April 2014 - firedrake - imperial.ac.uk

PETSc versions
by Lawrence Mitchell 16 Apr '14

16 Apr '14

Dear firedrakers, you will (with current firedrake master) need the PETSc "next" branch installed. This is due to us requiring some functionality that has not yet hit master at their end. I have updated the installation instructions to reflect this. When we drop back down to just requiring master, I'll update again. Cheers, Lawrence

1 0

Jobs in ECMWF scalability programme
by Cotter, Colin J 15 Apr '14

15 Apr '14

These jobs may be of interest: http://www.ecmwf.int/newsevents/employment/en/vn14-05-en.html Particularly attractive is that you wouldn't pay any income tax! --cjc

1 0

FEniCS/PDESoft Abstracts
by Florian Rathgeber 14 Apr '14

14 Apr '14

Hi all, Here are my drafts of potential talks to submit: FEniCS'14: https://gist.github.com/kynan/9c9c048b0c05fa8d0188 "Firedrake: Composing Domain-specific Abstractions for Productive and Efficient Computational Science" PDESoft 14: https://gist.github.com/kynan/025857adaa21bfabdd59 "PyOP2: a performance-portable framework for parallel computations on unstructured meshes" Cheers, Florian

3 4

Re: [firedrake] ObjectCached
by Andrew McRae 11 Apr '14

11 Apr '14

The former (at least that's what I do), with -U --no-deps, I think (check the website instructions) You might get away with just updating petsc4py (obtaining the proper petsc repository takes unpleasantly long) On 11 April 2014 23:28, Cotter, Colin J <colin.cotter(a)imperial.ac.uk> wrote: > Do I just do that via "pip install" again or is there an "update"? > > --cjc > ------------------------------ > *From:* firedrake-bounces(a)imperial.ac.uk [firedrake-bounces(a)imperial.ac.uk] > on behalf of Andrew McRae [a.mcrae12(a)imperial.ac.uk] > *Sent:* 11 April 2014 23:27 > > *To:* firedrake > *Subject:* Re: [firedrake] ObjectCached > > If your PETSc is more than 3 hours old, you might need to update it. > > ^^^ the rule I've been following recently > > > On 11 April 2014 23:25, Cotter, Colin J <colin.cotter(a)imperial.ac.uk>wrote: > >> ...ah, but now a different problem: >> >> >> --------------------------------------------------------------------------- >> TypeError Traceback (most recent call >> last) >> /home/cjc1/firedrake/firedrake/<ipython-input-2-9db8ddc4ad8d> in >> <module>() >> ----> 1 mesh = UnitSquareMesh(5,5) >> >> /home/cjc1/firedrake/firedrake/firedrake/mesh.pyc in __init__(self, nx, >> ny) >> 134 dmplex.setLabelValue("boundary_ids", face, 4) >> 135 >> --> 136 super(UnitSquareMesh, self).__init__(self.name, >> plex=dmplex) >> 137 >> 138 >> >> /home/cjc1/firedrake/firedrake/firedrake/core_types.so in >> firedrake.core_types.Mesh.__init__ (firedrake/core_types.c:8571)() >> >> /home/cjc1/firedrake/firedrake/firedrake/core_types.so in >> firedrake.core_types.Mesh._from_dmplex (firedrake/core_types.c:9978)() >> >> /usr/local/lib/python2.7/dist-packages/petsc4py/lib/PETSc.so in >> petsc4py.PETSc.DMPlex.createSection (src/petsc4py.PETSc.c:196190)() >> >> TypeError: createSection() got an unexpected keyword argument 'perm' >> >> ------------------------------ >> *From:* firedrake-bounces(a)imperial.ac.uk [ >> firedrake-bounces(a)imperial.ac.uk] on behalf of Cotter, Colin J [ >> colin.cotter(a)imperial.ac.uk] >> *Sent:* 11 April 2014 23:22 >> *To:* firedrake >> *Subject:* Re: [firedrake] ObjectCached >> >> Oh, as usual, I had forgotten to also pull from pyOP2. >> >> --cjc >> ------------------------------ >> *From:* firedrake-bounces(a)imperial.ac.uk [ >> firedrake-bounces(a)imperial.ac.uk] on behalf of Cotter, Colin J [ >> colin.cotter(a)imperial.ac.uk] >> *Sent:* 11 April 2014 23:18 >> *To:* firedrake >> *Subject:* [firedrake] ObjectCached >> >> Dear firedrakers, >> I just pulled from master, and now get the following. What do I have to >> do? >> >> --cjc >> >> File "/home/cjc1/firedrake/firedrake/firedrake/__init__.py", line 2, in >> <module> >> from core_types import * >> File "core_types.pyx", line 17, in init firedrake.core_types >> (firedrake/core_types.c:36280) >> ImportError: cannot import name ObjectCached >> cjc1@richardson:~/mercurial/vertical-tests$ firedrake >> cjc1@richardson:~/mercurial/vertical-tests$ python hydrostatic_bendy.py >> Traceback (most recent call last): >> File "hydrostatic_bendy.py", line 1, in <module> >> from firedrake import * >> File "/home/cjc1/firedrake/firedrake/firedrake/__init__.py", line 2, in >> <module> >> from core_types import * >> File "core_types.pyx", line 17, in init firedrake.core_types >> (firedrake/core_types.c:36280) >> ImportError: cannot import name ObjectCached >> >> >

