October 2015 - firedrake - imperial.ac.uk

PyOP2 pull request #465
by Eike Mueller 12 Oct '15

12 Oct '15

Dear firedrake, what's the status of PyOP2 pull request #465 (https://github.com/OP2/PyOP2/pull/465)<https://github.com/OP2/PyOP2/pull/465>? Is this about to be merged? I've got some changes in my code that depend on it (both in the geometric multigrid code and the performance logging branch), so would be great to have support for C++ kernels in the master. Thanks, Eike

1 0

Re: [firedrake] firedrake-install fails with hypre
by David Ham 09 Oct '15

09 Oct '15

It seems I spoke too soon. Firedrake-install works for me but the manual configure Tuomas posted fails. Intriguing... On Fri, 9 Oct 2015 at 10:49 Ham, David A <david.ham(a)imperial.ac.uk> wrote: > I can reproduce this error. Bizarrely it's an unresolved reference to ceil > somewhere deep in Hypre. I have no idea why. Looking further now. > david > > On Thu, 8 Oct 2015 at 22:39 Justin Chang <jychang48(a)gmail.com> wrote: > >> Oh, I think it would only show up if you successfully installed petsc. >> Could be wrong though >> >> On Thu, Oct 8, 2015 at 3:33 PM, Tuomas Karna <tuomas.karna(a)gmail.com> >> wrote: >> >>> Hi Justin, >>> >>> I haven't used hypre with PETSc before. I'll try configuring it with the >>> same options. >>> >>> Actually, show us what's in >>> /home/tuomas/firedrake/lib/python2.7/site-packages/petsc/lib/petsc/conf/configure.log >>> >>> I tried to look for the cofigure.log but it's not under >>> /home/tuomas/firedrake/lib/python2.7/site-packages/petsc/lib/ ... >>> >>> - Tuomas >>> >>> >>> On 10/08/2015 02:29 PM, Justin Chang wrote: >>> >>> Actually, show us what's in >>> /home/tuomas/firedrake/lib/python2.7/site-packages/petsc/lib/petsc/conf/configure.log >>> >>> On Thu, Oct 8, 2015 at 3:28 PM, Justin Chang <jychang48(a)gmail.com> >>> wrote: >>> >>>> In your already installed PETSc version, did you have HYPRE configured >>>> with the same MPI compilers? If not, see if you can reproduce this error >>>> and see what's going on in the configure.log >>>> >>>> On Thu, Oct 8, 2015 at 3:15 PM, Tuomas Karna <tuomas.karna(a)gmail.com> >>>> wrote: >>>> >>>>> Hi all, >>>>> >>>>> I'm trying to install firedrake with >>>>> python firedrake-install --developer --log --disable_ssh >>>>> --no_package_manager >>>>> >>>>> and I get the following error when building PETSc (full log attached): >>>>> >>>>> >>>>> ******************************************************************************* >>>>> >>>>> UNABLE to CONFIGURE with GIVEN OPTIONS (see configure.log >>>>> for details): >>>>> >>>>> >>>>> ------------------------------------------------------------------------------- >>>>> >>>>> Downloaded hypre could not be used. Please check install in >>>>> /home/tuomas/firedrake/lib/python2.7/site-packages/petsc >>>>> >>>>> >>>>> ******************************************************************************* >>>>> >>>>> Any ideas? I'm on Ubuntu 15.04, I have PETSc installed but I've unset >>>>> PETSC_DIR. >>>>> >>>>> >>>>> - Tuomas >>>>> >>>>> >>>>> _______________________________________________ >>>>> firedrake mailing list >>>>> firedrake(a)imperial.ac.uk >>>>> https://mailman.ic.ac.uk/mailman/listinfo/firedrake >>>>> >>>>> >>>> >>> >>> >>> _______________________________________________ >>> firedrake mailing listfiredrake@imperial.ac.ukhttps://mailman.ic.ac.uk/mailman/listinfo/firedrake >>> >>> >>> >>> _______________________________________________ >>> firedrake mailing list >>> firedrake(a)imperial.ac.uk >>> https://mailman.ic.ac.uk/mailman/listinfo/firedrake >>> >>> >>

4 5

Re: [firedrake] Modifying diagonal entries of Jacobian (preconditioning)
by David Ham 09 Oct '15

