- firedrake - imperial.ac.uk

Re: [firedrake] operations on matrices
by David Ham 05 Jan '15

05 Jan '15

On 5 January 2015 at 09:37, Cotter, Colin J <colin.cotter(a)imperial.ac.uk> wrote: > Oh yes, good isolation of the problem. > > If alpha also depends on values of A, do we have a problem there too? > > Yes. Same problem. There is no current way for a parallel loop to read the entries of a matrix. > -cjc > > On 05/01/15 09:33, David Ham wrote: > > Hi Colin, > > > > There is no way for a parallel loop to read from a matrix. However the > > operation you describe appears to be: > > > > assemble A > > assemble alpha > > > > scale entries of A by the corresponding entries of alpha. > > > > The last step is clearly the problem. I wonder if this could be achieved > > by some PETSc operation on the matrices. > > > > On 5 January 2015 at 09:02, Cotter, Colin J <colin.cotter(a)imperial.ac.uk > > <mailto:colin.cotter@imperial.ac.uk>> wrote: > > > > Dear all, > > Happy New Year! > > > > Perhaps I made the mistake of making some complex explanation before > > asking my question. > > > > What is the best way to make adjustments to matrix entries as part > of a > > loop over elements? > > > > cheers > > --cjc > > > > On 22/12/14 11:13, Cotter, Colin J wrote: > > > Dear Firedrakers, > > > I've been recently revisiting the "algebraic flux correction" > > schemes > > > of Dmitri Kuzmin, with the aim of getting a conservative+bounded > > > advection scheme for temperature in our NWP setup. These schemes > > involve > > > the following steps: > > > > > > 1) Forming the consistent mass matrix (which is column-diagonal) > > M_C for > > > the temperature space. > > > 2) Constructing the following matrix with the same sparsity as > M_C: > > > > > > A_{ij} = (M_C)_{ij}(T_i-T_j) > > > > > > where T_i is the value of temperature at node i. > > > > > > 3) "Limiting" the matrix by replacing > > > > > > A_{ij} -> A_{ij}\alpha_{ij} > > > > > > where \alpha_{ij} depends on various field values at nodes i and > > j (only > > > needs to be evaluated when nodes i and j share an element). > > > > > > 4) Evaluating Ax where x is the vector containing 1s, and adding > x to > > > the RHS of mass-matrix projection equation before solving. > > > > > > My question is: how to implement this in an efficient and > > parallel-safe > > > way in the Firedrake/PyOP2 framework? In particular, step (3) > > involves > > > looping over elements, and correcting matrix entries. Also, I'm > > not sure > > > of the best way to assemble A. > > > > > > all the best > > > --Colin > > > > > > _______________________________________________ > > firedrake mailing list > > firedrake(a)imperial.ac.uk <mailto:firedrake@imperial.ac.uk> > > https://mailman.ic.ac.uk/mailman/listinfo/firedrake > > > > > > > > > > -- > > Dr David Ham > > Departments of Mathematics and Computing > > Imperial College London > > > > http://www.imperial.ac.uk/people/david.ham > > > > > > _______________________________________________ > > 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 > -- Dr David Ham Departments of Mathematics and Computing Imperial College London http://www.imperial.ac.uk/people/david.ham

3 3

Re: [firedrake] operations on matrices
by David Ham 05 Jan '15

05 Jan '15

Hi Colin, There is no way for a parallel loop to read from a matrix. However the operation you describe appears to be: assemble A assemble alpha scale entries of A by the corresponding entries of alpha. The last step is clearly the problem. I wonder if this could be achieved by some PETSc operation on the matrices. On 5 January 2015 at 09:02, Cotter, Colin J <colin.cotter(a)imperial.ac.uk> wrote: > Dear all, > Happy New Year! > > Perhaps I made the mistake of making some complex explanation before > asking my question. > > What is the best way to make adjustments to matrix entries as part of a > loop over elements? > > cheers > --cjc > > On 22/12/14 11:13, Cotter, Colin J wrote: > > Dear Firedrakers, > > I've been recently revisiting the "algebraic flux correction" schemes > > of Dmitri Kuzmin, with the aim of getting a conservative+bounded > > advection scheme for temperature in our NWP setup. These schemes involve > > the following steps: > > > > 1) Forming the consistent mass matrix (which is column-diagonal) M_C for > > the temperature space. > > 2) Constructing the following matrix with the same sparsity as M_C: > > > > A_{ij} = (M_C)_{ij}(T_i-T_j) > > > > where T_i is the value of temperature at node i. > > > > 3) "Limiting" the matrix by replacing > > > > A_{ij} -> A_{ij}\alpha_{ij} > > > > where \alpha_{ij} depends on various field values at nodes i and j (only > > needs to be evaluated when nodes i and j share an element). > > > > 4) Evaluating Ax where x is the vector containing 1s, and adding x to > > the RHS of mass-matrix projection equation before solving. > > > > My question is: how to implement this in an efficient and parallel-safe > > way in the Firedrake/PyOP2 framework? In particular, step (3) involves > > looping over elements, and correcting matrix entries. Also, I'm not sure > > of the best way to assemble A. > > > > all the best > > --Colin > > > _______________________________________________ > firedrake mailing list > firedrake(a)imperial.ac.uk > https://mailman.ic.ac.uk/mailman/listinfo/firedrake > -- Dr David Ham Departments of Mathematics and Computing Imperial College London http://www.imperial.ac.uk/people/david.ham

