June 2021 - firedrake - imperial.ac.uk

Re: [firedrake] Understanding solver parameters in NonlinearVariationalSolver
by Bhavesh Shrimali 29 Jun '21

29 Jun '21

I think I mistakenly clicked reply instead of reply all. Attached is my response: -- Thanks Matt, Thanks for the swift response. These are the logs with basic, backtracking, cp and nleqerr. If I don't specify 'snes_linesearch_type' it defaults to 'basic'. At least that's what I could gather from the log. - `cp` seemingly diverges the fastest. Of course I haven't touched the other options. - `nleqerr` goes the furthest I believe, but not that far from `basic` or `bt`. My concern was that since the forms are essentially the same, and I am passing "seemingly" the same parameters in FEniCS and Firedrake (both of which are using PETSc's SNES), what is it that is different which causes the solver to fail in one but not the other. Edit: Just saw Ed's message. Indeed it defaults to `basic` For definiteness this was the dictionary of solver parameters I am passing to NLVS in Firedrake solver_params={ "snes_type": "newtonls", "ksp_type": "preonly", "pc_type": "lu", "snes_max_it": 100, # "snes_stol": 1.e-6, # "snes_rtol": 1.e-6, # "snes_atol": 1.e-6, "snes_view": None, # "ksp_atol": 1.e-10, # "ksp_rtol":1.e-8, "ksp_view": None, "snes_linesearch_type": "bt", "snes_converged_reason": None, "snes_monitor": None, "ksp_converged_reason": None, "ksp_monitor_true_residual": None } On Sun, 27 Jun 2021 at 11:22, Bhavesh Shrimali <bhavesh.shrimali(a)gmail.com> wrote: > Thanks Matt, > > > Thanks for the swift response. > These are the logs with basic, backtracking, cp and nleqerr. If I don't > specify 'snes_linesearch_type' it defaults to 'basic'. At least that's what > I could gather from the log. > > > - `cp` seemingly diverges the fastest. Of course I haven't touched the > other options. > - `nleqerr` goes the furthest I believe, but not that far from `basic` > or `bt`. > > My concern was that since the forms are essentially the same, and I am > passing "seemingly" the same parameters in FEniCS and Firedrake (both of > which are using PETSc's SNES), what is it that is different which causes > the solver to fail in one but not the other. > > Edit: Just saw Ed's message. Indeed it defaults to `basic` > > > For definiteness this was the dictionary of solver parameters I am passing > to NLVS in Firedrake > > solver_params={ > "snes_type": "newtonls", > "ksp_type": "preonly", > "pc_type": "lu", > "snes_max_it": 100, > # "snes_stol": 1.e-6, > # "snes_rtol": 1.e-6, > # "snes_atol": 1.e-6, > "snes_view": None, > # "ksp_atol": 1.e-10, > # "ksp_rtol":1.e-8, > "ksp_view": None, > "snes_linesearch_type": "bt", > "snes_converged_reason": None, > "snes_monitor": None, > "ksp_converged_reason": None, > "ksp_monitor_true_residual": None > } > > > On Sun, 27 Jun 2021 at 06:00, Matthew Knepley <knepley(a)gmail.com> wrote: > >> Give >> >> -snes_view -snes_converged_reason -snes_monitor -ksp_converged_reason >> -ksp_monitor_true_residual >> >> and send all the output. That way I might be able to see what is >> happening when you do not converge. Also, >> line search 'basic' is not a good choice usually. I would either use the >> default, or 'cp', or maybe 'nleqerr' that >> Patrick added. >> >> Thanks, >> >> Matt >> >> On Sat, Jun 26, 2021 at 10:14 PM Bhavesh Shrimali < >> bhavesh.shrimali(a)gmail.com> wrote: >> >>> Hi everyone, >>> >>> I am trying to understand the usage of solver parameters that are passed >>> to the NonlinearVariationalSolver in firedrake. In order to do some tests, >>> I decided to reproduce some of my previous calculations done in FEniCS for >>> nonlinear viscoelasticity. In old dolfin, I had the following for solver >>> parameters to be passed to PETSc's