09 Oct '15

Hi Henrik, You can pass an alternative Jacobian (ie not the true derivative(J, u) ) to the solver by explicitly passing a form for J. The critical person who is the expert on this is Lawrence, who is away for the next week. I'm not sure if he's on email. Regards, David On Fri, 9 Oct 2015 at 14:16 Buesing, Henrik <HBuesing(a)eonerc.rwth-aachen.de> wrote: > Dear Firedrakers, > > > > if I wanted to modify the diagonal entries of my Jacobian in a Newton step > during a time step… How would I do this? > Do I have to define my own preconditioner here? > > Henrik > > > > > > -- > > Dipl.-Math. Henrik Büsing > > Institute for Applied Geophysics and Geothermal Energy > > E.ON Energy Research Center > > RWTH Aachen University > > ------------------------------------------------------ > > Mathieustr. 10 | Tel +49 (0)241 80 49907 > > 52074 Aachen, Germany | Fax +49 (0)241 80 49889 > > ------------------------------------------------------ > > http://www.eonerc.rwth-aachen.de/GGE > > hbuesing(a)eonerc.rwth-aachen.de > > ------------------------------------------------------ > > >

1 0

Re: [firedrake] Dirichlet BC on extruded mesh with discontinuous function space
by David Ham 09 Oct '15

09 Oct '15

Hi Henrik, There is a pull request open, however we currently have a bug. Top boundary conditions are being applied in the wrong place. Once I hunt down that issue the feature will get merged. Regards, David On Fri, 9 Oct 2015 at 09:06 Buesing, Henrik <HBuesing(a)eonerc.rwth-aachen.de> wrote: > > > *Von:* firedrake-bounces(a)imperial.ac.uk [mailto: > firedrake-bounces(a)imperial.ac.uk] *Im Auftrag von *David Ham > > > *Gesendet:* 23 September 2015 13:23 > *An:* firedrake > > *Betreff:* Re: [firedrake] Dirichlet BC on extruded mesh with > discontinuous function space > > > > I believe this functionality exists in Tuomas' branch. I'll be in Oregon > next week and merging this is on my to do list. > > *[Buesing, Henrik] @David: Did you manage to integrate this functionality > in the main branch? And do you have a small example how to set the > Dirichlet boundary conditions?* > > * Thank you!* > > > > On Wed, 23 Sep 2015 at 11:34 Buesing, Henrik < > HBuesing(a)eonerc.rwth-aachen.de> wrote: > > > > *Von:* firedrake-bounces(a)imperial.ac.uk [mailto: > firedrake-bounces(a)imperial.ac.uk] *Im Auftrag von *Andrew McRae > *Gesendet:* 23 September 2015 12:16 > > > *An:* firedrake(a)imperial.ac.uk > *Betreff:* Re: [firedrake] Dirichlet BC on extruded mesh with > discontinuous function space > > > > Yes, but it should still be possible to set the value of the function in > those cells. Didn't Tuomas make some progress on this functionality? I > don't remember what the current status is. > > *[Buesing, Henrik] That would be nice. Then I can just prescribe the right > value at those cells. If this is enforced during the KSP, SNES solves > everything should be fine. * > > As a workaround, Miklos's suggestion may suffice. > > On 23 September 2015 at 11:09, Miklos Homolya <m.homolya14(a)imperial.ac.uk> > wrote: > > Dear Henrik, > > > > You’ve got a DG0 space, i.e. all your nodes are in the cell centres. > Neither topologically nor geometrically have you any nodes on the boundary > faces. > > > > Maybe you want to impose your BC weakly? > > > > Regards, > > Miklos > > > > On 23 Sep 2015, at 12:05, Buesing, Henrik <HBuesing(a)eonerc.rwth-aachen.de> > wrote: > > > > Dear Firedrakers, > > > > I’m constructing a DG space on an extruded mesh. > > > > > > > > *meshbase = RectangleMesh(Nx, Ny, Lx, Ly, quadrilateral=True) mesh = > ExtrudedMesh(meshbase, Nz, Delta_z) horiz_elt = FiniteElement("DG", > quadrilateral, 0) vert_elt = FiniteElement("DG", interval, 0) elt = > OuterProductElement(horiz_elt, vert_elt)* > > *DG = FunctionSpace(mesh, elt)* > > *W=DG*DG* > > > > Now I would like to impose Dirchlet BC on one of the boundaries. > > I tried > > *bc0 = DirichletBC(W.sub(0), Expression("1.0e7"), 2,method="geometric")* > > > > and get a > > > > *ValueError: Geometric boundary conditions are not yet supported on > extruded meshes* > > > > Using the standard topological way, seems to have no effect at all. > > > > Is there a way around this? I would be fine with fixing the value on the > most outer cell layer, if that is possible… > > > > Henrik > > > > > > > > > > > > -- > > Dipl.-Math. Henrik Büsing > > Institute for Applied Geophysics and Geothermal Energy > > E.ON Energy Research Center > > RWTH Aachen University > > ------------------------------------------------------ > > Mathieustr. 10 | Tel +49 (0)241 80 49907 > > 52074 Aachen, Germany | Fax +49 (0)241 80 49889 > > ------------------------------------------------------ > > http://www.eonerc.rwth-aachen.de/GGE > > hbuesing(a)eonerc.rwth-aachen.de > > ------------------------------------------------------ > > > > _______________________________________________ > firedrake mailing list > firedrake(a)imperial.ac.uk > https://mailman.ic.ac.uk/mailman/listinfo/firedrake > > > >

