

[8]:

lams = TestPreconditioner (10, condense=True)
max(lams)/min(lams)

[8]:

13.97495834396635

Combine BDDC with AMG for large problems

The coarse inverse \({\,\widetilde{A}\,}^{-1}\) of BDDC is expensive on fine meshes. Using the option coarsetype=h1amg flag, we can ask BDDC to replace \({\,\widetilde{A}\,}^{-1}\) by an Algebraic MultiGrid (AMG) preconditioner. Since NGSolve’s h1amg is well-suited
for the lowest order Lagrange space, we use the option wb_withedges=False to ensure that \(\widetilde{A}\) is made solely with vertex unknowns.
[9]:

p = 5
mesh = Mesh(unit_cube.GenerateMesh(maxh=0.05))
fes = H1(mesh, order=p, dirichlet="left|bottom|back",
         wb_withedges=False)
print("NDOF = ", fes.ndof)
u,v = fes.TnT()
a = BilinearForm(fes)
a += grad(u)*grad(v)*dx
f = LinearForm(fes)
f += v*dx

with TaskManager():
    pre = Preconditioner(a, "bddc", coarsetype="h1amg")
    a.Assemble()
    f.Assemble()

    gfu = GridFunction(fes)
    solvers.CG(mat=a.mat, rhs=f.vec, sol=gfu.vec,
               pre=pre, maxsteps=500)
Draw(gfu)

NDOF =  839901

WARNING: kwarg 'coarsetype' is an undocumented flags option for class <class 'ngsolve.comp.Preconditioner'>, maybe there is a typo?