SNES: >>> >>> snes_solver_parameters = {"nonlinear_solver": "snes", >>> "snes_solver": {"linear_solver": "lu", >>> "maximum_iterations": 50, >>> "report": True, >>> "line_search": 'basic', >>> "error_on_nonconvergence": False, >>> "relative_tolerance": 1e-7, >>> "absolute_tolerance": 1e-8}} >>> ... >>> ... >>> solver = NonlinearVariationalSolver(problem) >>> solver.parameters.update(snes_solver_parameters) >>> >>> Now as I understand, in Firedrake the options passed to the solver are >>> consistent with the corresponding names in PETSc (correct me if I am >>> wrong), and therefore I did: >>> >>> solver_params={ >>> "snes_type": "newtonls", >>> "ksp_type": "preonly", >>> "pc_type": "lu", >>> "snes_max_it": 100, >>> "snes_rtol": 1.e-7, >>> "snes_atol": 1.e-8, >>> "snes_view": None, >>> "ksp_atol": 1.e-10, # not needed? >>> "ksp_rtol":1.e-8, >>> "ksp_view": None, >>> "snes_linesearch_type": "basic", >>> "snes_no_convergence_test": 1 # this has no effect >>> } >>> solver = NonlinearVariationalSolver(problem, >>> solver_parameters=solver_params) >>> >>> However, for some reason, I am unable to get the solver to converge when >>> trying to solve the exact same problem as FEniCS. I experimented with >>> different quadrature degrees as well as tweaked the solver parameters for >>> SNES to best match what I am doing in dolfin. I understand that the form >>> compiler is different hence the precise code that is generated may be >>> different (I am not an expert on code generation) even though the python >>> scripts may look similar due to UFL syntax. >>> >>> In fact the weird part is that the solver seems to converge fine till a >>> certain point in my input (loading). You can see this in the plot below >>> where I plot the final stress-stretch relation obtained at a point in my >>> domain from dolfin and firedrake.. >>> >>> [image: firedrake_vs_fenics.png] >>> >>> >>> I have marked the abscissa beyond which the solver fails to converge in >>> Firedrake. The complete codes are here: >>> https://github.com/bhaveshshrimali/comparingFiredrakeFEniCS >>> >>> I tried setting the option `snes_no_convergence_test` to 1/True but >>> that had no effect as the SNES solver still diverges with `DIVERGED_DTOL`. >>> In the end I put the call to solve inside a try/catch block to at least be >>> able to see the solution till that point. And it surprisingly is accurate >>> till that point. >>> >>> Would anyone have any clue of what could be going wrong? I wouldn't >>> expect that you go through the entire code(s) as that could be fairly long, >>> but any pointers would be helpful :) >>> >>> I am using firedrake via Docker on windows. >>> >>> Thanks >>> >>> -- >>> >>> _______________________________________________ >>> firedrake mailing list >>> firedrake(a)imperial.ac.uk >>> https://mailman.ic.ac.uk/mailman/listinfo/firedrake >>> >> >> >> -- >> What most experimenters take for granted before they begin their >> experiments is infinitely more interesting than any results to which their >> experiments lead. >> -- Norbert Wiener >> >> https://www.cse.buffalo.edu/~knepley/ >> <http://www.cse.buffalo.edu/~knepley/> >> > > > -- > > *Bhavesh Shrimali* > *Ph.D. Student, **Civil and Environmental Engineering,* > > *University of Illinois Urbana-Champaign* > > > *Champaign, IL, United States* > > *Email : bshrima2(a)illinois.edu <bshrima2(a)illinois.edu>* > *Web : **https://bhaveshshrimali.github.io/ > <https://bhaveshshrimali.github.io/>* > -- *Bhavesh Shrimali* *Ph.D. Student, **Civil and Environmental Engineering,* *University of Illinois Urbana-Champaign* *Champaign, IL, United States* *Email : bshrima2(a)illinois.edu <bshrima2(a)illinois.edu>* *Web : **https://bhaveshshrimali.github.io/ <https://bhaveshshrimali.github.io/>*