3 2

operations on matrices
by Cotter, Colin J 05 Jan '15

05 Jan '15

Dear Firedrakers, I've been recently revisiting the "algebraic flux correction" schemes of Dmitri Kuzmin, with the aim of getting a conservative+bounded advection scheme for temperature in our NWP setup. These schemes involve the following steps: 1) Forming the consistent mass matrix (which is column-diagonal) M_C for the temperature space. 2) Constructing the following matrix with the same sparsity as M_C: A_{ij} = (M_C)_{ij}(T_i-T_j) where T_i is the value of temperature at node i. 3) "Limiting" the matrix by replacing A_{ij} -> A_{ij}\alpha_{ij} where \alpha_{ij} depends on various field values at nodes i and j (only needs to be evaluated when nodes i and j share an element). 4) Evaluating Ax where x is the vector containing 1s, and adding x to the RHS of mass-matrix projection equation before solving. My question is: how to implement this in an efficient and parallel-safe way in the Firedrake/PyOP2 framework? In particular, step (3) involves looping over elements, and correcting matrix entries. Also, I'm not sure of the best way to assemble A. all the best --Colin

2 1

BCs on velocity for mixed gravity wave system solver
by Eike Mueller 30 Dec '14

30 Dec '14

Dear firedrakers, I’ve got a first, lowest-order version of the matrix-free solver for the mixed 3d gravity-wave problem working now (pushed to “vertical" branch of firedrake-helmholtzsolver). I precondition with the matrix-free multigrid in pressure space, and it all converges nicely (although it is probably horribly inefficient in places). The equations for velocity u, pressure p and buoyancy b are du/dt = \grad(p) + b*\hat{z} dp/dt = -c^2 \div(u) db/dt = -N^2 u.\hat{z} Full details on the solver are in GravityWaves.tex in the directory https://github.com/firedrakeproject/firedrake-helmholtzsolver/tree/vertical… <https://github.com/firedrakeproject/firedrake-helmholtzsolver/tree/vertical…> However, I’m struggling to incorporate the boundary condition u.n=0 at the bottom and top of the domain (this is a spherical shell in my case - but should all also work for a 2d annulus) into the solver. I presume it is a Dirichlet BC, but of the form n_0*u_0 + n_1*u_1 + n_2*u_2 = 0, and I do not set the values explicity as in, for example, (u_0,u_1,u_2) = (0,0,0). How can I specify this? Also, where in the solver do I actually have to enforce it? In the mixed solver I apply the mixed operator, so I think I need to enforce it in the u-equation there. Also, in the mixed preconditioner I do a back-substitution for u after the pressure solve, so I think I have to do it there as well. Thanks a lot, Eike -- Dr Eike Hermann Mueller Research Associate (PostDoc) Department of Mathematical Sciences University of Bath Bath BA2 7AY, United Kingdom +44 1225 38 5803 e.mueller(a)bath.ac.uk http://people.bath.ac.uk/em459/

1 1

Re: [firedrake] BCs on velocity for mixed gravity wave system solver
by Andrew McRae 28 Dec '14

