We probably need to think about the preconditioner. First please could you confirm that the timing is dominated by the linear solver? On Wednesday, 5 October 2016, Onno Bokhove <O.Bokhove@leeds.ac.uk> wrote:
Great news! Curious about the next step ...!?!
------------------------------ *From:* firedrake-bounces@imperial.ac.uk <javascript:_e(%7B%7D,'cvml','firedrake-bounces@imperial.ac.uk');> < firedrake-bounces@imperial.ac.uk <javascript:_e(%7B%7D,'cvml','firedrake-bounces@imperial.ac.uk');>> on behalf of Anna Kalogirou <a.kalogirou@leeds.ac.uk <javascript:_e(%7B%7D,'cvml','a.kalogirou@leeds.ac.uk');>> *Sent:* Wednesday, October 5, 2016 6:43 PM *To:* firedrake@imperial.ac.uk <javascript:_e(%7B%7D,'cvml','firedrake@imperial.ac.uk');> *Subject:* Re: [firedrake] Mixed system
Dear all,
So now that the real function spaces are up and running (thanks for this!) I can confirm that my code runs without errors and gives the same result as the one I had before, which was using linear algebra techniques to take care of the constraint enforced with a Lagrange multiplier.
However, I noticed that the old code <https://bitbucket.org/annakalog/buoy2d/src> (without mixed function spaces) takes 27.5 minutes for a given final time T, while the new <https://bitbucket.org/annakalog/buoy2d/src/a1760633fb44f99b3b5ba4d09682b7c76c856098/Mixed%20system/?at=master> code takes almost twice as much to run up to the same T, namely 47.5 minutes.
Any ideas why this is happening?
Best wishes,
Anna.
On 16/09/16 17:22, David Ham wrote:
Hi Anna,
Are you certain that the test you are running is the one you pointed to on the web. Line 89 from your error message does not appear in the version on bitbucket, and the last commit was 4 days ago.
Regards,
David
On Fri, 16 Sep 2016 at 17:03 Anna Kalogirou <A.Kalogirou@leeds.ac.uk <javascript:_e(%7B%7D,'cvml','A.Kalogirou@leeds.ac.uk');>> wrote:
When I run
firedrake-zenodo -m "Versions I am using now”
I get
Failed to retrieve git hash for optional component 'SLOPE', continuing without it
The file firedrake.json contains the following (up to ufl it’s the same as yours):
{"fiat": "1bc440ca34d4b127897cf5a36d63bf8a41a3d55c", "COFFEE": " 8e03328d659fd8d8871bf5b7bc3f08b769f0ab9d", "ufl": " 2fc6d00f7e4b5e0639c6d01f27f9b2bbc0d4696e", "firedrake": " 71cf933f05508427287dc6464e15d2d7adf1b179", "tsfc": " f867938c58f475f1cd332421e2bf14cffa2f5aef", "petsc4py": " 0b590afa9d3c42e0b4f863061b6e082d6f0227f6", "petsc": " 21143a8c817f512a23f8e18da648891eb40f5217", "PyOP2": " d9ea13e1e17fcb1e05adcf1db31bd8190974520d", "message": "Versions I am using now"}
On 16 Sep 2016, at 16:49, David Ham <David.Ham@imperial.ac.uk <javascript:_e(%7B%7D,'cvml','David.Ham@imperial.ac.uk');>> wrote:
OK. Let's do some sanity checking.
1. Please confirm that the current firedrake checkout is in developer mode and has always been in developer mode (I can see that it is in developer mode from the error message but I want to make sure there cannot be stale files around from a non-developer mode past).
2. Run firedrake-clean and check that the error still occurs.
2. Please run the following command:
firedrake-zenodo -m "Versions I am using now"
This will create a file firedrake.json . Please check that its contents exactly match what is below:
{"fiat": "1bc440ca34d4b127897cf5a36d63bf8a41a3d55c", "COFFEE": " 8e03328d659fd8d8871bf5b7bc3f08b769f0ab9d", "ufl": " 2fc6d00f7e4b5e0639c6d01f27f9b2bbc0d4696e", "firedrake": " 71cf933f05508427287dc6464e15d2d7adf1b179", "tsfc": " f867938c58f475f1cd332421e2bf14cffa2f5aef", "petsc4py": " 0b590afa9d3c42e0b4f863061b6e082d6f0227f6", "petsc": " 21143a8c817f512a23f8e18da648891eb40f5217", "PyOP2": " d9ea13e1e17fcb1e05adcf1db31bd8190974520d", "message": "Versions I am using now"}
On Fri, 16 Sep 2016 at 16:28 Anna Kalogirou <A.Kalogirou@leeds.ac.uk <javascript:_e(%7B%7D,'cvml','A.Kalogirou@leeds.ac.uk');>> wrote:
That is the same version I have.