2 6

Understanding solver parameters in NonlinearVariationalSolver
by Bhavesh Shrimali 27 Jun '21

27 Jun '21

Hi everyone, I am trying to understand the usage of solver parameters that are passed to the NonlinearVariationalSolver in firedrake. In order to do some tests, I decided to reproduce some of my previous calculations done in FEniCS for nonlinear viscoelasticity. In old dolfin, I had the following for solver parameters to be passed to PETSc's SNES: snes_solver_parameters = {"nonlinear_solver": "snes", "snes_solver": {"linear_solver": "lu", "maximum_iterations": 50, "report": True, "line_search": 'basic', "error_on_nonconvergence": False, "relative_tolerance": 1e-7, "absolute_tolerance": 1e-8}} ... ... solver = NonlinearVariationalSolver(problem) solver.parameters.update(snes_solver_parameters) Now as I understand, in Firedrake the options passed to the solver are consistent with the corresponding names in PETSc (correct me if I am wrong), and therefore I did: solver_params={ "snes_type": "newtonls", "ksp_type": "preonly", "pc_type": "lu", "snes_max_it": 100, "snes_rtol": 1.e-7, "snes_atol": 1.e-8, "snes_view": None, "ksp_atol": 1.e-10, # not needed? "ksp_rtol":1.e-8, "ksp_view": None, "snes_linesearch_type": "basic", "snes_no_convergence_test": 1 # this has no effect } solver = NonlinearVariationalSolver(problem, solver_parameters=solver_params) However, for some reason, I am unable to get the solver to converge when trying to solve the exact same problem as FEniCS. I experimented with different quadrature degrees as well as tweaked the solver parameters for SNES to best match what I am doing in dolfin. I understand that the form compiler is different hence the precise code that is generated may be different (I am not an expert on code generation) even though the python scripts may look similar due to UFL syntax. In fact the weird part is that the solver seems to converge fine till a certain point in my input (loading). You can see this in the plot below where I plot the final stress-stretch relation obtained at a point in my domain from dolfin and firedrake.. [image: firedrake_vs_fenics.png] I have marked the abscissa beyond which the solver fails to converge in Firedrake. The complete codes are here: https://github.com/bhaveshshrimali/comparingFiredrakeFEniCS I tried setting the option `snes_no_convergence_test` to 1/True but that had no effect as the SNES solver still diverges with `DIVERGED_DTOL`. In the end I put the call to solve inside a try/catch block to at least be able to see the solution till that point. And it surprisingly is accurate till that point. Would anyone have any clue of what could be going wrong? I wouldn't expect that you go through the entire code(s) as that could be fairly long, but any pointers would be helpful :) I am using firedrake via Docker on windows. Thanks --