28 Dec '14

Hi Eike, In general, we don't actually have proper functionality for that. However, assuming you're using some H(div) space for u, including an HDiv(U2xV0) + HDiv(U1xV1) sort of thing, you can obtain the BC u.n=0 by doing, for example, bcs = [DirichletBC(W2, 0.0, "bottom"), DirichletBC(W2, 0.0, "top")] Why does this work? Well, this causes any degrees of freedom on the 'bottom' and 'top' boundaries to be set to 0. But the degrees of freedom on the boundary in an H(div) space are precisely u.n (technically with some sort of scaling, but scaled 0 is still 0). And any interior DOFs, or DOFs on other facets, have vanishing normal component on the original facet. So indeed, this does the right thing. If your velocity was in something like VectorFunctionSpace(mesh, "CG", n), unfortunately this approach wouldn't work, and I think you'd have to apply the BC weakly by adding on a surface integral. But in the H(div) case, what I described above happens to work. Hope this helps! Best, Andrew On 27 December 2014 at 14:14, Eike Mueller <e.mueller(a)bath.ac.uk> wrote: > Dear firedrakers, > > I’ve got a first, lowest-order version of the matrix-free solver for the > mixed 3d gravity-wave problem working now (pushed to “vertical" branch of > firedrake-helmholtzsolver). I precondition with the matrix-free multigrid > in pressure space, and it all converges nicely (although it is probably > horribly inefficient in places). The equations for velocity u, pressure p > and buoyancy b are > > du/dt = \grad(p) + b*\hat{z} > dp/dt = -c^2 \div(u) > db/dt = -N^2 u.\hat{z} > > Full details on the solver are in GravityWaves.tex in the directory > https://github.com/firedrakeproject/firedrake-helmholtzsolver/tree/vertical… > > However, I’m struggling to incorporate the boundary condition u.n=0 at > the bottom and top of the domain (this is a spherical shell in my case - > but should all also work for a 2d annulus) into the solver. I presume it is > a Dirichlet BC, but of the form n_0*u_0 + n_1*u_1 + n_2*u_2 = 0, and I do > not set the values explicity as in, for example, (u_0,u_1,u_2) = (0,0,0). > How can I specify this? Also, where in the solver do I actually have to > enforce it? > > In the mixed solver I apply the mixed operator, so I think I need to > enforce it in the u-equation there. > Also, in the mixed preconditioner I do a back-substitution for u after the > pressure solve, so I think I have to do it there as well. > > Thanks a lot, > > Eike > > -- > > Dr Eike Hermann Mueller > Research Associate (PostDoc) > > Department of Mathematical Sciences > University of Bath > Bath BA2 7AY, United Kingdom > > +44 1225 38 5803 > e.mueller(a)bath.ac.uk > http://people.bath.ac.uk/em459/ > >

3 2

1+1 dimensional hierarchical meshes and function spaces
by Eike Mueller 22 Dec '14

22 Dec '14

Dear firedrakers, do the hierarchical meshes and function spaces currently only work in 2+1 dimension and not for 1+1? If I run the code below it works for dimension=3, but if I replace this by dimension=2 it crashes with [0]PETSC ERROR: DMPlexGetCellRefiner_Internal() line 6777 in /Users/eikemueller/PostDocBath/EllipticSolvers/petsc/src/dm/impls/plex/plexrefine.c Unknown dimension 1 for cell refiner -------------------------------------------------------------------------- MPI_ABORT was invoked on rank 0 in communicator MPI_COMM_WORLD with errorcode 62. NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes. You may or may not see output from other processes, depending on exactly when Open MPI kills them. -------------------------------------------------------------------------- I want to work in dimension=2+1, but having dimension=1+1 would be useful for testing. Thanks a lot, Eike from firedrake import * dimension = 2 D = 0.1 nlayers=4 nlevel = 4 if (dimension == 2): ncells=3 host_mesh = CircleManifoldMesh(ncells) else: refcount = 0 host_mesh = UnitIcosahedralSphereMesh(refcount) host_mesh_hierarchy = MeshHierarchy(host_mesh,nlevel) mesh_hierarchy = ExtrudedMeshHierarchy(host_mesh_hierarchy, layers=nlayers, extrusion_type='radial', layer_height=D/nlayers) if (dimension == 2): U2 = FiniteElement('DG',interval,1) V1 = FiniteElement('DG',interval,1) else: U2 = FiniteElement('DG',triangle,0) V1 = FiniteElement('DG',interval,0) W3_elt = OuterProductElement(U2,V1) W3 = FunctionSpaceHierarchy(mesh_hierarchy,W3_elt) f = FunctionHierarchy(W3) f[-1].interpolate(Expression('x[0]*x[1]')) print norm(f[-1]) -- Dr Eike Hermann Mueller Research Associate (PostDoc) Department of Mathematical Sciences University of Bath Bath BA2 7AY, United Kingdom +44 1225 38 5803 e.mueller(a)bath.ac.uk http://people.bath.ac.uk/em459/

3 11

Re: [firedrake] Problem with Nonlinear solvers after update
by Fabio Luporini 22 Dec '14

