April 2016 - firedrake - imperial.ac.uk

Re: [firedrake] problems on my mac os x
by David Ham 09 Apr '16

09 Apr '16

Hi Francis, If you ran firedrake-update with PYTHONPATH set then it's possible six was found in your PYTHONPATH and therefore not installed by the upgrade script. Once you unset PYTHONPATH then it would not be found. Try rerunning firedrake-install with PYTHONPATH unset so hopefully six will get installed (six is the Python 2 to 3 compatibility module) Regards, David On Sat, 9 Apr 2016 at 17:34 Francis Poulin <fpoulin(a)uwaterloo.ca> wrote: > Minor correction. > > I remembered from previous installations that I should clear my PYTHONPATH > before starting firedrake. I did that and then got a slightly different > error. > > >>> import firedrake > > Traceback (most recent call last): > > File "<stdin>", line 1, in <module> > > File > "/Users/fpoulin/software/firedrake/lib/python2.7/site-packages/firedrake/__init__.py", > line 19, in <module> > > import ufl > > File > "/Users/fpoulin/software/firedrake/lib/python2.7/site-packages/ufl/__init__.py", > line 193, in <module> > > from ufl.utils.sequences import product > > File > "/Users/fpoulin/software/firedrake/lib/python2.7/site-packages/ufl/utils/sequences.py", > line 21, in <module> > > from six.moves import zip > > ImportError: No module named six.moves > > It can import petsc4y and other stuff, just not sure what six.moves is and > why the module is not found. > > Francis >

2 1

problems on my mac os x
by Francis Poulin 09 Apr '16

09 Apr '16

Hello, I recently reinstalled firedrake on my Mac OS X because I saw there was an update recently. Unfortunately, now I am not able to import firedrake. >>> import firedrake Traceback (most recent call last): File "<stdin>", line 1, in <module> File "firedrake/__init__.py", line 31, in <module> from firedrake.assemble import * File "firedrake/assemble.py", line 8, in <module> from firedrake import assembly_cache File "firedrake/assembly_cache.py", line 41, in <module> from firedrake.parameters import parameters File "firedrake/parameters.py", line 4, in <module> from tsfc.constants import default_parameters File "/Users/fpoulin/software/firedrake/lib/python2.7/site-packages/tsfc/__init__.py", line 3, in <module> from tsfc.driver import compile_form # noqa File "/Users/fpoulin/software/firedrake/lib/python2.7/site-packages/tsfc/driver.py", line 10, in <module> from tsfc.quadrature import create_quadrature, QuadratureRule File "/Users/fpoulin/software/firedrake/lib/python2.7/site-packages/tsfc/quadrature.py", line 31, in <module> from FIAT.reference_element import UFCInterval, UFCTriangle, UFCTetrahedron, \ ImportError: cannot import name FiredrakeQuadrilateral I am sorry for the bother but any help would be greatly appreciated. I also want to update it on my ubuntu machine but thought I should do one at a time. Cheers, Francis

2 2

dynamical core name
by Colin Cotter 08 Apr '16

08 Apr '16

Dear all, The Firedrake dynamical core is currently un-named and we have just been referring to the project as dcore. I have now come up with a plausible name, Gusto. If nobody has a better suggestion in the next 7 days, I'll rename the project to Gusto and start to make a project site under firedrakeproject.org . Warning: I share David's dislike of acronyms but not his tendency towards nerdy D&D-type mythical creatures. cheers --cjc -- http://www.imperial.ac.uk/people/colin.cotter www.cambridge.org/9781107663916

