October 2018 - firedrake - imperial.ac.uk

Re: [firedrake] Installation with my own PETSc repository
by Andrew McRae 06 Nov '18

06 Nov '18

Line 5406 _configtest.c:2:10: fatal error: 'mpi.h' file not found On Wed, 31 Oct 2018 at 20:27, Matthew Knepley <knepley(a)gmail.com> wrote: > Thanks folks! I made it farther: > > :/PETSc3/fem$ ~/MacSoftware/bin/python3 ./firedrake-install > --honour-petsc-dir --verbose --no-package-manager > > Running ./firedrake-install --honour-petsc-dir --verbose > --no-package-manager > > Xcode and homebrew installation disabled. Proceeding on the rash > assumption that packaged dependencies are in place. > > Creating firedrake venv in '/PETSc3/fem/firedrake'. > > Cloning firedrake > > > Successfully cloned repository firedrake > > Checking out branch master > > Successfully checked out branch master > > Cloning the dependencies of firedrake > > Cloning ufl > > > Successfully cloned repository ufl > > Checking out branch master > > Successfully checked out branch master > > Cloning fiat > > > Successfully cloned repository fiat > > Checking out branch master > > Successfully checked out branch master > > Cloning FInAT > > > Successfully cloned repository FInAT > > Checking out branch master > > Successfully checked out branch master > > Cloning tsfc > > > Successfully cloned repository tsfc > > Checking out branch master > > Successfully checked out branch master > > Cloning PyOP2 > > > Successfully cloned repository PyOP2 > > Checking out branch master > > Successfully checked out branch master > > Cloning the dependencies of PyOP2 > > Using existing petsc installation > > > Cloning petsc4py > > > Successfully cloned repository petsc4py > > Checking out branch firedrake > > Successfully checked out branch firedrake > > Cloning COFFEE > > > Successfully cloned repository COFFEE > > Checking out branch master > > Successfully checked out branch master > > Installing pip dependencies for petsc4py > > No dependencies found. Skipping. > > Installing pip dependencies for COFFEE > > Installing pip dependencies for ufl > > No dependencies found. Skipping. > > Installing pip dependencies for fiat > > No dependencies found. Skipping. > > Installing pip dependencies for FInAT > > Installing pip dependencies for tsfc > > Installing pip dependencies for PyOP2 > > Traceback (most recent call last): > > File "./firedrake-install", line 1124, in <module> > > pip_requirements(p) > > File "./firedrake-install", line 607, in pip_requirements > > run_pip_install(["-r", "%s/requirements-ext.txt" % package]) > > File "./firedrake-install", line 597, in run_pip_install > > check_call(pipinstall + pipargs) > > File "./firedrake-install", line 435, in check_call > > log.debug(subprocess.check_output(arguments, > stderr=subprocess.STDOUT, env=env).decode()) > > File "/Users/knepley/MacSoftware/lib/python3.6/subprocess.py", line > 336, in check_output > > **kwargs).stdout > > File "/Users/knepley/MacSoftware/lib/python3.6/subprocess.py", line > 418, in run > > output=stdout, stderr=stderr) > > subprocess.CalledProcessError: Command > '['/PETSc3/fem/firedrake/bin/python', '-m', 'pip', 'install', > '--no-binary', 'mpi4py', '-vvv', '-r', 'PyOP2/requirements-ext.txt']' > returned non-zero exit status 1. > > > > Install log saved in firedrake-install.log > > 1:/PETSc3/fem$ vi firedrake-install.log > > > It looks like mpi4py is failing, but I do not know where to look for the > actual error message. Log attached. > > Thanks, > > Matt > > On Tue, Oct 30, 2018 at 12:51 PM Lawrence Mitchell <wence(a)gmx.li> wrote: > >> >> >> > On 30 Oct 2018, at 16:30, Matthew Knepley <knepley(a)gmail.com> wrote: >> > >> > On Fri, Oct 26, 2018 at 11:26 AM Matthew Knepley <knepley(a)gmail.com> >> wrote: >> > Dear Firedrake, >> > >> > Long-time listener, first time mailer. >> > >> > I would like to build using my own repository for PETSc. I am reading >> this page: >> > >> > https://www.firedrakeproject.org/download.html >> > >> > but it is not clear to me how to tell the installer to use mine, rather >> than install its own. >> > Do I just point to it using PETSC_DIR? >> > >> > Its not going to install itself by osmosis. >> >> Aha, you're not subscribed and people don't automatically CC. You would >> normally get a bounce notification asking you to subscribe, but I think >> someone in the past (probably me) added you to a list of addresses that get >> accepted anyway. >> >> Answers from David: >> >> > Nearly: there’s a small piece of idiot-proofing to prevent the unwary >> new user from accidentally using the wrong PETSc. You need to pass >> –honour-petsc-dir (note UK spelling) to firedrake-install >> > >> > You also need to check the install script to make sure your petsc has >> our minimal set of options. I apologise that you have to RTFS for that. >> From memory, the absolute minimum is that you need –download-hdf5 and >> –download-eigen >> >> Nicolas said: >> >> > >> > You need to define PETSC_DIR and PETSC_ARCH, then run the install >> script with the --honour-petsc-dir option. >> > >> > If you call the script with --help, you may also discover a couple of >> options you might want to use (or not) such as --disable-ssh or >> --no-package-manager >> >> And Miklos pointed out: >> >> > You don’t actually need to read the script, >> > >> > firedrake-install --show-petsc-configure-options >> > >> > tells you which options the Firedrake install script passes to PETSc’s >> configure script. It is normally safe add things, but removing things may >> result in harmless test failures (e.g. some solver packages), or in broken >> Firedrake functionality. >> >> This last option doesn't exit the installer after showing them, so it's >> not quite as useful as it seems. >> >> FWIW, we by default build with: >> >> --download-chaco --download-metis --download-parmetis >> --download-scalapack --download-hypre --download-mumps --with-zlib >> --download-netcdf --download-hdf5 --download-pnetcdf --download-exodusii >> --with-debugging=0 --with-shared-libraries=1 >> >> Cheers, >> >> Lawrence >> >> > > -- > 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/> >