22 Dec '14

Hi Anna, Yes, this problem is clearly caused by coffee; I'll work on this after the holidays. Thanks. -- Fabio 2014-12-22 15:28 GMT+01:00 Anna Kalogirou <A.Kalogirou(a)leeds.ac.uk>: > Thank you very much Andrew. That fixed my problems! > > Anna. > > P.S. Happy holidays and Merry Christmas to all of you :) > > > On 22 Dec 2014, at 14:22, Eike Mueller <e.mueller(a)bath.ac.uk<mailto: > e.mueller(a)bath.ac.uk>> wrote: > > Oh, I should have done what Andrew suggests before sending my email... > > op2.init() > parameters["coffee"]["O2"] = False > > seems to fix at least part of my problem :-) > > Sorry for the noise, > > Eike > > -- > > Dr Eike Hermann Mueller > Research Associate (PostDoc) > > Department of Mathematical Sciences > University of Bath > Bath BA2 7AY, United Kingdom > > +44 1225 38 5803 > e.mueller(a)bath.ac.uk<mailto:e.mueller@bath.ac.uk> > http://people.bath.ac.uk/em459/ > > On 22 Dec 2014, at 13:58, Andrew McRae <a.mcrae12(a)imperial.ac.uk<mailto: > a.mcrae12(a)imperial.ac.uk>> wrote: > > Ugh, one of the recent changes we made was to turn on COFFEE's > optimisations by default. Unfortunately, this has led to some teething > problems. > > I assume that all components -- COFFEE, Firedrake, PyOP2, FFC, UFL, and > FIAT -- are up to date, and that you're running with a clean cache (from > the Firedrake base directory, run "./scripts/firedrake-clean"). > > If the problem persists, as a short-term workaround, can you add > > op2.init() > parameters["coffee"]["O2"] = False > > to the start of your file, just after the imports? This will result in > sub-optimal performance during the assembly stage, so isn't ideal in the > long-term. > > Fabio, our resident barista, will have some fixes after the new year, I'm > sure. > > Andrew > > On 22 December 2014 at 13:47, Anna Kalogirou <A.Kalogirou(a)leeds.ac.uk > <mailto:A.Kalogirou@leeds.ac.uk>> wrote: > Dear all, > > I very recently updated firedrake, pyop2 and installed coffee, but since > then none of my codes include nonlinear solvers don’t work (they used to > work fine before the update). > > I even tried to run the burgers.py code from the demo and even that > diverges. > > I get something like: > > Traceback (most recent call last): > File "benney-luke.py", line 57, in <module> > phi_solver = solver_phi(phi1, phi1, phi0, eta0, etaR, phi, gamma, dt, > mu, epsilon, is_linear); > File "/Users/matak/Documents/Firedrake/modules/solvers.py", line 26, in > solver_phi > phi_solver = > NonlinearVariationalSolver(phi_problem,solver_parameters={'snes_monitor': > True,'ksp_monitor': True,'snes_linesearch_monitor': True}) > File "/Users/matak/firedrake/firedrake/variational_solver.py", line 97, > in __init__ > form_compiler_parameters=self._problem.form_compiler_parameters) > File "/Users/matak/firedrake/firedrake/assemble.py", line 58, in assemble > form_compiler_parameters=form_compiler_parameters) > File "/Users/matak/firedrake/firedrake/assemble.py", line 79, in > _assemble > kernels = ffc_interface.compile_form(f, "form", > parameters=form_compiler_parameters) > File "/Users/matak/firedrake/firedrake/ffc_interface.py", line 268, in > compile_form > ffc_kernel = FFCKernel(f, name + str(i) + str(j), parameters) > File > "/usr/local/Cellar/python/2.7.8_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PyOP2-0.11.0_148_g8a1955c-py2.7-macosx-10.9-x86_64.egg/pyop2/caching.py", > line 202, in __new__ > obj = make_obj() > File > "/usr/local/Cellar/python/2.7.8_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PyOP2-0.11.0_148_g8a1955c-py2.7-macosx-10.9-x86_64.egg/pyop2/caching.py", > line 192, in make_obj > obj.