2 1

Integral is missing an integration domain.
by Fryderyk Wilczynski 07 Apr '16

07 Apr '16

Dear all, I'm experiencing some problems setting up weak forms for the system of equations I'm trying to solve. Here's what I'm trying to do: ----- from firedrake import * # Model parameters dt = 5.0e-04 # time step Cn=0.01 # Cahn We=0.45 # Weber Pe=4.5 # Peclet Re=100 # Reynolds g_switch = 0 # switch for gravity 1 = ON, 0 = Off Fr2=0.1 # Froude number squared beta = 2 # stabilising term rho1 = Constant(1) rho2 = Constant(1) drho_hat = Constant(0) drho_hat.assign(0.5*(1 - (rho2/rho1))) # d rho_hat / d phi alpha = (rho2-rho1)/(rho2+rho1) # Create mesh and define function spaces mesh = Mesh("square_mesh.msh") P1 = FunctionSpace(mesh, "Lagrange", 1) P2 = VectorFunctionSpace(mesh, "Lagrange", 2) # Taylor-Hood elements for velocity and pressure, P1-P1 elements for phi and mu W = P2*P1*P1*P1 # Define test functions chi, theta, omega, psi = TestFunctions(W) # Define functions u = Function(W) # current solution u0 = Function(W) # solution from previous converged step # Split mixed functions v, p, phi, mu = split(u) v0, p0, phi0, mu0 = split(u0) # Compute the chemical potential df/dphi phi = variable(phi) f = 0.25*(phi**2 - 1)**2 dfdphi = diff(f, phi) # create a functions for the two definitions of density rho_hat = Function(P1) rho = Function(P1) ex, ey = unit_vectors(2) # Weak statement of the equations a1 = (1/dt)*omega*phi*dx + omega*div(v*phi0)*dx + (1/Pe)*inner(grad(omega),grad(mu))*dx a2 = psi*mu*dx - (beta/Cn)*phi*psi*dx - Cn*inner(grad(phi),grad(psi))*dx - (1/rho_hat)*drho_hat*psi*p*dx a3 = rho*(1/dt)*inner(chi, v)*dx + rho*inner(dot(v0,nabla_grad(v)), chi)*dx +(1/We)*(-p*div(chi) + inner(chi, phi*grad(mu))-(1/rho_hat)*drho_hat*p*inner(chi, grad(phi0)) )*dx + (1/Re)*inner((nabla_grad(v)[i,j] + nabla_grad(v)[j,i]), nabla_grad(chi)[i,j] )*dx - (2/(3*Re))*div(v)*div(chi)*dx a4 = theta*div(v)*dx + (alpha/Pe)*inner(grad(theta), grad(mu))*dx a = a1 + a2 + a3 + a4 L1 = (1/dt)*omega*phi0*dx L2 = - (beta/Cn)*phi0*psi*dx + (1/Cn)*dfdphi*psi*dx L3 = - (g_switch/Fr2)*rho_hat*inner(chi,ey)*dx + rho*(1/dt)*inner(chi, v0)*dx L = L1 + L2 + L3 v0, p0, phi0, mu0 = u0.split() # Initial conditions for phi and velocity phi_ini = Expression('''1 - tanh( ( sqrt( pow(x[0]-x0, 2)+pow(x[1]-y0, 2) ) - r0 )/C*sqrt(2) )- tanh( ( sqrt( pow(x[0]-x1, 2)+pow(x[1]-y1, 2) ) - r1 )/C*sqrt(2) )''', x0=0.4,y0=0.5,r0 = 0.25,x1=0.78,y1=0.5,r1 = 0.1,C=0.01 ) phi0.interpolate(phi_ini) vel_ini = Expression(("0.0", "0.0")) v0.interpolate(vel_ini) # initialise density rho.interpolate( Expression('''0.5*((1+phi_ini) + (rho2/rho1)*(1-phi_ini))''') ) outfile = File("./Results_NSCH/nsch_res.pvd") t = 0 tend = 1.0 outfile.write(u0) while t <= tend: t+=dt ; print t # compute density using algebraic expression rho_hat = 0.5*( (1+phi) + (rho2/rho1)*(1-phi)) rho_hat.interpolate(Expression(' 0.5*( (1+phi0) + (rho2/rho1)*(1-phi0))')) # Solve the system of equations solve(a == L, u) # update rho rho -= dt*div(rho*v0) u0.assign(u) outfile.write(u0) ----- When I try to run the code I get an exception: This integral is missing an integration domain. Traceback (most recent call last): File "navier_stokes_CH.py", line 54, in <module> a3 = rho*(1/dt)*inner(chi, v)*dx + rho*inner(dot(v0,nabla_grad(v)), chi)*dx +(1/We)*(-p*div(chi) + inner(chi, phi*grad(mu))-(1/rho_hat)*drho_hat*p*inner(chi, grad(phi0)) )*dx + (1/Re)*inner((nabla_grad(v)[i,j] + nabla_grad(v)[j,i]), nabla_grad(chi)[i,j] )*dx - (2/(3*Re))*div(v)*div(chi)*dx File "/home/cserv1_a/apps/install/firedrake/2016-03-el7/firedrake/lib/python2.7/site-packages/ufl/measure.py", line 398, in __rmul__ error("This integral is missing an integration domain.") File "/home/cserv1_a/apps/install/firedrake/2016-03-el7/firedrake/lib/python2.7/site-packages/ufl/log.py", line 158, in error raise self._exception_type(self._format_raw(*message)) ufl.log.UFLException: This integral is missing an integration domain. Any help on fixing that would be greatly appreciated I assume that this is not the only bug in my code, so I'd also appreciate other corrections that need to be made. I have attached a pdf with a description of exactly what I want to do. Thanks, Fryderyk

