Re: [firedrake] Lagrange multipliers
Hi All, Real function spaces have now landed in master. Please attempt your simulations using current master Firedrake and report any bugs you find. Regards, David On Sat, 29 Apr 2017 at 21:19 Anna Kalogirou <A.Kalogirou@leeds.ac.uk> wrote:
I still get the same "Mismatching function spaces” error, unfortunately. The code I am running is basically the same as James’ but with a few more lines added at the end.
On 29 Apr 2017, at 16:50, James Jackaman <wb816921@live.reading.ac.uk> wrote:
Thank you Anna, this was really useful. And thank you David (attached message), the current Firedrake branch and real-function-space are now working for me.
On 25/04/17 21:58, Anna Kalogirou wrote:
Dear James,
I haven’t updated firedrake on my old laptop for a few months (I think since December). The version I have at the moment is:
"fiat": "5bdea9eba05571791b05e763f69e318ccd57006e†, "COFFEE": "0a2a19a377aeba3b9bcdb15952fe6c0da99714a2†, "ufl": "013e6929bc56fc03e019a4f0f3dfd9ceee2eb199†, "firedrake": "e051f502a94edf68ec9e062d94d8a02fb5cc6624†, "tsfc": "e05d750d9daeea53c7a4bc8c71e8d5fcc406b2bb†, "petsc4py": "7770e68fec1953a7315ed187986248af7b4dfb6a†, "petsc": "afae6a5591f0bed801d208622b499bf7ea2fde7e†, "PyOP2": "dd9d40ab7b28c07b7d9547dbb8830f5a87f0d9df†,
Best, Anna.
On 25 Apr 2017, at 16:44, James Jackaman <wb816921@live.reading.ac.uk> wrote:
Dear Anna,
Thank you for looking into my issue further. While the error is being investigated by the Firedrake team would you be able to let me know exactly which version of Firedrake is working on your laptop (via firedrake-zenodo )?
Thanks again,
James On 24/04/17 11:34, Anna Kalogirou wrote:
Dear James (and firedrake team),
I get the same errors as you do when I run your neumann-poisson.py code on my desktop machine, on which I installed firedrake last week. The fact that it runs on my laptop which has a older firedrake version means that something that changed recently is causing the error. I believe it is related to COFFEE, please see .log and .err files attached.
Best, Anna.
On 23 Apr 2017, at 18:25, Anna Kalogirou <A.Kalogirou@leeds.ac.uk> wrote:
Dear James,
Your neumann-poisson.py code also works on my laptop. I’ll leave it to the firedrake developers to help you with your errors.
Best wishes,
Anna.
On 21 Apr 2017, at 16:41, James Jackaman <wb816921@live.reading.ac.uk> wrote:
Hi Anna,
Thanks :), this fixes the real_solver.py test, but unfortunately on this branch of Firedrake the neumann-poisson.py test does not compile (output attached). Is this a bug or am I doing something silly?
All the best,
James On 21/04/17 13:47, Anna Kalogirou wrote:
Hi James,
Your test file for the RealFunctionSpace returns u=1 when I run it. You need to install firedrake and use the correct branch of real-function-space, i.e.
python firedrake-install --package-branch firedrake real-function-space
Regards, Anna.
On 19 Apr 2017, at 15:09, James Jackaman <wb816921@live.reading.ac.uk> wrote:
Dear all,
I'm currently working on an Firedrake implementation which utilises Lagrange multipliers, in the spirit of the Poisson equation with pure Neumann boundary conditions where the average value of the solution is constrained by a Lagrange multiplier.
I have attempted to translate the Fenics demo, see https://fenicsproject.org/olddocs/dolfin/1.6.0/python/demo/documented/neuman..., into Firedrake with no success, see neumann-poisson.py in the attached gist.
I believe there may be a problem with the function space of real constants *FunctionSpace(mesh,"R",0)*, see real_solver.py in the attached gist which returns u=0 instead of u=1.
How can I modify my example codes in the gist such that they behave as expected, or alternatively how can I implement the Poisson equation with pure Neumann boundary conditions using Lagrange multipliers in Firedrake?