On 16 Sep 2016, at 16:23, David Ham <David.Ham@imperial.ac.uk <javascript:_e(%7B%7D,'cvml','David.Ham@imperial.ac.uk');>> wrote:
Hi Anna,
Please confirm the git revision of your PyOP2. You can retrieve this by typing the following command in the PyOP2 directory:
git rev-parse HEAD
You should have: d9ea13e1e17fcb1e05adcf1db31bd8190974520d
Regards,
David
On Fri, 16 Sep 2016 at 15:59 Anna Kalogirou <A.Kalogirou@leeds.ac.uk <javascript:_e(%7B%7D,'cvml','A.Kalogirou@leeds.ac.uk');>> wrote:
Hi David,
Thank you for the reply. Indeed, the problem was with the virtualenv; I had to deactivate and activate it again. Unfortunately I still get errors, which are probably due to the firedrake I have installed (since you said that it works on your computer):
File "buoy-swe.py", line 89, in <module> F_solver = solver_F(phi0_5, eta1, mu0_5, I, w, phi0, eta0, Z0, W0, step_b, etaR, phi_t, eta_t, mu_t, I_t, v1, v2, v3, v4, dt, Hb, H0, L, dR_dt, g, rho, Mass, solvers_print); File "/Users/matak/Documents/Simulations/Firedrake/Ship/Modules/Mixed system/solvers.py", line 41, in solver_F F_solver = LinearVariationalSolver(F_problem, solver_parameters=solvers_print) File "/Users/matak/firedrake/src/firedrake/firedrake/variational_solver.py", line 264, in __init__ super(LinearVariationalSolver, self).__init__(*args, **kwargs) File "/Users/matak/firedrake/src/firedrake/firedrake/variational_solver.py", line 127, in __init__ ctx = solving_utils._SNESContext(problem) File "/Users/matak/firedrake/src/firedrake/firedrake/solving_utils.py", line 107, in __init__ for problem in problems) File "/Users/matak/firedrake/src/firedrake/firedrake/solving_utils.py", line 107, in <genexpr> for problem in problems) File "/Users/matak/firedrake/src/firedrake/firedrake/assemble.py", line 67, in assemble inverse=inverse, nest=nest) File "<decorator-gen-294>", line 2, in _assemble File "/Users/matak/firedrake/src/firedrake/firedrake/utils.py", line 62, in wrapper return f(*args, **kwargs) File "/Users/matak/firedrake/src/firedrake/firedrake/assemble.py", line 179, in _assemble nest=nest) File "/Users/matak/firedrake/src/PyOP2/pyop2/backends.py", line 118, in __call__ return t(*args, **kwargs) File "/Users/matak/firedrake/src/PyOP2/pyop2/caching.py", line 122, in __new__ args, kwargs = cls._process_args(*args, **kwargs) File "<decorator-gen-258>", line 2, in _process_args File "/Users/matak/firedrake/src/PyOP2/pyop2/utils.py", line 130, in wrapper return f(*args, **kwargs) File "/Users/matak/firedrake/src/PyOP2/pyop2/base.py", line 3591, in _process_args if not (pair[0].toset == dsets[0].set and File "/Users/matak/firedrake/src/PyOP2/pyop2/utils.py", line 64, in __get__ obj.__dict__[self.__name__] = result = self.fget(obj) File "/Users/matak/firedrake/src/PyOP2/pyop2/base.py", line 3372, in toset m.toset for m in self._maps)) File "/Users/matak/firedrake/src/PyOP2/pyop2/caching.py", line 156, in __new__ obj = make_obj() File "/Users/matak/firedrake/src/PyOP2/pyop2/caching.py", line 137, in make_obj obj.__init__(*args, **kwargs) File "/Users/matak/firedrake/src/PyOP2/pyop2/base.py", line 977, in __init__ "All components of a MixedSet must have the same number of layers." AssertionError: All components of a MixedSet must have the same number of layers.