2 1

Re: [firedrake] firedrake-install fails with hypre
by David Ham 09 Oct '15

09 Oct '15

I can reproduce this error. Bizarrely it's an unresolved reference to ceil somewhere deep in Hypre. I have no idea why. Looking further now. david On Thu, 8 Oct 2015 at 22:39 Justin Chang <jychang48(a)gmail.com> wrote: > Oh, I think it would only show up if you successfully installed petsc. > Could be wrong though > > On Thu, Oct 8, 2015 at 3:33 PM, Tuomas Karna <tuomas.karna(a)gmail.com> > wrote: > >> Hi Justin, >> >> I haven't used hypre with PETSc before. I'll try configuring it with the >> same options. >> >> Actually, show us what's in >> /home/tuomas/firedrake/lib/python2.7/site-packages/petsc/lib/petsc/conf/configure.log >> >> I tried to look for the cofigure.log but it's not under >> /home/tuomas/firedrake/lib/python2.7/site-packages/petsc/lib/ ... >> >> - Tuomas >> >> >> On 10/08/2015 02:29 PM, Justin Chang wrote: >> >> Actually, show us what's in >> /home/tuomas/firedrake/lib/python2.7/site-packages/petsc/lib/petsc/conf/configure.log >> >> On Thu, Oct 8, 2015 at 3:28 PM, Justin Chang <jychang48(a)gmail.com> wrote: >> >>> In your already installed PETSc version, did you have HYPRE configured >>> with the same MPI compilers? If not, see if you can reproduce this error >>> and see what's going on in the configure.log >>> >>> On Thu, Oct 8, 2015 at 3:15 PM, Tuomas Karna <tuomas.karna(a)gmail.com> >>> wrote: >>> >>>> Hi all, >>>> >>>> I'm trying to install firedrake with >>>> python firedrake-install --developer --log --disable_ssh >>>> --no_package_manager >>>> >>>> and I get the following error when building PETSc (full log attached): >>>> >>>> >>>> ******************************************************************************* >>>> >>>> UNABLE to CONFIGURE with GIVEN OPTIONS (see configure.log >>>> for details): >>>> >>>> >>>> ------------------------------------------------------------------------------- >>>> >>>> Downloaded hypre could not be used. Please check install in >>>> /home/tuomas/firedrake/lib/python2.7/site-packages/petsc >>>> >>>> >>>> ******************************************************************************* >>>> >>>> Any ideas? I'm on Ubuntu 15.04, I have PETSc installed but I've unset >>>> PETSC_DIR. >>>> >>>> >>>> - Tuomas >>>> >>>> >>>> _______________________________________________ >>>> firedrake mailing list >>>> firedrake(a)imperial.ac.uk >>>> https://mailman.ic.ac.uk/mailman/listinfo/firedrake >>>> >>>> >>> >> >> >> _______________________________________________ >> firedrake mailing listfiredrake@imperial.ac.ukhttps://mailman.ic.ac.uk/mailman/listinfo/firedrake >> >> >> >> _______________________________________________ >> firedrake mailing list >> firedrake(a)imperial.ac.uk >> https://mailman.ic.ac.uk/mailman/listinfo/firedrake >> >> >

1 0

Re: [firedrake] Jacobian evaluation with initial conditions
by David Ham 09 Oct '15