__init__(*args, **kwargs) > File "/Users/matak/firedrake/firedrake/ffc_interface.py", line 194, in > __init__ > (name, it.integral_type(), it.subdomain_id()), opts, inc), > File > "/usr/local/Cellar/python/2.7.8_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PyOP2-0.11.0_148_g8a1955c-py2.7-macosx-10.9-x86_64.egg/pyop2/backends.py", > line 118, in __call__ > return t(*args, **kwargs) > File > "/usr/local/Cellar/python/2.7.8_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PyOP2-0.11.0_148_g8a1955c-py2.7-macosx-10.9-x86_64.egg/pyop2/caching.py", > line 202, in __new__ > obj = make_obj() > File > "/usr/local/Cellar/python/2.7.8_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PyOP2-0.11.0_148_g8a1955c-py2.7-macosx-10.9-x86_64.egg/pyop2/caching.py", > line 192, in make_obj > obj.__init__(*args, **kwargs) > File > "/usr/local/Cellar/python/2.7.8_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PyOP2-0.11.0_148_g8a1955c-py2.7-macosx-10.9-x86_64.egg/pyop2/base.py", > line 3632, in __init__ > self._code = self._ast_to_c(code, opts) > File > "/usr/local/Cellar/python/2.7.8_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PyOP2-0.11.0_148_g8a1955c-py2.7-macosx-10.9-x86_64.egg/pyop2/host.py", > line 63, in _ast_to_c > ast_handler.plan_cpu(opts) > File "/usr/local/lib/python2.7/site-packages/coffee/plan.py", line 405, > in plan_cpu > loop_opts = _generate_cpu_code(self, **params) > File "/usr/local/lib/python2.7/site-packages/coffee/plan.py", line 241, > in _generate_cpu_code > loop_opt.rewrite(rewrite) > File "/usr/local/lib/python2.7/site-packages/coffee/optimizer.py", line > 110, in rewrite > ew.expand() > File "/usr/local/lib/python2.7/site-packages/coffee/rewriter.py", line > 168, in expand > ee.expand(self.stmt.children[1], self.stmt, itvars_occs, itvar_exp) > File "/usr/local/lib/python2.7/site-packages/coffee/rewriter.py", line > 565, in expand > return self.expand(node.children[0], node, itvars_occs, itvar_exp) > File "/usr/local/lib/python2.7/site-packages/coffee/rewriter.py", line > 567, in expand > l_node, l_type = self.expand(node.children[0], node, itvars_occs, > itvar_exp) > File "/usr/local/lib/python2.7/site-packages/coffee/rewriter.py", line > 565, in expand > return self.expand(node.children[0], node, itvars_occs, itvar_exp) > File "/usr/local/lib/python2.7/site-packages/coffee/rewriter.py", line > 599, in expand > l_node, l_type = self.expand(node.children[0], node, itvars_occs, > itvar_exp) > File "/usr/local/lib/python2.7/site-packages/coffee/rewriter.py", line > 565, in expand > return self.expand(node.children[0], node, itvars_occs, itvar_exp) > File "/usr/local/lib/python2.7/site-packages/coffee/rewriter.py", line > 608, in expand > raise RuntimeError("Expansion error: unknown node: %s" % str(node)) > RuntimeError: Expansion error: unknown node: (LI_IPK_1_0[k] * > LI_IPJ_1_1[j]) / LI_C_1_0 > > > Thank you. > > Anna. > > > > _______________________________________________ > firedrake mailing list > firedrake(a)imperial.ac.uk<mailto:firedrake@imperial.ac.uk> > https://mailman.ic.ac.uk/mailman/listinfo/firedrake > > _______________________________________________ > firedrake mailing list > firedrake(a)imperial.ac.uk<mailto:firedrake@imperial.ac.uk> > https://mailman.ic.ac.uk/mailman/listinfo/firedrake > > _______________________________________________ > firedrake mailing list > firedrake(a)imperial.ac.uk<mailto:firedrake@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 >