On 16 Sep 2016, at 15:47, David Ham <David.Ham@imperial.ac.uk <javascript:_e(%7B%7D,'cvml','David.Ham@imperial.ac.uk');>> wrote:
Hi Anna,
Your code works for me (in the sense that it appears to run without error. I have no idea whether the answer is right ;)
Please also note that Miklos is correct: your error is definitely from a global firedrake install. You can tell because the paths start /usr/local. Your previous error messages are from a local install. You can tell that because the paths start /Users/matak . Are you sure your virtualenv is active?
Regards,
David
On Fri, 16 Sep 2016 at 15:16 Anna Kalogirou <A.Kalogirou@leeds.ac.uk <javascript:_e(%7B%7D,'cvml','A.Kalogirou@leeds.ac.uk');>> wrote:
During the past few days I have been working on my laptop (not the desktop at Leeds) and I don’t believe there are multiple installations of firedrake on this machine.
Can someone please try to run my code <https://bitbucket.org/annakalog/buoy2d/src/673fef3466b06c98ac11f5a120ff42117679d41e/Mixed%20system/?at=master> so that we know for sure if the problem is in my code or related with the installation? Thanks.
On 16 Sep 2016, at 14:54, Colin Cotter <colin.cotter@imperial.ac.uk <javascript:_e(%7B%7D,'cvml','colin.cotter@imperial.ac.uk');>> wrote:
Ah yes, that tallies with Daniel Ruprecht's reports that Firedrake is part of the standard installation at Leeds!
On 16 September 2016 at 14:53, Homolya, Miklós < m.homolya14@imperial.ac.uk> wrote:
It seems that the Firedrake you're importing is a system wide installation. No wonder that updating your own installation doesn't fix it.
Get Outlook for Android <https://aka.ms/ghei36>
On Thu, Sep 15, 2016 at 8:33 PM +0100, "Anna Kalogirou" < A.Kalogirou@leeds.ac.uk> wrote:
Ok thanks, I get the error below:
Traceback (most recent call last):
File "buoy-swe.py", line 6, in <module> from firedrake import * File "/usr/local/lib/python2.7/site-packages/firedrake/__init__.py", line 30, in <module> from assemble import * File "/usr/local/lib/python2.7/site-packages/firedrake/assemble.py", line 8, in <module> import assembly_cache File "/usr/local/lib/python2.7/site-packages/firedrake/assembly_cache.py", line 40, in <module> import function File "/usr/local/lib/python2.7/site-packages/firedrake/function.py", line 9, in <module> import assemble_expressions File "/usr/local/lib/python2.7/site-packages/firedrake/assemble_expressions.py", line 15, in <module> import functionspace File "/usr/local/lib/python2.7/site-packages/firedrake/functionspace.py", line 10, in <module> import dmplex ImportError: dlopen(/usr/local/lib/python2.7/site-packages/firedrake/dmplex.so, 2): Library not loaded: /usr/local/lib/python2.7/site- packages/petsc/lib/libpetsc.3.05.2.dylib Referenced from: /usr/local/lib/python2.7/site- packages/firedrake/dmplex.so Reason: image not found
On 15 Sep 2016, at 18:21, Andrew McRae <A.T.T.McRae@bath.ac.uk> wrote:
To give credit where it's due,
s/Andrew's patch/Miklós's patch/
On 15 September 2016 at 18:16, David Ham <David.Ham@imperial.ac.uk> wrote:
I suspect Andrew is correct. However the deeper issue is that you edited the installed file, which is never a safe thing to do. One must always edit the file in the source directory and then (if necessary) re-install.
Anyway, I have committed Andrew's patch to the repository, so hopefully the problem will go away if you do a git update.
Regards,
David
On Wed, 14 Sep 2016 at 18:39 Andrew McRae <A.T.T.McRae@bath.ac.uk> wrote:
Probably related to installing Firedrake in development mode or not.
I can't help any further here; I don't know if someone else can.
On 14 September 2016 at 16:13, Anna Kalogirou <A.Kalogirou@leeds.ac.uk
wrote:
Ok, so I replace line 3366 in file “/Users/matak/firedrake/lib/ python2.7/site-packages/pyop2/base.py” with the suggested patch, but I now get the error below.