3 2

Re: [firedrake] firedrake Digest, Vol 53, Issue 4
by Bhavesh Shrimali 25 Jun '21

25 Jun '21

Thank you for the swift response Prof. Farrell. Indeed, I sloppily read through the spaces and changing the internal bubble to degree 4 no longer gives an error. However the resulting system cannot be solved by PETSc. That is, changing the code to: P2 = FiniteElement("CG", "tetrahedron", 2) B4 = FiniteElement("B", "tetrahedron", 4) FB3 = FiniteElement("FacetBubble", "tetrahedron", 3) Vh = VectorFunctionSpace(mesh, P2 + B4 + FB3) Qh = FunctionSpace(mesh, "DG", 1) gives the following traceback: File "/home/firedrake/firedrake/src/firedrake/firedrake/adjoint/variational_solver.py", line 75, in wrapper out = solve(self, **kwargs) File "/home/firedrake/firedrake/src/firedrake/firedrake/variational_solver.py", line 281, in solve solving_utils.check_snes_convergence(self.snes) File "/home/firedrake/firedrake/src/firedrake/firedrake/solving_utils.py", line 140, in check_snes_convergence raise ConvergenceError(r"""Nonlinear solve failed to converge after %d nonlinear iterations. firedrake.exceptions.ConvergenceError: Nonlinear solve failed to converge after 3 nonlinear iterations. Reason: DIVERGED_DTOL I would expect that for larger problems this could be an issue, but for a small test mesh such as this (UnitCubeMesh(5,5,5)) it shouldn't be, unless of course I am making some grave error. I am attaching a complete log with the monitors in case if someone wants to take a look and maybe spot what could be going wrong. On Fri, 25 Jun 2021 at 06:00, <firedrake-request(a)imperial.ac.uk> wrote: > Send firedrake mailing list submissions to > firedrake(a)imperial.ac.uk > > To subscribe or unsubscribe via the World Wide Web, visit > https://mailman.ic.ac.uk/mailman/listinfo/firedrake > or, via email, send a message with subject or body 'help' to > firedrake-request(a)imperial.ac.uk > > You can reach the person managing the list at > firedrake-owner(a)imperial.ac.uk > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of firedrake digest..." > > > Today's Topics: > > 1. Mixed formulation for incompressible hyperelasticity > (Bhavesh Shrimali) > 2. Re: Mixed formulation for incompressible hyperelasticity > (Bhavesh Shrimali) > 3. Re: Mixed formulation for incompressible hyperelasticity > (Patrick Farrell) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Thu, 24 Jun 2021 22:11:25 -0500 > From: Bhavesh Shrimali <bhavesh.shrimali(a)gmail.com> > To: firedrake(a)imperial.ac.uk > Subject: [firedrake] Mixed formulation for incompressible > hyperelasticity > Message-ID: > < > CANw4EA2dOXvzvGtWvphFHBA6azcD9yunrhu_Zad96fdT8rrJGg(a)mail.gmail.com> > Content-Type: text/plain; charset="utf-8" > > Hi everyone, > > I was trying to use a mixed-formulation for an incompressible > hyperelasticity problem. See the following MWE which uses a Taylor-Hood > approximation -- P2 for displacement (u) and P1 for pressure (p): > > mesh = UnitCubeMesh(5,5,5) > Vh = VectorFunctionSpace(mesh, "CG", 2) > Qh = FunctionSpace(mesh, "CG", 1) > Wh = Vho * Qh > w = Function(Wh) > dw = TrialFunction(Wh) > v = TestFunction(Wh) > u, p = split(w) > mu, lmbda = Constant(1.), Constant(1.e3) > Js = (lmbda + p + sqrt((lmbda + p)**2. + 4.*lmbda*mu ))/(2.