7 28

Re: [firedrake] local discontinuity
by Ham, David A 05 Nov '18

05 Nov '18

Dear Floriane, I think you can do this in a manner analogous to the way that Thomasz did fluid-structure interaction. We define a mixed function space with two continuous components: (V_d, V_p). We’re going to use the first component for the solution in the disk and the second component for the solution outside the disk. We further define two DG0 functions I_d and I_p such that I_d is 1 inside the disk and 0 outside, and I_p = 1 – I_d. You can now write the Laplace equation for the two parts using essentially normal Firedrake code except that you multiply the test function by the appropriate indicator function. This ensures that you only actually assemble integral contributions on the correct part of the domain. Now we just need the surface integral over the facets on the edge of the disk. Observe that this is the only place in the domain where both indicator functions are positive. You can write your jump integrals using the *dS measure. You can restrict the intervals to the relevant edges by multiplying the integrand by 4*avg(I_p)*avg(I_d). If W = V_d * V_p and u_d, u_p = TrialFunctions(W) then you can write the jump as: avg(u_d) – avg(u_p) The avg is mathematically unnecessary and does nothing, however in UFL all terminals that appear in dS integrals have to be restricted (you could just as well write u_d(‘+’) – u_p(‘+’)). Finally, when you come to the solve, you need to ensure that all the V_d nodes outside the disk are eliminated, and all the V_p nodes inside the disk too. You do this using the same DirichletBC subclassing trick which is in https://www.firedrakeproject.org/demos/linear_fluid_structure_interaction.p… I realise that is slightly involved, but it should work and the syntax will be fairly clean once you have done it. Feel free to ask for clarifications. Regards, David From: <firedrake-bounces(a)imperial.ac.uk> on behalf of "Floriane Gidel [RPG]" <mmfg(a)leeds.ac.uk> Reply-To: firedrake <firedrake(a)imperial.ac.uk> Date: Friday, 5 October 2018 at 15:33 To: firedrake <firedrake(a)imperial.ac.uk> Subject: [firedrake] local discontinuity Dear Firedrakers, I would like to solve equations in a domain made of two subdomains: - a disk, inside which test functions are continuous (Omega_d) - a plane around the disk, in which test functions are continuous (Omega_p) However, I would like the test functions to be discontinuous across the interface Gamma between the disk and the surrounding plane. This is to solve Laplace equation in each domain, augmented by a dynamic boundary condition on the jump between the two domains, that is Delta u = 0 in Omega_d Delta u = 0 in Omega_p partial_t [u] = ... on Gamma, where [u]= u(+)-u(-) is the jump Is there a way to define a functionspace on the full mesh, with continuous basis functions in each subdomain but discontinuous across the interface ? Or is there any better way to solve this kind of system ? Thank you for your help, Floriane