1 0

Re: [firedrake] Problem with Nonlinear solvers after update
by Andrew McRae 22 Dec '14

22 Dec '14

Ugh, one of the recent changes we made was to turn on COFFEE's optimisations by default. Unfortunately, this has led to some teething problems. I assume that all components -- COFFEE, Firedrake, PyOP2, FFC, UFL, and FIAT -- are up to date, and that you're running with a clean cache (from the Firedrake base directory, run "./scripts/firedrake-clean"). If the problem persists, as a short-term workaround, can you add op2.init() parameters["coffee"]["O2"] = False to the start of your file, just after the imports? This will result in sub-optimal performance during the assembly stage, so isn't ideal in the long-term. Fabio, our resident barista, will have some fixes after the new year, I'm sure. Andrew On 22 December 2014 at 13:47, Anna Kalogirou <A.Kalogirou(a)leeds.ac.uk> wrote: > Dear all, > > I very recently updated firedrake, pyop2 and installed coffee, but since > then none of my codes include nonlinear solvers don’t work (they used to > work fine before the update). > > I even tried to run the burgers.py code from the demo and even that > diverges. > > I get something like: > > Traceback (most recent call last): > File "benney-luke.py", line 57, in <module> > phi_solver = solver_phi(phi1, phi1, phi0, eta0, etaR, phi, gamma, dt, > mu, epsilon, is_linear); > File "/Users/matak/Documents/Firedrake/modules/solvers.py", line 26, in > solver_phi > phi_solver = > NonlinearVariationalSolver(phi_problem,solver_parameters={'snes_monitor': > True,'ksp_monitor': True,'snes_linesearch_monitor': True}) > File "/Users/matak/firedrake/firedrake/variational_solver.py", line 97, > in __init__ > form_compiler_parameters=self._problem.form_compiler_parameters) > File "/Users/matak/firedrake/firedrake/assemble.py", line 58, in assemble > form_compiler_parameters=form_compiler_parameters) > File "/Users/matak/firedrake/firedrake/assemble.py", line 79, in > _assemble > kernels = ffc_interface.compile_form(f, "form", > parameters=form_compiler_parameters) > File "/Users/matak/firedrake/firedrake/ffc_interface.py", line 268, in > compile_form > ffc_kernel = FFCKernel(f, name + str(i) + str(j), parameters) > File > "/usr/local/Cellar/python/2.7.8_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PyOP2-0.11.0_148_g8a1955c-py2.7-macosx-10.9-x86_64.egg/pyop2/caching.py", > line 202, in __new__ > obj = make_obj() > File > "/usr/local/Cellar/python/2.7.8_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PyOP2-0.11.0_148_g8a1955c-py2.7-macosx-10.9-x86_64.egg/pyop2/caching.py", > line 192, in make_obj > obj.__init__(*args, **kwargs) > File "/Users/matak/firedrake/firedrake/ffc_interface.py", line 194, in > __init__ > (name, it.integral_type(), it.subdomain_id()), opts, inc), > File > "/usr/local/Cellar/python/2.7.8_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PyOP2-0.11.0_148_g8a1955c-py2.7-macosx-10.9-x86_64.egg/pyop2/backends.py", > line 118, in __call__ > return t(*args, **kwargs) > File > "/usr/local/Cellar/python/2.7.8_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PyOP2-0.11.0_148_g8a1955c-py2.7-macosx-10.9-x86_64.egg/pyop2/caching.py", > line 202, in __new__ > obj = make_obj() > File > "/usr/local/Cellar/python/2.7.8_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PyOP2-0.11.0_148_g8a1955c-py2.7-macosx-10.9-x86_64.egg/pyop2/caching.py", > line 192, in make_obj > obj.__init__(*args, **kwargs) > File > "/usr/local/Cellar/python/2.7.8_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PyOP2-0.11.0_148_g8a1955c-py2.7-macosx-10.9-x86_64.egg/pyop2/base.py", > line 3632, in __init__ > self._code = self._ast_to_c(code, opts) > File > "/usr/local/Cellar/python/2.7.8_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PyOP2-0.11.0_148_g8a1955c-py2.7-macosx-10.9-x86_64.egg/pyop2/host.py", > line 63, in _ast_to_c > ast_handler.plan_cpu(opts) > File "/usr/local/lib/python2.7/site-packages/coffee/plan.py", line 405, > in plan_cpu > loop_opts = _generate_cpu_code(self, **params) > File "/usr/local/lib/python2.7/site-packages/coffee/plan.py", line 241, > in _generate_cpu_code > loop_opt.rewrite(rewrite) > File "/usr/local/lib/python2.7/site-packages/coffee/optimizer.py", line > 110, in rewrite > ew.expand() > File "/usr/local/lib/python2.7/site-packages/coffee/rewriter.py", line > 168, in expand > ee.expand(self.stmt.children[1], self.stmt, itvars_occs, itvar_exp) > File "/usr/local/lib/python2.7/site-packages/coffee/rewriter.py", line > 565, in expand > return self.expand(node.children[0], node, itvars_occs, itvar_exp) > File "/usr/local/lib/python2.7/site-packages/coffee/rewriter.py", line > 567, in expand > l_node, l_type = self.expand(node.children[0], node, itvars_occs, > itvar_exp) > File "/usr/local/lib/python2.7/site-packages/coffee/rewriter.py", line > 565, in expand > return self.expand(node.children[0], node, itvars_occs, itvar_exp) > File "/usr/local/lib/python2.7/site-packages/coffee/rewriter.py", line > 599, in expand > l_node, l_type = self.expand(node.children[0], node, itvars_occs, > itvar_exp) > File "/usr/local/lib/python2.7/site-packages/coffee/rewriter.py", line > 565, in expand > return self.expand(node.children[0], node, itvars_occs, itvar_exp) > File "/usr/local/lib/python2.7/site-packages/coffee/rewriter.py", line > 608, in expand > raise RuntimeError("Expansion error: unknown node: %s" % str(node)) > RuntimeError: Expansion error: unknown node: (LI_IPK_1_0[k] * > LI_IPJ_1_1[j]) / LI_C_1_0 > > > Thank you. > > Anna. > > > > _______________________________________________ > firedrake mailing list > firedrake(a)imperial.ac.uk > https://mailman.ic.ac.uk/mailman/listinfo/firedrake >