*lmbda) > F = Identity(mesh.geometric_dimension()) + grad(u) > J = det(F) > I1 = inner(F, F) > Pi = p * (J - Js) + mu/2 * (I1 - 3) - mu * ln(Js) + lmbda/2 * (Js - 1)**2 > Pi *= dx > a = derivative(Pi, w, v) > Jac = derivative(a, w, dw) > bcs = [ > DirichletBC(Wh.sub(0).sub(0), Constant(0.), 1), > DirichletBC(Wh.sub(0).sub(1), Constant(0.), 3), > DirichletBC(Wh.sub(0).sub(2), Constant(0.), 5), > DirichletBC(Wh.sub(0).sub(0), Constant(1.), 2) > ] > problem = NonlinearVariationalProblem(a, w, bcs, J=Jac) > solver = NonlinearVariationalSolver(problem) > solver.solve() > u, p = w.split() > File("displacement.pvd").write(u) > > > Everything works fine as expected. Now, I wanted to try another pair of > approximations that is inf-sup stable, namely the one in the celebrated > paper by Crouzeix-Raviart (1973). See (4.36) in > http://archive.numdam.org/article/M2AN_1973__7_3_33_0.pdf. > > This amounts to approximating `u` by a "P2 + Bubble" and `p` by a "DG-1" > space. This is also given in (3.2.3) in > https://link.springer.com/content/pdf/10.1007%2F978-3-642-87047-7_5.pdf > and (2.38) in > https://link.springer.com/content/pdf/10.1007%2F978-3-642-61623-5_2.pdf > > Now, in 3D as far as I can tell this amounds to approximating `u` with > "P2 + an internal bubble + FacetBubble". Would anyone know how this could > be achieved within firedrake? Because as far as I can see, firedrake > supports `FacetBubble`. However, when I try something like: > > B3 = FiniteElement("Bubble", "tetrahedron", 3) > FB3 = FiniteElement("FacetBubble", "tetrahedron", 3) > P2 = FiniteElement("CG", "tetrahedron", 2) > Vh = FunctionSpace(mesh, P2 + B3 + FB3) > > I get the following error > > File "/usr/lib/python3.8/functools.py", line 875, in wrapper > return dispatch(args[0].__class__)(*args, **kw) > File "/home/firedrake/firedrake/src/tsfc/tsfc/finatinterface.py", line > 189, in convert_enrichedelement > elements, deps = zip(*[_create_element(elem, **kwargs) > File "/home/firedrake/firedrake/src/tsfc/tsfc/finatinterface.py", line > 189, in <listcomp> > elements, deps = zip(*[_create_element(elem, **kwargs) > File "/home/firedrake/firedrake/src/tsfc/tsfc/finatinterface.py", line > 310, in _create_element > finat_element, deps = convert(ufl_element, **kwargs) > File "/usr/lib/python3.8/functools.py", line 875, in wrapper > return dispatch(args[0].__class__)(*args, **kw) > File "/home/firedrake/firedrake/src/tsfc/tsfc/finatinterface.py", line > 177, in convert_finiteelement > return lmbda(cell, element.degree()), set() > File "/home/firedrake/firedrake/src/FInAT/finat/fiat_elements.py", line > 318, in __init__ > super(Bubble, self).__init__(FIAT.Bubble(cell, degree)) > File "/home/firedrake/firedrake/src/fiat/FIAT/bubble.py", line 33, in > __init__ > super(Bubble, self).__init__(ref_el, degree, codim=0) > File "/home/firedrake/firedrake/src/fiat/FIAT/bubble.py", line 24, in > __init__ > raise RuntimeError('Bubble element of degree %d and codimension %d has > no dofs' % (degree, codim)) > RuntimeError: Bubble element of degree 3 and codimension 0 has no dofs > > Any pointers would be appreciated > -- > > *Bhavesh Shrimali* > *Ph.D. Student, **Civil and Environmental Engineering,* > > *University of Illinois Urbana-Champaign* > > > *Champaign, IL, United States* > > *Email : bshrima2(a)illinois.edu <bshrima2(a)illinois.edu>* > *Web : **https://bhaveshshrimali.github.io/ > <https://bhaveshshrimali.github.io/>* >