Note that the error I get comes from file firedrake/lib/python2.7/ site-packages/pyop2/base.py while in the reported issue the error comes from firedrake/src/PyOP2/ pyop2/base.py so the two might not be related after all.
Thanks, Anna.
Traceback (most recent call last): File "buoy-swe.py", line 89, in <module> F_solver = solver_F(phi0_5, eta1, mu0_5, I, w, phi0, eta0, Z0, W0, step_b, etaR, phi_t, eta_t, mu_t, I_t, v1, v2, v3, v4, dt, Hb, H0, L, dR_dt, g, rho, Mass, solvers_print); File "/Users/matak/Documents/Simulations/Firedrake/Ship/Modules/Mixed system/solvers.py", line 41, in solver_F F_solver = LinearVariationalSolver(F_problem, solver_parameters=solvers_print) File "/Users/matak/firedrake/lib/python2.7/site-packages/ firedrake/variational_solver.py", line 264, in __init__ super(LinearVariationalSolver, self).__init__(*args, **kwargs) File "/Users/matak/firedrake/lib/python2.7/site-packages/ firedrake/variational_solver.py", line 127, in __init__ ctx = solving_utils._SNESContext(problem) File "/Users/matak/firedrake/lib/python2.7/site-packages/firedrake/solving_utils.py", line 107, in __init__ for problem in problems) File "/Users/matak/firedrake/lib/python2.7/site-packages/firedrake/solving_utils.py", line 107, in <genexpr> for problem in problems) File "/Users/matak/firedrake/lib/python2.7/site-packages/firedrake/assemble.py", line 67, in assemble inverse=inverse, nest=nest) File "<decorator-gen-294>", line 2, in _assemble File "/Users/matak/firedrake/lib/python2.7/site-packages/firedrake/utils.py", line 62, in wrapper return f(*args, **kwargs) File "/Users/matak/firedrake/lib/python2.7/site-packages/firedrake/assemble.py", line 179, in _assemble nest=nest) File "/Users/matak/firedrake/lib/python2.7/site-packages/pyop2/backends.py", line 118, in __call__ return t(*args, **kwargs) File "/Users/matak/firedrake/lib/python2.7/site-packages/pyop2/caching.py", line 122, in __new__ args, kwargs = cls._process_args(*args, **kwargs) File "<decorator-gen-258>", line 2, in _process_args File "/Users/matak/firedrake/lib/python2.7/site-packages/pyop2/utils.py", line 130, in wrapper return f(*args, **kwargs) File "/Users/matak/firedrake/lib/python2.7/site-packages/pyop2/base.py", line 3592, in _process_args if not (pair[0].toset == dsets[0].set and File "/Users/matak/firedrake/lib/python2.7/site-packages/pyop2/utils.py", line 64, in __get__ obj.__dict__[self.__name__] = result = self.fget(obj) File "/Users/matak/firedrake/lib/python2.7/site-packages/pyop2/base.py", line 3373, in toset m.toset for m in self._maps)) File "/Users/matak/firedrake/lib/python2.7/site-packages/pyop2/caching.py", line 156, in __new__ obj = make_obj() File "/Users/matak/firedrake/lib/python2.7/site-packages/pyop2/caching.py", line 137, in make_obj obj.__init__(*args, **kwargs) File "/Users/matak/firedrake/lib/python2.7/site-packages/pyop2/base.py", line 977, in __init__ "All components of a MixedSet must have the same number of layers." AssertionError: All components of a MixedSet must have the same number of layers.
**********************************************************
Dr Anna Kalogirou Research Fellow
Department of Applied Mathematics University of Leeds
Web: http://www1.maths.leeds.ac.uk/~matak/
Youtube: https://www.youtube.com/channel/UCUZAYHtVoiMqepQflisp66g
********************************************************** Leverhulme Trust Early Career Fellow 2017-2019
**********************************************************
On 14 Sep 2016, at 16:00, Andrew McRae <A.T.T.McRae@bath.ac.uk> wrote:
Not as a branch, but you can change PyOP2 locally using the code at the bottom of that post, "Suggested patch"
On 14 September 2016 at 15:54, Anna Kalogirou <A.Kalogirou@leeds.ac.uk
wrote:
Hi Andrew,
Yes indeed, the two might be related. Was there a fix for the issue reported on github?
Best, Anna.