3 2

Problem with Nonlinear solvers after update
by Anna Kalogirou 22 Dec '14

22 Dec '14

Dear all, I very recently updated firedrake, pyop2 and installed coffee, but since then none of my codes include nonlinear solvers don’t work (they used to work fine before the update). I even tried to run the burgers.py code from the demo and even that diverges. I get something like: Traceback (most recent call last): File "benney-luke.py", line 57, in <module> phi_solver = solver_phi(phi1, phi1, phi0, eta0, etaR, phi, gamma, dt, mu, epsilon, is_linear); File "/Users/matak/Documents/Firedrake/modules/solvers.py", line 26, in solver_phi phi_solver = NonlinearVariationalSolver(phi_problem,solver_parameters={'snes_monitor': True,'ksp_monitor': True,'snes_linesearch_monitor': True}) File "/Users/matak/firedrake/firedrake/variational_solver.py", line 97, in __init__ form_compiler_parameters=self._problem.form_compiler_parameters) File "/Users/matak/firedrake/firedrake/assemble.py", line 58, in assemble form_compiler_parameters=form_compiler_parameters) File "/Users/matak/firedrake/firedrake/assemble.py", line 79, in _assemble kernels = ffc_interface.compile_form(f, "form", parameters=form_compiler_parameters) File "/Users/matak/firedrake/firedrake/ffc_interface.py", line 268, in compile_form ffc_kernel = FFCKernel(f, name + str(i) + str(j), parameters) File "/usr/local/Cellar/python/2.7.8_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PyOP2-0.11.0_148_g8a1955c-py2.7-macosx-10.9-x86_64.egg/pyop2/caching.py", line 202, in __new__ obj = make_obj() File "/usr/local/Cellar/python/2.7.8_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PyOP2-0.11.0_148_g8a1955c-py2.7-macosx-10.9-x86_64.egg/pyop2/caching.py", line 192, in make_obj obj.__init__(*args, **kwargs) File "/Users/matak/firedrake/firedrake/ffc_interface.py", line 194, in __init__ (name, it.integral_type(), it.subdomain_id()), opts, inc), File "/usr/local/Cellar/python/2.7.8_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PyOP2-0.11.0_148_g8a1955c-py2.7-macosx-10.9-x86_64.egg/pyop2/backends.py", line 118, in __call__ return t(*args, **kwargs) File "/usr/local/Cellar/python/2.7.8_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PyOP2-0.11.0_148_g8a1955c-py2.7-macosx-10.9-x86_64.egg/pyop2/caching.py", line 202, in __new__ obj = make_obj() File "/usr/local/Cellar/python/2.7.8_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PyOP2-0.11.0_148_g8a1955c-py2.7-macosx-10.9-x86_64.egg/pyop2/caching.py", line 192, in make_obj obj.__init__(*args, **kwargs) File "/usr/local/Cellar/python/2.7.8_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PyOP2-0.11.0_148_g8a1955c-py2.7-macosx-10.9-x86_64.egg/pyop2/base.py", line 3632, in __init__ self._code = self._ast_to_c(code, opts) File "/usr/local/Cellar/python/2.7.8_2/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PyOP2-0.11.0_148_g8a1955c-py2.7-macosx-10.9-x86_64.egg/pyop2/host.py", line 63, in _ast_to_c ast_handler.plan_cpu(opts) File "/usr/local/lib/python2.7/site-packages/coffee/plan.py", line 405, in plan_cpu loop_opts = _generate_cpu_code(self, **params) File "/usr/local/lib/python2.7/site-packages/coffee/plan.py", line 241, in _generate_cpu_code loop_opt.rewrite(rewrite) File "/usr/local/lib/python2.7/site-packages/coffee/optimizer.py", line 110, in rewrite ew.expand() File "/usr/local/lib/python2.7/site-packages/coffee/rewriter.py", line 168, in expand ee.expand(self.stmt.children[1], self.stmt, itvars_occs, itvar_exp) File "/usr/local/lib/python2.7/site-packages/coffee/rewriter.py", line 565, in expand return self.expand(node.children[0], node, itvars_occs, itvar_exp) File "/usr/local/lib/python2.7/site-packages/coffee/rewriter.py", line 567, in expand l_node, l_type = self.expand(node.children[0], node, itvars_occs, itvar_exp) File "/usr/local/lib/python2.7/site-packages/coffee/rewriter.py", line 565, in expand return self.expand(node.children[0], node, itvars_occs, itvar_exp) File "/usr/local/lib/python2.7/site-packages/coffee/rewriter.py", line 599, in expand l_node, l_type = self.expand(node.children[0], node, itvars_occs, itvar_exp) File "/usr/local/lib/python2.7/site-packages/coffee/rewriter.py", line 565, in expand return self.expand(node.children[0], node, itvars_occs, itvar_exp) File "/usr/local/lib/python2.7/site-packages/coffee/rewriter.py", line 608, in expand raise RuntimeError("Expansion error: unknown node: %s" % str(node)) RuntimeError: Expansion error: unknown node: (LI_IPK_1_0[k] * LI_IPJ_1_1[j]) / LI_C_1_0 Thank you. Anna.