09 Oct '15

Dear Henrik, J = derivative(F, u) results in J being a Form, which is to say a symbolic representation of the Jacobian. If F is nonlinear then u will still feature in F. At this point the Jacobian isn't evaluated so it doesn't really make sense to talk about the Jacobian at u=u0. Typing: M = assemble(J) evaluates the Jacobian resulting in a matrix. Whatever value the field u holds when the assemble call occurs will be the value at which the Jacobian is assembled. You can project, interpolate or assign initial values into u at any stage before calling assemble in order to cause the evaluation to occur at your choice of initial condition. Regards, David On Fri, 9 Oct 2015 at 08:52 Buesing, Henrik <HBuesing(a)eonerc.rwth-aachen.de> wrote: > Dear Firedrakers, > > > > I’m evaluating the Jacobian dF/du of my nonlinear function F by, > > > > J = derivative(F,u) > > > > This seems to evaluate with zero u: dF(0)/du. > > What I would like to have is dF(u0)/du, the Jacobian evaluated at the > initial conditions. > > > > How would I do that? Thank you! > > Henrik > > > > > > -- > > Dipl.-Math. Henrik Büsing > > Institute for Applied Geophysics and Geothermal Energy > > E.ON Energy Research Center > > RWTH Aachen University > > ------------------------------------------------------ > > Mathieustr. 10 | Tel +49 (0)241 80 49907 > > 52074 Aachen, Germany | Fax +49 (0)241 80 49889 > > ------------------------------------------------------ > > http://www.eonerc.rwth-aachen.de/GGE > > hbuesing(a)eonerc.rwth-aachen.de > > ------------------------------------------------------ > > >

2 1

Re: [firedrake] Dirichlet BC on extruded mesh with discontinuous function space
by David Ham 09 Oct '15

09 Oct '15

I believe this functionality exists in Tuomas' branch. I'll be in Oregon next week and merging this is on my to do list. On Wed, 23 Sep 2015 at 11:34 Buesing, Henrik <HBuesing(a)eonerc.rwth-aachen.de> wrote: > > > *Von:* firedrake-bounces(a)imperial.ac.uk [mailto: > firedrake-bounces(a)imperial.ac.uk] *Im Auftrag von *Andrew McRae > *Gesendet:* 23 September 2015 12:16 > > > *An:* firedrake(a)imperial.ac.uk > *Betreff:* Re: [firedrake] Dirichlet BC on extruded mesh with > discontinuous function space > > > > Yes, but it should still be possible to set the value of the function in > those cells. Didn't Tuomas make some progress on this functionality? I > don't remember what the current status is. > > *[Buesing, Henrik] That would be nice. Then I can just prescribe the right > value at those cells. If this is enforced during the KSP, SNES solves > everything should be fine. * > > As a workaround, Miklos's suggestion may suffice. > > On 23 September 2015 at 11:09, Miklos Homolya <m.homolya14(a)imperial.ac.uk> > wrote: > > Dear Henrik, > > > > You’ve got a DG0 space, i.e. all your nodes are in the cell centres. > Neither topologically nor geometrically have you any nodes on the boundary > faces. > > > > Maybe you want to impose your BC weakly? > > > > Regards, > > Miklos > > > > On 23 Sep 2015, at 12:05, Buesing, Henrik <HBuesing(a)eonerc.rwth-aachen.de> > wrote: > > > > Dear Firedrakers, > > > > I’m constructing a DG space on an extruded mesh. > > > > > > > > *meshbase = RectangleMesh(Nx, Ny, Lx, Ly, quadrilateral=True) mesh = > ExtrudedMesh(meshbase, Nz, Delta_z) horiz_elt = FiniteElement("DG", > quadrilateral, 0) vert_elt = FiniteElement("DG", interval, 0) elt = > OuterProductElement(horiz_elt, vert_elt)* > > *DG = FunctionSpace(mesh, elt)* > > *W=DG*DG* > > > > Now I would like to impose Dirchlet BC on one of the boundaries. > > I tried > > *bc0 = DirichletBC(W.sub(0), Expression("1.0e7"), 2,method="geometric")* > > > > and get a > > > > *ValueError: Geometric boundary conditions are not yet supported on > extruded meshes* > > > > Using the standard topological way, seems to have no effect at all. > > > > Is there a way around this? I would be fine with fixing the value on the > most outer cell layer, if that is possible… > > > > Henrik > > > > > > > > > > > > -- > > Dipl.-Math. Henrik Büsing > > Institute for Applied Geophysics and Geothermal Energy > > E.ON Energy Research Center > > RWTH Aachen University > > ------------------------------------------------------ > > Mathieustr. 10 | Tel +49 (0)241 80 49907 > > 52074 Aachen, Germany | Fax +49 (0)241 80 49889 > > ------------------------------------------------------ > > http://www.eonerc.rwth-aachen.de/GGE > > hbuesing(a)eonerc.rwth-aachen.de > > ------------------------------------------------------ > > > > _______________________________________________ > firedrake mailing list > firedrake(a)imperial.ac.uk > https://mailman.ic.ac.uk/mailman/listinfo/firedrake > > >