**********************************************************
Dr Anna Kalogirou Research Fellow
Department of Applied Mathematics University of Leeds
Web: http://www1.maths.leeds.ac.uk/~matak/
Youtube: https://www.youtube.com/channel/UCUZAYHtVoiMqepQflisp66g
********************************************************** Leverhulme Trust Early Career Fellow 2017-2019
**********************************************************
On 13 Sep 2016, at 16:34, Andrew McRae <A.T.T.McRae@bath.ac.uk> wrote:
Probably related to https://github.com/firedrakeproject/firedrake/ issues/862 (third post, not first). You might be able to adapt the fix; I haven't looked in detail.
On 12 September 2016 at 16:34, Anna Kalogirou <A.Kalogirou@leeds.ac.uk
wrote:
Dear all,
Regarding the problem described in my email sent of Friday, I fixed a minor typo and now get the error below. Any ideas what might be causing it? Thanks!
Best, Anna.
Traceback (most recent call last): File "buoy-swe.py", line 89, in <module> F_solver = solver_F(phi0_5, eta1, mu0_5, I, w, phi0, eta0, Z0, W0, step_b, etaR, phi_t, eta_t, mu_t, I_t, v1, v2, v3, v4, dt, Hb, H0, L, dR_dt, g, rho, Mass, solvers_print); File "/Users/matak/Documents/Simulations/Firedrake/Ship/Modules/Mixed system/solvers.py", line 41, in solver_F F_solver = LinearVariationalSolver(F_problem, solver_parameters=solvers_print) File "/Users/matak/firedrake/lib/python2.7/site-packages/ firedrake/variational_solver.py", line 264, in __init__ super(LinearVariationalSolver, self).__init__(*args, **kwargs) File "/Users/matak/firedrake/lib/python2.7/site-packages/ firedrake/variational_solver.py", line 127, in __init__ ctx = solving_utils._SNESContext(problem) File "/Users/matak/firedrake/lib/python2.7/site-packages/firedrake/solving_utils.py", line 107, in __init__ for problem in problems) File "/Users/matak/firedrake/lib/python2.7/site-packages/firedrake/solving_utils.py", line 107, in <genexpr> for problem in problems) File "/Users/matak/firedrake/lib/python2.7/site-packages/firedrake/assemble.py", line 67, in assemble inverse=inverse, nest=nest) File "<decorator-gen-294>", line 2, in _assemble File "/Users/matak/firedrake/lib/python2.7/site-packages/firedrake/utils.py", line 62, in wrapper return f(*args, **kwargs) File "/Users/matak/firedrake/lib/python2.7/site-packages/firedrake/assemble.py", line 163, in _assemble map_pairs.append((op2.DecoratedMap(test.cell_node_map(), cell_domains), File "/Users/matak/firedrake/lib/python2.7/site-packages/ firedrake/functionspaceimpl.py", line 646, in cell_node_map for i, s in enumerate(self._spaces)) File "/Users/matak/firedrake/lib/python2.7/site-packages/pyop2/backends.py", line 118, in __call__ return t(*args, **kwargs) File "/Users/matak/firedrake/lib/python2.7/site-packages/pyop2/caching.py", line 156, in __new__ obj = make_obj() File "/Users/matak/firedrake/lib/python2.7/site-packages/pyop2/caching.py", line 137, in make_obj obj.__init__(*args, **kwargs) File "/Users/matak/firedrake/lib/python2.7/site-packages/pyop2/base.py", line 3342, in __init__ if not all(m is None or m.iterset == self.iterset for m in self._maps): File "/Users/matak/firedrake/lib/python2.7/site-packages/pyop2/base.py", line 3342, in <genexpr> if not all(m is None or m.iterset == self.iterset for m in self._maps): File "/Users/matak/firedrake/lib/python2.7/site-packages/pyop2/utils.py", line 64, in __get__ obj.__dict__[self.__name__] = result = self.fget(obj) File "/Users/matak/firedrake/lib/python2.7/site-packages/pyop2/base.py", line 3366, in iterset return reduce(lambda a, b: a if a is None else a.iterset or b if b is None else b.iterset, self._maps) File "/Users/matak/firedrake/lib/python2.7/site-packages/pyop2/base.py", line 3366, in <lambda> return reduce(lambda a, b: a if a is None else a.iterset or b if b is None else b.iterset, self._maps) File "/Users/matak/firedrake/lib/python2.7/site-packages/pyop2/base.py", line 799, in __getattr__ return getattr(self._parent, name) AttributeError: 'Set' object has no attribute ‘iterset'
**********************************************************
Dr Anna Kalogirou Research Fellow
Department of Applied Mathematics University of Leeds
Web: http://www1.maths.leeds.ac.uk/~matak/
Youtube: https://www.youtube.com/channel/UCUZAYHtVoiMqepQflisp66g
********************************************************** Leverhulme Trust Early Career Fellow 2017-2019
**********************************************************
On 9 Sep 2016, at 17:37, Anna Kalogirou <a.kalogirou@leeds.ac.uk> wrote:
Hi David,
I have tried to implement my problem (attached) using the real function spaces but I get the error below. The code can be found here <https://bitbucket.org/annakalog/buoy2d/src/d6a805b6a8d0/Mixed%20system/?at=master> .