1 0

Re: [firedrake] borked my update
by David Ham 20 Dec '14

20 Dec '14

On Saturday, December 20, 2014, Rathgeber, Florian < florian.rathgeber(a)imperial.ac.uk> wrote: > > > > I hope this is not a silly question, but in which directory I am > supposed be in when I write “ sudo python setup.py install“ during the > COFFEE installation? > > That should not matter. The only instance where it could matter is if > you're on a file system where the local superuser has no write > permission (which is often the case for NFS mounted file systems). > > I beg to differ. The command as typed will only work from the COFFEE source directory > HTH, > Florian > > > Thanks, > > > > Anna. > > > > > > > > On 19 Dec 2014, at 10:36, Fabio Luporini <fabilupo(a)gmail.com > <javascript:;><mailto:fabilupo@gmail.com <javascript:;>>> wrote: > > > > Some modules are still looking for ast_base which in the new version of > COFFEE has a different name. This may be due to some packages that still > need be updated. Could be ffc, pyop2, or firedrake; are you sure you've > updated and reinstalled all of these packages? > > > > If you run python and try : from coffee import base, what does it say? > > > > Thanks > > > > -- Fabio > > > > Sent from my iPhone > > > > On 19/dic/2014, at 10:20, Anna Kalogirou <A.Kalogirou(a)leeds.ac.uk > <javascript:;><mailto:A.Kalogirou@leeds.ac.uk <javascript:;>>> wrote: > > > > I already installed COFFEE and updated PyOP2, so now there are no COFFEE > leftovers in the PyOP2/pyop2 directory... > > > > > > On 19 Dec 2014, at 10:16, David Ham <David.Ham(a)imperial.ac.uk > <javascript:;><mailto:David.Ham@imperial.ac.uk <javascript:;>><mailto: > David.Ham(a)imperial.ac.uk <javascript:;>>> wrote: > > > > Dear Anna, > > > > My first guess is that you are seeing exactly the same error as Colin so > the same solution should apply. > > > > Basically, COFFEE is now a separate package. This means you need to > install the COFFEE package and remove any left over COFFEE files from your > PyOP2 repository. > > > > You can install COFFEE with: > > > > sudo pip install git+https://github.com/coneoproject/COFFEE.git > > > > and you can remove any leftovers by typing the following in the > PyOP2/pyop2 directory: > > > > rm -r coffee > > > > Regards, > > > > David > > > > > > On 19 December 2014 at 10:09, Anna Kalogirou <A.Kalogirou(a)leeds.ac.uk > <javascript:;><mailto:A.Kalogirou@leeds.ac.uk <javascript:;>><mailto: > A.Kalogirou(a)leeds.ac.uk <javascript:;>>> wrote: > > Hi all, > > > > I just tried to update firedrake on my laptop (including the coffee > installation), and get similar errors as Colin did a few weeks ago. It > seems that there is no module named coffee.ast_base. Colin, how did you fix > the problem after all? > > > > Thank you. > > > > Best, > > > > Anna. > > -- Dr David Ham Departments of Mathematics and Computing Imperial College London http://www.imperial.ac.uk/people/david.ham

2 1