4 11

Re: [firedrake] pytest error in parallel
by Hiroe Yamazaki 06 Apr '16

06 Apr '16

Hi Lawrence, I added a path to the nonlinear slice code in my PYTHONPATH and passed all the tests! I still have 1 pytest-warnings. Should I care about it? All the best, Hiroe On 6 April 2016 at 09:52, Mitchell, Lawrence <lawrence.mitchell(a)imperial.ac.uk> wrote: > Dear Hiroe, > > On 05/04/16 17:03, Hiroe Yamazaki wrote: >> Dear All, >> >> I re-installed firedrake with the install script. After it has been >> installed, I get the following error when I do a parrallel test with >> pytest. >> >> E CalledProcessError: Command 'mpiexec -n 2 >> /home/hyamazak/firedrake/bin/python -c "import >> test_nonlinear_gw_nh_novort; >> test_nonlinear_gw_nh_novort.test_total_energy_parallel()"' returned >> non-zero exit status 1 >> >> How can I fix it? The full error message is found here: >> >> https://gist.github.com/anonymous/b87288483d6b1d22543e267c9142c315 > > It looks like the problem is to do with the spawned subprocess not > finding your nonlinear slice code. Is this package available in your > PYTHONPATH, or do you somehow rely on running the tests from a > particular directory? > > Cheers, > > Lawrence >

2 1

pytest error in parallel
by Hiroe Yamazaki 06 Apr '16

06 Apr '16

Dear All, I re-installed firedrake with the install script. After it has been installed, I get the following error when I do a parrallel test with pytest. E CalledProcessError: Command 'mpiexec -n 2 /home/hyamazak/firedrake/bin/python -c "import test_nonlinear_gw_nh_novort; test_nonlinear_gw_nh_novort.test_total_energy_parallel()"' returned non-zero exit status 1 How can I fix it? The full error message is found here: https://gist.github.com/anonymous/b87288483d6b1d22543e267c9142c315 All the best, Hiroe

2 1

Re: [firedrake] pytest error in parallel
by Hiroe Yamazaki 05 Apr '16

05 Apr '16

Oh yes, they are the tests for my slicemodel. > The error seems to be "ImportError: No module named nonlinear_slice", so do > you not get this in serial as well? No, that's the weird thing. I pass all the serial tests but fail the same tests with @pytest.mark.parallel(nprocs=2). I get the same errors on my macbook and Ubuntu computer. Hiroe On 5 April 2016 at 17:27, Andrew McRae <A.T.T.McRae(a)bath.ac.uk> wrote: > Worth mentioning that this is an error when running tests in Hiroe's > repository, not in the Firedrake test suite. > > The error seems to be "ImportError: No module named nonlinear_slice", so do > you not get this in serial as well? > > On 5 April 2016 at 17:03, Hiroe Yamazaki <h.yamazaki(a)imperial.ac.uk> wrote: >> >> Dear All, >> >> I re-installed firedrake with the install script. After it has been >> installed, I get the following error when I do a parrallel test with >> pytest. >> >> E CalledProcessError: Command 'mpiexec -n 2 >> /home/hyamazak/firedrake/bin/python -c "import >> test_nonlinear_gw_nh_novort; >> test_nonlinear_gw_nh_novort.test_total_energy_parallel()"' returned >> non-zero exit status 1 >> >> How can I fix it? The full error message is found here: >> >> https://gist.github.com/anonymous/b87288483d6b1d22543e267c9142c315 >> >> All the best, >> Hiroe >> >> _______________________________________________ >> firedrake mailing list >> firedrake(a)imperial.ac.uk >> https://mailman.ic.ac.uk/mailman/listinfo/firedrake > >

1 0

Re: [firedrake] pytest error in parallel
by Andrew McRae 05 Apr '16

05 Apr '16

Worth mentioning that this is an error when running tests in Hiroe's repository, not in the Firedrake test suite. The error seems to be "ImportError: No module named nonlinear_slice", so do you not get this in serial as well? On 5 April 2016 at 17:03, Hiroe Yamazaki <h.yamazaki(a)imperial.ac.uk> wrote: > Dear All, > > I re-installed firedrake with the install script. After it has been > installed, I get the following error when I do a parrallel test with > pytest. > > E CalledProcessError: Command 'mpiexec -n 2 > /home/hyamazak/firedrake/bin/python -c "import > test_nonlinear_gw_nh_novort; > test_nonlinear_gw_nh_novort.test_total_energy_parallel()"' returned > non-zero exit status 1 > > How can I fix it? The full error message is found here: > > https://gist.github.com/anonymous/b87288483d6b1d22543e267c9142c315 > > All the best, > Hiroe > > _______________________________________________ > firedrake mailing list > firedrake(a)imperial.ac.uk > https://mailman.ic.ac.uk/mailman/listinfo/firedrake >

1 0