CG iteration 1, residual = 0.7217953115360729
CG iteration 2, residual = 0.3050248203792491
CG iteration 3, residual = 0.2749055243142275
CG iteration 4, residual = 0.3254836128333248
CG iteration 5, residual = 0.3197726120226225
CG iteration 6, residual = 0.24005556840618275
CG iteration 7, residual = 0.18373566498596774
CG iteration 8, residual = 0.14887596601780836
CG iteration 9, residual = 0.11776205172130383
CG iteration 10, residual = 0.09415746715001225
CG iteration 11, residual = 0.07003619211581447
CG iteration 12, residual = 0.053374347102851716
CG iteration 13, residual = 0.044173909848494355
CG iteration 14, residual = 0.03617025551380328
CG iteration 15, residual = 0.027632823230957906
CG iteration 16, residual = 0.021697513734227663
CG iteration 17, residual = 0.01813717835960824
CG iteration 18, residual = 0.014942565945605515
CG iteration 19, residual = 0.01186786990408173
CG iteration 20, residual = 0.008882592897356066
CG iteration 21, residual = 0.006862801077328545
CG iteration 22, residual = 0.005164356408633748
CG iteration 23, residual = 0.003953665409810014
CG iteration 24, residual = 0.003036117884227966
CG iteration 25, residual = 0.00238248062693759
CG iteration 26, residual = 0.001890013420182118
CG iteration 27, residual = 0.0015173967401730392
CG iteration 28, residual = 0.0011513838097701222
CG iteration 29, residual = 0.0009015669397151688
CG iteration 30, residual = 0.0007164959640586643
CG iteration 31, residual = 0.0005663664655473003
CG iteration 32, residual = 0.0004327728364206818
CG iteration 33, residual = 0.00033534779198470834
CG iteration 34, residual = 0.00026236876368009856
CG iteration 35, residual = 0.0002025651066229893
CG iteration 36, residual = 0.00016357418064306012
CG iteration 37, residual = 0.00013001522985554227
CG iteration 38, residual = 9.827935845585473e-05
CG iteration 39, residual = 7.620034796382275e-05
CG iteration 40, residual = 6.005682339731345e-05
CG iteration 41, residual = 4.736199375008367e-05
CG iteration 42, residual = 3.706921239205016e-05
CG iteration 43, residual = 2.877010949602475e-05
CG iteration 44, residual = 2.2259449023109942e-05
CG iteration 45, residual = 1.6972553712084623e-05
CG iteration 46, residual = 1.3339825547311691e-05
CG iteration 47, residual = 1.0603308347311004e-05
CG iteration 48, residual = 8.35816042307099e-06
CG iteration 49, residual = 6.421504202703166e-06
CG iteration 50, residual = 5.001730912146577e-06
CG iteration 51, residual = 3.922675090427644e-06
CG iteration 52, residual = 3.239398380170967e-06
CG iteration 53, residual = 2.8865921042922155e-06
CG iteration 54, residual = 2.11139217370341e-06
CG iteration 55, residual = 1.660434237260289e-06
CG iteration 56, residual = 1.2837196402791172e-06
CG iteration 57, residual = 9.705364915010282e-07
CG iteration 58, residual = 7.565091660741372e-07
CG iteration 59, residual = 5.909277634067599e-07
CG iteration 60, residual = 4.717085375806342e-07
CG iteration 61, residual = 4.2041300288445e-07
CG iteration 62, residual = 3.2417023395452897e-07
CG iteration 63, residual = 2.430796138321275e-07
CG iteration 64, residual = 1.866206646937136e-07
CG iteration 65, residual = 1.4665283693813282e-07
CG iteration 66, residual = 1.1723327301542398e-07
CG iteration 67, residual = 9.843535219079966e-08
CG iteration 68, residual = 8.293912434191132e-08
CG iteration 69, residual = 6.318005133936699e-08
CG iteration 70, residual = 4.846516756177291e-08
CG iteration 71, residual = 3.7051437698781584e-08
CG iteration 72, residual = 2.8281596986246695e-08
CG iteration 73, residual = 2.1749572038129366e-08
CG iteration 74, residual = 1.6588024223306982e-08
CG iteration 75, residual = 1.2866322271714595e-08
CG iteration 76, residual = 9.922235948995858e-09
CG iteration 77, residual = 7.682984251443676e-09
CG iteration 78, residual = 5.870522233328118e-09
CG iteration 79, residual = 4.587412991398981e-09
CG iteration 80, residual = 3.669305692941072e-09
CG iteration 81, residual = 3.0626387014976424e-09
CG iteration 82, residual = 2.4601522826905715e-09
CG iteration 83, residual = 1.8406976726561403e-09
CG iteration 84, residual = 1.4392029087228812e-09
CG iteration 85, residual = 1.088031028483148e-09
CG iteration 86, residual = 8.265601822417413e-10
CG iteration 87, residual = 6.49033952730564e-10
CG iteration 88, residual = 5.013809981236942e-10
CG iteration 89, residual = 3.88003825508102e-10
CG iteration 90, residual = 2.9673548346998894e-10
CG iteration 91, residual = 2.259723829304555e-10
CG iteration 92, residual = 1.7641650835799896e-10
CG iteration 93, residual = 1.3955796054934607e-10
CG iteration 94, residual = 1.197890873260174e-10
CG iteration 95, residual = 1.032910487143179e-10
CG iteration 96, residual = 7.353962142799821e-11
CG iteration 97, residual = 5.552665085251104e-11
CG iteration 98, residual = 4.313561623126917e-11
CG iteration 99, residual = 3.3235116420943526e-11
CG iteration 100, residual = 2.5425417870649966e-11
CG iteration 101, residual = 1.9359326725203247e-11
CG iteration 102, residual = 1.4969883425390265e-11
CG iteration 103, residual = 1.2049004679655104e-11
CG iteration 104, residual = 1.0230191323655324e-11
CG iteration 105, residual = 8.277004499287447e-12
CG iteration 106, residual = 6.6584306648569545e-12
CG iteration 107, residual = 5.293919250189958e-12
CG iteration 108, residual = 4.1350545363226086e-12
CG iteration 109, residual = 3.07966250720792e-12
CG iteration 110, residual = 2.356992328063745e-12
CG iteration 111, residual = 1.826593031930661e-12
CG iteration 112, residual = 1.4243602744778434e-12
CG iteration 113, residual = 1.1073733199470907e-12
CG iteration 114, residual = 9.789750575047045e-13
CG iteration 115, residual = 7.966088440099294e-13
CG iteration 116, residual = 5.836736676325778e-13

[9]:

BaseWebGuiScene

Postscript

By popular demand, here is the code to draw the figure found at the beginning of this tutorial:

[10]:

from netgen.geom2d import unit_square
mesh = Mesh(unit_square.GenerateMesh(maxh=0.1))
fes_ho = Discontinuous(H1(mesh, order=10))
fes_lo = H1(mesh, order=1, dirichlet=".*")
fes_lam = Discontinuous(H1(mesh, order=1))
fes = fes_ho*fes_lo*fes_lam
uho, ulo, lam = fes.TrialFunction()

a = BilinearForm(fes)
a += Variation(0.5 * grad(uho)*grad(uho)*dx
               - 1*uho*dx
               + (uho-ulo)*lam*dx(element_vb=BBND))
gfu = GridFunction(fes)
solvers.Newton(a=a, u=gfu)
Draw(gfu.components[0],deformation=True, settings={"camera": {"transformations": [{"type": "rotateX", "angle": -45}]}})

Newton iteration  0
err =  0.39346141669994983
Newton iteration  1
err =  1.8488788182878563e-15

[10]:

BaseWebGuiScene