5 18

Solving nonlinear algebraic equation
by Amireh Mousavi 01 Nov '18

01 Nov '18

Dear Firedrake team, I'm trying to solve Hamilton-Jacobi-Bellman equation. It is a fully nonlinear equation, so after linearization, in each iteration in first we need to determine a parameter which is obtained from solving nonlinear algebraic equation and after that we solve the linear PDE. How can I solve nonlinear algebraic equation in Firedrake? Consider the simple example: mesh = UnitSquareMesh(4, 4)R = FunctionSpace(mesh, "R", 0)Ralpha = Function(R)alpha = variable(Ralpha)A_alpha = as_tensor([[cos(alpha) , -sin(alpha)] , [sin(alpha) , cos(alpha)]])A = Constant([[20 ,1] , [1 , 0.1]])A_alphaA = inner(A_alpha , A) #dA_alphaA = solve(diff(A_alphaA , alpha ), alpha) How can I solve "diff(A_alphaA , alpha)=0" in Firedrake ? Thank you in advance. All the best, Amireh

3 4

Installation with my own PETSc repository
by Matthew Knepley 31 Oct '18

31 Oct '18

Dear Firedrake, Long-time listener, first time mailer. I would like to build using my own repository for PETSc. I am reading this page: https://www.firedrakeproject.org/download.html but it is not clear to me how to tell the installer to use mine, rather than install its own. Do I just point to it using PETSC_DIR? Thanks, Matt -- 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/>

5 6

.at() and PointNotInDomainError
by Floriane Gidel [RPG] 26 Oct '18

26 Oct '18

Dear all, When trying to evaluate a function u on some mesh points with mesh = Mesh('coarse/Omega_coarse.msh') V = FunctionSpace(mesh, "CG",1) V_Vec = VectorFunctionSpace(mesh, "CG", 1) u = Function(V) points = Function(V_vec).interpolate(SpatialCoordinate(mesh)) dat = u.at(points.dat.data_ro) I get the following error: File "lin_dyn_unicell.py", line 117, in <module> dat = u.at(points.dat.data_ro) File "/Users/gidel/Documents/Firedrake/firedrake/src/firedrake/firedrake/function.py", line 585, in at raise PointNotInDomainError(self.function_space().mesh(), points[i].reshape(-1)) firedrake.function.PointNotInDomainError: domain <Mesh #1> does not contain point [200. 66.66666667] How is that possible while both function spaces are defined on the same mesh? Note that in this example I am using the same domain to make sure the points are the same in the two cases, but I used to run this with different meshes and it worked on my laptop with an old firedrake version but not on my new machine on which I installed Firedrake yesterday. Thank you for your feedback, Floriane

3 3

Sanity check
by Jon Hill 24 Oct '18

24 Oct '18

Hi folks, Can I just get a sanity check on this bit of code to check I'm not doing something insane (I probably am) or stupid (equally likely). The problem is a basic diffusion equation where a pile of sediment is eroded and transport by diffusion only. The un-erodable topographic surface (land) + sediment (phi) make up the land surface and the slope over which the sediment is moved. The attached figure might show it more clearly. phi is therefore moved according to the 2nd order (phi+land). However, I've also had to "turn off" diffusion by setting the diffusion coefficient (nu) to zero where there is no sediment or volume is definitely not conserved. The equation is set up using: tiny = 1e-10 surface.interpolate((((land+phi) + land)+abs((land+phi)-land)) / 2) thickness.interpolate(surface-land) limit.interpolate((surface - land)/(surface-land+tiny)) F = (inner((phi - phi_)/timestep, v) + limit*nu*inner(grad(phi+land), grad(v)))*dx And then simply solved: solve(F == 0, phi) in the time loop, updating limit and surface as we go. Hopefully, the above is understandable, but if anyone has any opinions on my (in)sanity, then I'm happy to hear them! Cheers, J. -- I'm currently in Australia and replies will not be in standard UK business hours. Please do not feel you have to respond outside your normal work hours. Dr Jon Hill Senior Lecturer in Physical Geography Department of Environment and Geography University of York M: +44(0)7748254812 P: +44(0)1904 324480 Twitter: @jonxhill Web: https://jonxhill.wordpress.com/ Web: https://envmodellinggroup.github.io/