2 3

Mixed formulation for incompressible hyperelasticity
by Bhavesh Shrimali 25 Jun '21

25 Jun '21

Hi everyone, I was trying to use a mixed-formulation for an incompressible hyperelasticity problem. See the following MWE which uses a Taylor-Hood approximation -- P2 for displacement (u) and P1 for pressure (p): mesh = UnitCubeMesh(5,5,5) Vh = VectorFunctionSpace(mesh, "CG", 2) Qh = FunctionSpace(mesh, "CG", 1) Wh = Vho * Qh w = Function(Wh) dw = TrialFunction(Wh) v = TestFunction(Wh) u, p = split(w) mu, lmbda = Constant(1.), Constant(1.e3) Js = (lmbda + p + sqrt((lmbda + p)**2. + 4.*lmbda*mu ))/(2.*lmbda) F = Identity(mesh.geometric_dimension()) + grad(u) J = det(F) I1 = inner(F, F) Pi = p * (J - Js) + mu/2 * (I1 - 3) - mu * ln(Js) + lmbda/2 * (Js - 1)**2 Pi *= dx a = derivative(Pi, w, v) Jac = derivative(a, w, dw) bcs = [ DirichletBC(Wh.sub(0).sub(0), Constant(0.), 1), DirichletBC(Wh.sub(0).sub(1), Constant(0.), 3), DirichletBC(Wh.sub(0).sub(2), Constant(0.), 5), DirichletBC(Wh.sub(0).sub(0), Constant(1.), 2) ] problem = NonlinearVariationalProblem(a, w, bcs, J=Jac) solver = NonlinearVariationalSolver(problem) solver.solve() u, p = w.split() File("displacement.pvd").write(u) Everything works fine as expected. Now, I wanted to try another pair of approximations that is inf-sup stable, namely the one in the celebrated paper by Crouzeix-Raviart (1973). See (4.36) in http://archive.numdam.org/article/M2AN_1973__7_3_33_0.pdf. This amounts to approximating `u` by a "P2 + Bubble" and `p` by a "DG-1" space. This is also given in (3.2.3) in https://link.springer.com/content/pdf/10.1007%2F978-3-642-87047-7_5.pdf and (2.38) in https://link.springer.com/content/pdf/10.1007%2F978-3-642-61623-5_2.pdf Now, in 3D as far as I can tell this amounds to approximating `u` with "P2 + an internal bubble + FacetBubble". Would anyone know how this could be achieved within firedrake? Because as far as I can see, firedrake supports `FacetBubble`. However, when I try something like: B3 = FiniteElement("Bubble", "tetrahedron", 3) FB3 = FiniteElement("FacetBubble", "tetrahedron", 3) P2 = FiniteElement("CG", "tetrahedron", 2) Vh = FunctionSpace(mesh, P2 + B3 + FB3) I get the following error File "/usr/lib/python3.8/functools.py", line 875, in wrapper return dispatch(args[0].__class__)(*args, **kw) File "/home/firedrake/firedrake/src/tsfc/tsfc/finatinterface.py", line 189, in convert_enrichedelement elements, deps = zip(*[_create_element(elem, **kwargs) File "/home/firedrake/firedrake/src/tsfc/tsfc/finatinterface.py", line 189, in <listcomp> elements, deps = zip(*[_create_element(elem, **kwargs) File "/home/firedrake/firedrake/src/tsfc/tsfc/finatinterface.py", line 310, in _create_element finat_element, deps = convert(ufl_element, **kwargs) File "/usr/lib/python3.8/functools.py", line 875, in wrapper return dispatch(args[0].__class__)(*args, **kw) File "/home/firedrake/firedrake/src/tsfc/tsfc/finatinterface.py", line 177, in convert_finiteelement return lmbda(cell, element.degree()), set() File "/home/firedrake/firedrake/src/FInAT/finat/fiat_elements.py", line 318, in __init__ super(Bubble, self).__init__(FIAT.Bubble(cell, degree)) File "/home/firedrake/firedrake/src/fiat/FIAT/bubble.py", line 33, in __init__ super(Bubble, self).__init__(ref_el, degree, codim=0) File "/home/firedrake/firedrake/src/fiat/FIAT/bubble.py", line 24, in __init__ raise RuntimeError('Bubble element of degree %d and codimension %d has no dofs' % (degree, codim)) RuntimeError: Bubble element of degree 3 and codimension 0 has no dofs Any pointers would be appreciated -- *Bhavesh Shrimali* *Ph.D. Student, **Civil and Environmental Engineering,* *University of Illinois Urbana-Champaign* *Champaign, IL, United States* *Email : bshrima2(a)illinois.edu <bshrima2(a)illinois.edu>* *Web : **https://bhaveshshrimali.github.io/ <https://bhaveshshrimali.github.io/>*

2 2

Question from a New User
by Sean McNulty 20 Jun '21

20 Jun '21

Hello, I am a new user of your firedrake library. I'm running into an interesting issue where set_log_level() is not displaying all the information that I believe it should be. Unfortunately I haven't been able to find much in terms of help online. When setting the level to "DEBUG" I only get the message "pyop2:DEBUG Found existing compilation communicator". Would y'all happen to have an example on hand using this method that I could take a look at to make sure my syntax is correct? I've been trying to get set_log_level() to work with the firedrake introductory tutorials. Thanks! -Sean

1 0

Reading group
by Cotter, Colin J 08 Jun '21

08 Jun '21

Dear Firedrakers, Just to let everyone know that our informal reading group continues on Mondays at 4pm BST. The theme is quite broad in anything computational mathematics-y especially things that popped up on the ArXiV recently. It's mostly just a way to stay in touch whilst we are doing a lot of work from home. We seem to have some "senior staff" doing the reading over the next few weeks but it would be great to get more contributions from PhDs and postdocs. https://github.com/firedrakeproject/firedrake/wiki/Reading-group all the best --cjc Professor Colin Cotter (he/him) Department of Mathematics 755, Huxley Building Imperial College London South Kensington Campus United Kingdom of Great Britain and Northern Ireland +44 2075943468

1 0

Broken Installation
by Davies, Calvin 01 Jun '21

01 Jun '21

Hi, I am a MSc student at Imperial currently trying to use firedrake as part of my final research project. I am operating on a macOS Big Sur (11.2.3). I believe I have the necessary requirements on my system, however part way during building PETSc it broke, and appears to have an issue with 'pastix'. Please see attached the log files Any help would be much appreciated. Cheers, Calvin

2 1