I define the real space by V_R = FunctionSpace(mesh, "R", 0). Is that correct? Also, is it a problem that the linear part of the variational form does not depend on one of the test functions from the mixed function space (only depends on v1, v2, v3 and not v4)?
Best,
Anna.
Traceback (most recent call last): File "buoy-swe.py", line 89, in <module> F_solver = solver_F(phi0_5, eta1, mu0_5, I, w, phi0, eta0, Z0, W0, step_b, etaR, phi_t, eta_t, mu_t, I_t, v1, v2, v3, v4, dt, Hb, H0, L, dR_dt, g, rho, Mass, solvers_print); File "/Users/matak/Documents/Simulations/Firedrake/Ship/Modules/Mixed system/solvers.py", line 41, in solver_F F_solver = LinearVariationalSolver(F_problem, solver_parameters=solvers_print) File "/Users/matak/firedrake/lib/python2.7/site-packages/ firedrake/variational_solver.py", line 264, in __init__ super(LinearVariationalSolver, self).__init__(*args, **kwargs) File "/Users/matak/firedrake/lib/python2.7/site-packages/ firedrake/variational_solver.py", line 127, in __init__ ctx = solving_utils._SNESContext(problem) File "/Users/matak/firedrake/lib/python2.7/site-packages/firedrake/solving_utils.py", line 107, in __init__ for problem in problems) File "/Users/matak/firedrake/lib/python2.7/site-packages/firedrake/solving_utils.py", line 107, in <genexpr> for problem in problems) File "/Users/matak/firedrake/lib/python2.7/site-packages/firedrake/assemble.py", line 67, in assemble inverse=inverse, nest=nest) File "<decorator-gen-294>", line 2, in _assemble File "/Users/matak/firedrake/lib/python2.7/site-packages/firedrake/utils.py", line 62, in wrapper return f(*args, **kwargs) File "/Users/matak/firedrake/lib/python2.7/site-packages/firedrake/assemble.py", line 101, in _assemble inverse=inverse) File "/Users/matak/firedrake/lib/python2.7/site-packages/firedrake/tsfc_interface.py", line 307, in compile_form number_map).kernels File "/Users/matak/firedrake/lib/python2.7/site-packages/pyop2/caching.py", line 198, in __new__ obj = make_obj() File "/Users/matak/firedrake/lib/python2.7/site-packages/pyop2/caching.py", line 188, in make_obj obj.__init__(*args, **kwargs) File "/Users/matak/firedrake/lib/python2.7/site-packages/firedrake/tsfc_interface.py", line 212, in __init__ tree = tsfc_compile_form(form, prefix=name, parameters=parameters) File "/Users/matak/firedrake/lib/python2.7/site-packages/tsfc/driver.py", line 47, in compile_form do_estimate_degrees=True) File "/Users/matak/firedrake/lib/python2.7/site-packages/ufl/ algorithms/compute_form_data.py", line 386, in compute_form_data check_form_arity(preprocessed_form, self.original_form.arguments()) # Currently testing how fast this is File "/Users/matak/firedrake/lib/python2.7/site-packages/ufl/algorithms/check_arities.py", line 152, in check_form_arity check_integrand_arity(itg.integrand(), arguments) File "/Users/matak/firedrake/lib/python2.7/site-packages/ufl/algorithms/check_arities.py", line 145, in check_integrand_arity args = map_expr_dag(rules, expr, compress=False) File "/Users/matak/firedrake/lib/python2.7/site-packages/ufl/corealg/map_dag.py", line 37, in map_expr_dag result, = map_expr_dags(function, [expression], compress=compress) File "/Users/matak/firedrake/lib/python2.7/site-packages/ufl/corealg/map_dag.py", line 86, in map_expr_dags r = handlers[v._ufl_typecode_](v, *[vcache[u] for u in v.ufl_operands]) File "/Users/matak/firedrake/lib/python2.7/site-packages/ufl/algorithms/check_arities.py", line 42, in sum raise ArityMismatch("Adding expressions with non-matching form