2 1

problems updating
by Francis Poulin 23 Oct '18

23 Oct '18

Hello, I have tried updating firedrake for the first time in months but it seems to fail. The output suggests doing a firedrake-clean, which I have tried but that seems to have failed. I am attaching the log file. Can anyone suggest something that I might be able to try to get this working? Cheers, Francis ------------------ Francis Poulin Associate Dean, Undergraduate Studies Professor Department of Applied Mathematics University of Waterloo email: fpoulin(a)uwaterloo.ca Web: https://uwaterloo.ca/poulin-research-group/ Telephone: +1 519 888 4567 x32637 [X]

3 6

Re: [firedrake] Fix for problem with Monge-Ampere demo
by Ham, David A 11 Oct '18

11 Oct '18

Hi Dave, This is really rather odd. Do you happen to have the exception lying around somewhere? E.g. does the problem occur when the demo is run, or maybe when pylit is run on the input rst file? We ought to be able to fix this within Firedrake in an environment-independent way but it depends a bit on where the issue occurs. Regards, David From: <firedrake-bounces(a)imperial.ac.uk> on behalf of "Acreman, David" <D.M.Acreman(a)exeter.ac.uk> Reply-To: firedrake <firedrake(a)imperial.ac.uk> Date: Thursday, 11 October 2018 at 11:28 To: firedrake <firedrake(a)imperial.ac.uk> Subject: [firedrake] Fix for problem with Monge-Ampere demo Hi all, I ran into a slightly odd problem running one of the demos and thought it would be useful to post the solution in case anyone else is having the same problem. The problem was triggered by a non-ascii character in the Monge-Ampere demo (ma-demo) which showed up as a failure when running the test suite. The resulting error message is: UnicodeDecodeError: ’ascii’ codec can’t decode byte 0xc3 in position 41: ordinal not in range(128) This can be fixed by setting the LC_ALL environment variable e.g. for bash-like shells export LC_ALL=en_US.utf-8 You can check the encoding by running python3 -c 'import sys; print(sys.stdin.encoding)’ which should report "UTF-8” with the environment variable set. The encoding usually seems to work correctly but didn’t on our local cluster. Regards, Dave Dr. David Acreman, Senior Research Fellow, College of Engineering, Mathematics and Physical Sciences, University of Exeter, Stocker Road, Exeter EX4 4QL

2 2

Fix for problem with Monge-Ampere demo
by Acreman, David 11 Oct '18

11 Oct '18

Hi all, I ran into a slightly odd problem running one of the demos and thought it would be useful to post the solution in case anyone else is having the same problem. The problem was triggered by a non-ascii character in the Monge-Ampere demo (ma-demo) which showed up as a failure when running the test suite. The resulting error message is: UnicodeDecodeError: ’ascii’ codec can’t decode byte 0xc3 in position 41: ordinal not in range(128) This can be fixed by setting the LC_ALL environment variable e.g. for bash-like shells export LC_ALL=en_US.utf-8 You can check the encoding by running python3 -c 'import sys; print(sys.stdin.encoding)’ which should report "UTF-8” with the environment variable set. The encoding usually seems to work correctly but didn’t on our local cluster. Regards, Dave Dr. David Acreman, Senior Research Fellow, College of Engineering, Mathematics and Physical Sciences, University of Exeter, Stocker Road, Exeter EX4 4QL

1 0

local discontinuity
by Floriane Gidel [RPG] 05 Oct '18

05 Oct '18

Dear Firedrakers, I would like to solve equations in a domain made of two subdomains: - a disk, inside which test functions are continuous (Omega_d) - a plane around the disk, in which test functions are continuous (Omega_p) However, I would like the test functions to be discontinuous across the interface Gamma between the disk and the surrounding plane. This is to solve Laplace equation in each domain, augmented by a dynamic boundary condition on the jump between the two domains, that is Delta u = 0 in Omega_d Delta u = 0 in Omega_p partial_t [u] = ... on Gamma, where [u]= u(+)-u(-) is the jump Is there a way to define a functionspace on the full mesh, with continuous basis functions in each subdomain but discontinuous across the interface ? Or is there any better way to solve this kind of system ? Thank you for your help, Floriane

1 0