Gist: https://gist.github.com/j130792/64d12e84cf5ede2b327d1859a36dc36d
Thanks for your help,
James _______________________________________________ firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake
_______________________________________________ firedrake mailing listfiredrake@imperial.ac.ukhttps://mailman.ic.ac.uk/mailman/listinfo/firedrake
<real-function-space-neumann-output.txt> _______________________________________________ firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake
_______________________________________________ firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake
_______________________________________________ firedrake mailing listfiredrake@imperial.ac.ukhttps://mailman.ic.ac.uk/mailman/listinfo/firedrake
_______________________________________________ firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake
_______________________________________________ firedrake mailing listfiredrake@imperial.ac.ukhttps://mailman.ic.ac.uk/mailman/listinfo/firedrake
<Attached Message.eml>_______________________________________________
firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake
-- Dr David Ham Department of Mathematics Imperial College London
Dear all, I still get the Mismatching Function Spaces error I reported before. After solving a mixed system that involves (at least) one component from the R FunctionSpace, the result cannot be assigned to another function from R even though they both technically belong to the same function space, as seen in the code attached. On 9 May 2017, at 10:47, David Ham <David.Ham@imperial.ac.uk<mailto:David.Ham@imperial.ac.uk>> wrote: Hi All, Real function spaces have now landed in master. Please attempt your simulations using current master Firedrake and report any bugs you find. Regards, David On Sat, 29 Apr 2017 at 21:19 Anna Kalogirou <A.Kalogirou@leeds.ac.uk<mailto:A.Kalogirou@leeds.ac.uk>> wrote: I still get the same "Mismatching function spaces” error, unfortunately. The code I am running is basically the same as James’ but with a few more lines added at the end. On 29 Apr 2017, at 16:50, James Jackaman <wb816921@live.reading.ac.uk<mailto:wb816921@live.reading.ac.uk>> wrote: Thank you Anna, this was really useful. And thank you David (attached message), the current Firedrake branch and real-function-space are now working for me. On 25/04/17 21:58, Anna Kalogirou wrote: Dear James, I haven’t updated firedrake on my old laptop for a few months (I think since December). The version I have at the moment is: "fiat": "5bdea9eba05571791b05e763f69e318ccd57006e†, "COFFEE": "0a2a19a377aeba3b9bcdb15952fe6c0da99714a2†, "ufl": "013e6929bc56fc03e019a4f0f3dfd9ceee2eb199†, "firedrake": "e051f502a94edf68ec9e062d94d8a02fb5cc6624†, "tsfc": "e05d750d9daeea53c7a4bc8c71e8d5fcc406b2bb†, "petsc4py": "7770e68fec1953a7315ed187986248af7b4dfb6a†, "petsc": "afae6a5591f0bed801d208622b499bf7ea2fde7e†, "PyOP2": "dd9d40ab7b28c07b7d9547dbb8830f5a87f0d9df†, Best, Anna. On 25 Apr 2017, at 16:44, James Jackaman <wb816921@live.reading.ac.uk<mailto:wb816921@live.reading.ac.uk>> wrote: Dear Anna, Thank you for looking into my issue further. While the error is being investigated by the Firedrake team would you be able to let me know exactly which version of Firedrake is working on your laptop (via firedrake-zenodo)? Thanks again, James On 24/04/17 11:34, Anna Kalogirou wrote: Dear James (and firedrake team), I get the same errors as you do when I run your neumann-poisson.py code on my desktop machine, on which I installed firedrake last week. The fact that it runs on my laptop which has a older firedrake version means that something that changed recently is causing the error. I believe it is related to COFFEE, please see .log and .err files attached. Best, Anna. On 23 Apr 2017, at 18:25, Anna Kalogirou <A.Kalogirou@leeds.ac.uk<mailto:A.Kalogirou@leeds.ac.uk>> wrote: Dear James, Your neumann-poisson.py code also works on my laptop. I’ll leave it to the firedrake developers to help you with your errors. Best wishes, Anna. On 21 Apr 2017, at 16:41, James Jackaman <wb816921@live.reading.ac.uk<mailto:wb816921@live.reading.ac.uk>> wrote: Hi Anna, Thanks :), this fixes the real_solver.py test, but unfortunately on this branch of Firedrake the neumann-poisson.py test does not compile (output attached). Is this a bug or am I doing something silly? All the best, James On 21/04/17 13:47, Anna Kalogirou wrote: Hi James, Your test file for