2 2

Jacobian evaluation with initial conditions
by Buesing, Henrik 09 Oct '15

09 Oct '15

Dear Firedrakers, I'm evaluating the Jacobian dF/du of my nonlinear function F by, J = derivative(F,u) This seems to evaluate with zero u: dF(0)/du. What I would like to have is dF(u0)/du, the Jacobian evaluated at the initial conditions. How would I do that? Thank you! Henrik -- Dipl.-Math. Henrik Büsing Institute for Applied Geophysics and Geothermal Energy E.ON Energy Research Center RWTH Aachen University ------------------------------------------------------ Mathieustr. 10 | Tel +49 (0)241 80 49907 52074 Aachen, Germany | Fax +49 (0)241 80 49889 ------------------------------------------------------ http://www.eonerc.rwth-aachen.de/GGE hbuesing(a)eonerc.rwth-aachen.de ------------------------------------------------------

1 0

firedrake-install fails with hypre
by Tuomas Karna 08 Oct '15

08 Oct '15

Hi all, I'm trying to install firedrake with python firedrake-install --developer --log --disable_ssh --no_package_manager and I get the following error when building PETSc (full log attached): ******************************************************************************* UNABLE to CONFIGURE with GIVEN OPTIONS (see configure.log for details): ------------------------------------------------------------------------------- Downloaded hypre could not be used. Please check install in /home/tuomas/firedrake/lib/python2.7/site-packages/petsc ******************************************************************************* Any ideas? I'm on Ubuntu 15.04, I have PETSc installed but I've unset PETSC_DIR. - Tuomas

2 4

Re: [firedrake] weirdness on bendy branch
by Fabio Luporini 08 Oct '15

08 Oct '15

Hello Stephan, Can you pull the latest enhance-rewriter-mode3-logflops (i've just pushed a fix that seems to work), clean the caches (firedrake/scripts/firedrake-clean), and try it again? Thanks -- Fabio 2015-10-06 20:56 GMT+01:00 Luporini, Fabio <f.luporini12(a)imperial.ac.uk>: > Hi Stephen, > > I guess this is on me. Will look into this tomorrow and will get back to > you. Thanks. > > -- Fabio > > 2015-10-06 19:36 GMT+01:00 Kramer, Stephan <s.kramer(a)imperial.ac.uk>: > >> More weirdness, I'm afraid. If I change the function space V to: >> >> V = FunctionSpace(mesh2d, "RT", 1) >> >> I still get different answers between M and M1+M2. This is using the >> bendy branches for firedrake/ffc/ufl and >> enhance-rewriter-mode3-logflops for COFFEE. Again M==M1+M2 using >> parameters['coffee']={} and I get the same answer for M in "flat" >> firedrake. >> >> Any other magic branches I could try? ;-) >> Seriously though, any suggestions are welcome >> >> Cheers >> Stephan >> >> >> On 02/10/15 13:10, Stephan Kramer wrote: >> > Hey guys >> > >> > I've been trying to implement a simple IP viscosity scheme for DG, but >> > was getting weird results, which I think I've tracked down to a bug >> > related to fd_bendy. >> > >> > If I run the following: >> > >> > from firedrake import * >> > >> > mesh2d = UnitSquareMesh(1,1) >> > V = VectorFunctionSpace(mesh2d, "DG", 1) >> > uv = Function(V) >> > U_test = TestFunction(V) >> > >> > n = FacetNormal(mesh2d) >> > F1 = dot(jump(U_test[0], n), jump(uv[0], n))*dS >> > F2 = -dot(avg(grad(U_test[0])), jump(uv[0], n))*dS >> > F=F1+F2 >> > F0 = dot(jump(U_test[0],n)-avg(grad(U_test[0])), jump(uv[0], n))*dS >> > >> > M=assemble(derivative(F, uv)).M.values >> > M1=assemble(derivative(F1, uv)).M.values >> > M2=assemble(derivative(F2, uv)).M.values >> > M0=assemble(derivative(F0, uv)).M.values >> > >> > I really expect M, M1+M2 and M0 to be all the same. On firedrake master >> > (and in fact Dolfin) these are all the same indeed. But in my fd_bendy >> > enviroment (fd_bendy for ffc and ufl, bendy_changes for firedrake) the >> > values for M are different. Final piece of information is that if I use >> > parameters['coffee']={}, everything is fine again. Other than that I'm >> > stuck though, so I'd much appreciate any pointers into what I could >> > check next >> > >> > Cheers >> > Stephan >> > >> > _______________________________________________ >> > firedrake mailing list >> > firedrake(a)imperial.ac.uk >> > https://mailman.ic.ac.uk/mailman/listinfo/firedrake >> > >> > >> >> >> _______________________________________________ >> firedrake mailing list >> firedrake(a)imperial.ac.uk >> https://mailman.ic.ac.uk/mailman/listinfo/firedrake >> > >

2 1