arguments {0} vs {1}.".format(a, b)) ufl.algorithms.check_arities.ArityMismatch: Adding expressions with non-matching form arguments () vs (Argument(WithGeometry( IndexedProxyFunctionSpace(<firedrake.mesh.ExtrudedMeshTopology object at 0x113e6b890>, TensorProductElement(FiniteElement('Lagrange', interval, 1), FiniteElement('Lagrange', interval, 1), cell=TensorProductCell(interval, interval)), name=None, index=0, component=None), Mesh(VectorElement(TensorProductElement(FiniteElement('Lagrange', interval, 1), FiniteElement('Lagrange', interval, 1), cell=TensorProductCell(interval, interval)), dim=2), 4)), 1, None),). <coupled_system.pdf>
On 18 Aug 2016, at 16:03, David Ham <david.ham@imperial.ac.uk> wrote:
Hi Anna,
Real function spaces now pass some simple tests, at least on my machine. You need the real-function-space branch of firedrake, and the globals_matrices branch of PyOP2. It's not going to get merged before I disappear for a week - not least because I need to update documentation. However you should be able to test it out on the branches.
Regards,
David
On Wed, 10 Aug 2016 at 16:58 David Ham <David.Ham@imperial.ac.uk> wrote:
Unfortunately, It isn't as simple as saying "no mpi". MPI pervades Firedrake. Unless real function spaces play nicely with communicators, Firedrake will simply crash. I'm sorry that this feature is not available right now, but it actually does require significant work.
On Wed, 10 Aug 2016 at 16:51 Anna Kalogirou <A.Kalogirou@leeds.ac.uk> wrote:
Yes indeed, but I believe David would want to make it work for all cases. The merge is necessary in order to proceed with my simulations, so an update as soon as possible would be greatly appreciated.
Best, Anna.
On 10 Aug 2016, at 16:29, Onno Bokhove <O.Bokhove@leeds.ac.uk> wrote:
We would already be helped for the case without MPI, I think that is correct, Anna?
*From:* firedrake-bounces@imperial.ac.uk <firedrake-bounces@imperial. ac.uk> on behalf of David Ham <David.Ham@imperial.ac.uk>
*Sent:* Wednesday, August 10, 2016 10:01 AM
*To:* firedrake *Subject:* Re: [firedrake] Mixed system
The short version is it depends on when I find some time to work on it. It is not simply the case that we have new features lying around waiting to be merged. The current proximate issue is to make the real spaces work properly with MPI communicators. I can't guarantee that will be the last issue.
On Tue, 9 Aug 2016 at 17:37 Onno Bokhove <O.Bokhove@leeds.ac.uk> wrote:
It is possible to merge the branch for the mixed system relatively soon? Anna's results are relevant for the grant application Colin and I aim to submit in Sept. on further wave-ship modelling. Sorry for asking! ------------------------------ *From:* firedrake-bounces@imperial.ac.uk <firedrake- bounces@imperial.ac.uk> on behalf of David Ham < David.Ham@imperial.ac.uk> *Sent:* Friday, August 5, 2016 11:37:23 AM *To:* firedrake *Subject:* Re: [firedrake] Mixed system
On Fri, 5 Aug 2016 at 11:31 Anna Kalogirou <a.kalogirou@leeds.ac.uk> wrote:
Hi Kramer,
Thank you for the reply.
The only problem I see with the solution you suggest, is that both I and lambda will be trial functions, so both terms should be on the bilinear part of the variational form.
Not quite. I and lambda together form a Function in a mixed finite element space. You solve for both of them together, just like pressure and velocity in a fluid simulation.
I could, of course, define a residual and solve everything using nonlinear solvers, but I don't know how efficient that would be.