the RealFunctionSpace returns u=1 when I run it. You need to install firedrake and use the correct branch of real-function-space, i.e. python firedrake-install --package-branch firedrake real-function-space Regards, Anna. On 19 Apr 2017, at 15:09, James Jackaman <wb816921@live.reading.ac.uk<mailto:wb816921@live.reading.ac.uk>> wrote: Dear all, I'm currently working on an Firedrake implementation which utilises Lagrange multipliers, in the spirit of the Poisson equation with pure Neumann boundary conditions where the average value of the solution is constrained by a Lagrange multiplier. I have attempted to translate the Fenics demo, see https://fenicsproject.org/olddocs/dolfin/1.6.0/python/demo/documented/neuman..., into Firedrake with no success, see neumann-poisson.py in the attached gist. I believe there may be a problem with the function space of real constants FunctionSpace(mesh,"R",0), see real_solver.py in the attached gist which returns u=0 instead of u=1. How can I modify my example codes in the gist such that they behave as expected, or alternatively how can I implement the Poisson equation with pure Neumann boundary conditions using Lagrange multipliers in Firedrake? Gist: https://gist.github.com/j130792/64d12e84cf5ede2b327d1859a36dc36d Thanks for your help, James _______________________________________________ firedrake mailing list firedrake@imperial.ac.uk<mailto:firedrake@imperial.ac.uk> https://mailman.ic.ac.uk/mailman/listinfo/firedrake _______________________________________________ firedrake mailing list firedrake@imperial.ac.uk<mailto:firedrake@imperial.ac.uk> https://mailman.ic.ac.uk/mailman/listinfo/firedrake <real-function-space-neumann-output.txt>_______________________________________________ firedrake mailing list firedrake@imperial.ac.uk<mailto:firedrake@imperial.ac.uk> https://mailman.ic.ac.uk/mailman/listinfo/firedrake _______________________________________________ firedrake mailing list firedrake@imperial.ac.uk<mailto:firedrake@imperial.ac.uk> https://mailman.ic.ac.uk/mailman/listinfo/firedrake _______________________________________________ firedrake mailing list firedrake@imperial.ac.uk<mailto:firedrake@imperial.ac.uk> https://mailman.ic.ac.uk/mailman/listinfo/firedrake _______________________________________________ firedrake mailing list firedrake@imperial.ac.uk<mailto:firedrake@imperial.ac.uk> https://mailman.ic.ac.uk/mailman/listinfo/firedrake _______________________________________________ firedrake mailing list firedrake@imperial.ac.uk<mailto:firedrake@imperial.ac.uk> https://mailman.ic.ac.uk/mailman/listinfo/firedrake <Attached Message.eml>_______________________________________________ firedrake mailing list firedrake@imperial.ac.uk<mailto:firedrake@imperial.ac.uk> https://mailman.ic.ac.uk/mailman/listinfo/firedrake -- Dr David Ham Department of Mathematics Imperial College London _______________________________________________ firedrake mailing list firedrake@imperial.ac.uk<mailto:firedrake@imperial.ac.uk> https://mailman.ic.ac.uk/mailman/listinfo/firedrake
Dear Anna, On 09/05/17 15:47, Anna Kalogirou wrote:
Dear all,
I still get the Mismatching Function Spaces error I reported before. After solving a mixed system that involves (at least) one component from the R FunctionSpace, the result cannot be assigned to another function from R even though they both technically belong to the same function space, as seen in the code attached.
Thanks for your patience. This bug is now fixed in latest master (commit 26d813ea57). Thanks, Lawrence
Thank you so much Lawrence (and David), I have also fixed the convergence issue from yesterday — it was a typo (of a wrong sign) in the equations after all! Best, Anna.
On 10 May 2017, at 14:05, Lawrence Mitchell <lawrence.mitchell@imperial.ac.uk> wrote:
Dear Anna,
On 09/05/17 15:47, Anna Kalogirou wrote:
Dear all,
I still get the Mismatching Function Spaces error I reported before. After solving a mixed system that involves (at least) one component from the R FunctionSpace, the result cannot be assigned to another function from R even though they both technically belong to the same function space, as seen in the code attached.
Thanks for your patience. This bug is now fixed in latest master (commit 26d813ea57).
Thanks,
Lawrence
_______________________________________________ firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake
participants (3)
-
Anna Kalogirou
-
David Ham
-
Lawrence Mitchell