I think you are confused. None of the relevant issues change in the nonlinear case. After all, the nonlinear solver will differentiate the residual to create a bilinear form anyway.
Incidentally, there is no particular reason to believe that writing the problem in nonlinear form would be inefficient.
Regards,
David
Best,
Anna.
Dr Anna Kalogirou Research Fellow School of Mathematics University of Leeds
http://www1.maths.leeds.ac.uk/~matak/
On 04/08/16 17:29, Stephan Kramer wrote:
On 27/07/16 11:01, Anna Kalogirou wrote: > Dear all, > > I am trying to solve the system found in the attached pdf as a mixed > system. I want to solve the same > problem as the one found here > <https://bitbucket.org/annakalog/buoy2d/src/ 14334d3c20b9f10ed7c1246cde9e3cb60b1c75e4/Inequality% 20constraint/?at=master>, > > but with the use of Schur complements and not linear algebra. > > In a previous discussion about this problem, Lawrence mentioned that the > test function for integral(lambda*Theta(x-Lp)dx) needs to be considered > as coming from the real space of constant functions. Could you please > elaborate on this?
My guess at what Lawrence means is (it's how I would probably approach it), is that you add an additional equation that says
(1) I = integral(lambda*Theta(x-Lp)dx)
and then substitute I in your third equation. Equation (1) can be turned into a finite element weak formulation by considering the function space of functions that are constant over the entire domain This function space is just 1-dimensional and therefore equivalent to the space of reals R and hence we denote this function
-- http://www.imperial.ac.uk/people/colin.cotter www.cambridge.org/9781107663916
On 05/10/16 19:44, Colin Cotter wrote:
We probably need to think about the preconditioner. First please could you confirm that the timing is dominated by the linear solver?
To elaborate on this. Used timed_stages in your code: from pyop2.profiling import timed_stage while t < tend: with timed_stage("solve_A"): do stuff for A. with timed_stage("solve_B"): do stuff for B. ... And then run your models with "-log_view" python my_model.py -log_view And post the outputs in both cases. Cheers, Lawrence
Sorry for the delay, I was away from my desk all day today. I attach the two log files. Best, Anna.
Hi Anna,
On 6 Oct 2016, at 18:42, Anna Kalogirou <A.Kalogirou@leeds.ac.uk> wrote:
Sorry for the delay, I was away from my desk all day today. I attach the two log files.
Best, Anna.
<log_file_mixed.txt><log_file.txt>
OK great. We can see that both options have quite a large room for improvement in the linear solver. The original code does ~6000 KSPSolve calls, but calls PCApply ~395000 times. This tells me you're taking around 65 iterations per solve. Which is probably quite a lot unless your problem is very stiff. The new code (mixed system) is worse, you do only ~2000 KSPSolve calls, but call PCApply almost ~3000000 times, or around ~1500 iterations per solve. This is terrible :). What preconditioning, if any, are you using for these solves (in both cases?). As Colin says, we need to figure out what the linear system looks like to choose an appropriate preconditioner. Cheers, Lawrence
Hi Lawrence, I don’t use any preconditioning in either of the two sets of code. Anna. On 7 Oct 2016, at 14:33, Lawrence Mitchell <lawrence.mitchell@imperial.ac.uk> wrote:
Hi Anna,
On 6 Oct 2016, at 18:42, Anna Kalogirou <A.Kalogirou@leeds.ac.uk> wrote:
Sorry for the delay, I was away from my desk all day today. I attach the two log files.
Best, Anna.
<log_file_mixed.txt><log_file.txt>
OK great. We can see that both options have quite a large room for improvement in the linear solver.
The original code does ~6000 KSPSolve calls, but calls PCApply ~395000 times. This tells me you're taking around 65 iterations per solve. Which is probably quite a lot unless your problem is very stiff.
The new code (mixed system) is worse, you do only ~2000 KSPSolve calls, but call PCApply almost ~3000000 times, or around ~1500 iterations per solve. This is terrible :).
What preconditioning, if any, are you using for these solves (in both cases?). As Colin says, we need to figure out what the linear system looks like to choose an appropriate preconditioner.
Cheers,
Lawrence
_______________________________________________ firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake
participants (3)
-
Anna Kalogirou
-
Colin Cotter
-
Lawrence Mitchell