1 0

Re: [firedrake] ObjectCached
by Andrew McRae 11 Apr '14

11 Apr '14

If your PETSc is more than 3 hours old, you might need to update it. ^^^ the rule I've been following recently On 11 April 2014 23:25, Cotter, Colin J <colin.cotter(a)imperial.ac.uk> wrote: > ...ah, but now a different problem: > > --------------------------------------------------------------------------- > TypeError Traceback (most recent call last) > /home/cjc1/firedrake/firedrake/<ipython-input-2-9db8ddc4ad8d> in <module>() > ----> 1 mesh = UnitSquareMesh(5,5) > > /home/cjc1/firedrake/firedrake/firedrake/mesh.pyc in __init__(self, nx, ny) > 134 dmplex.setLabelValue("boundary_ids", face, 4) > 135 > --> 136 super(UnitSquareMesh, self).__init__(self.name, > plex=dmplex) > 137 > 138 > > /home/cjc1/firedrake/firedrake/firedrake/core_types.so in > firedrake.core_types.Mesh.__init__ (firedrake/core_types.c:8571)() > > /home/cjc1/firedrake/firedrake/firedrake/core_types.so in > firedrake.core_types.Mesh._from_dmplex (firedrake/core_types.c:9978)() > > /usr/local/lib/python2.7/dist-packages/petsc4py/lib/PETSc.so in > petsc4py.PETSc.DMPlex.createSection (src/petsc4py.PETSc.c:196190)() > > TypeError: createSection() got an unexpected keyword argument 'perm' > > ------------------------------ > *From:* firedrake-bounces(a)imperial.ac.uk [firedrake-bounces(a)imperial.ac.uk] > on behalf of Cotter, Colin J [colin.cotter(a)imperial.ac.uk] > *Sent:* 11 April 2014 23:22 > *To:* firedrake > *Subject:* Re: [firedrake] ObjectCached > > Oh, as usual, I had forgotten to also pull from pyOP2. > > --cjc > ------------------------------ > *From:* firedrake-bounces(a)imperial.ac.uk [firedrake-bounces(a)imperial.ac.uk] > on behalf of Cotter, Colin J [colin.cotter(a)imperial.ac.uk] > *Sent:* 11 April 2014 23:18 > *To:* firedrake > *Subject:* [firedrake] ObjectCached > > Dear firedrakers, > I just pulled from master, and now get the following. What do I have to > do? > > --cjc > > File "/home/cjc1/firedrake/firedrake/firedrake/__init__.py", line 2, in > <module> > from core_types import * > File "core_types.pyx", line 17, in init firedrake.core_types > (firedrake/core_types.c:36280) > ImportError: cannot import name ObjectCached > cjc1@richardson:~/mercurial/vertical-tests$ firedrake > cjc1@richardson:~/mercurial/vertical-tests$ python hydrostatic_bendy.py > Traceback (most recent call last): > File "hydrostatic_bendy.py", line 1, in <module> > from firedrake import * > File "/home/cjc1/firedrake/firedrake/firedrake/__init__.py", line 2, in > <module> > from core_types import * > File "core_types.pyx", line 17, in init firedrake.core_types > (firedrake/core_types.c:36280) > ImportError: cannot import name ObjectCached > >

2 1

ObjectCached
by Cotter, Colin J 11 Apr '14

11 Apr '14

Dear firedrakers, I just pulled from master, and now get the following. What do I have to do? --cjc File "/home/cjc1/firedrake/firedrake/firedrake/__init__.py", line 2, in <module> from core_types import * File "core_types.pyx", line 17, in init firedrake.core_types (firedrake/core_types.c:36280) ImportError: cannot import name ObjectCached cjc1@richardson:~/mercurial/vertical-tests$ firedrake cjc1@richardson:~/mercurial/vertical-tests$ python hydrostatic_bendy.py Traceback (most recent call last): File "hydrostatic_bendy.py", line 1, in <module> from firedrake import * File "/home/cjc1/firedrake/firedrake/firedrake/__init__.py", line 2, in <module> from core_types import * File "core_types.pyx", line 17, in init firedrake.core_types (firedrake/core_types.c:36280) ImportError: cannot import name ObjectCached

1 2

vertical slice + topography
by Cotter, Colin J 11 Apr '14

11 Apr '14

Dear Firedrakers, The attached image doesn't look very exciting but it shows that we managed to get a hydrostatic pressure calculation in the presence of topography (a hill!) working in the extruded FEEC framework. Things are starting to work! We took a temperature field that depends only on z, and computed the pressure on a terrain following mesh, the exact solution depends only on z too. This is with the lowest order spaces i.e. DG1 for pressure and P0xP1 for temperature. cheers --cjc

2 1

Fwd: Tim Palmer, Grantham Seminar Speaker - 9 April
by David Ham 08 Apr '14

08 Apr '14

Hi All, An interesting talk about a highly unconventional approach to precision in algorithms tomorrow... David Greater accuracy with less precision: a new paradigm for weather and climate prediction. A link to the abstract and the bio is below: http://www3.imperial.ac.uk/newsandeventspggrp/imperialcollege/naturalscienc… We hope to see many of you here. Best wishes Gosia Gosia Gayer Grantham Institute for Climate Change Imperial College London South Kensington Campus London SW7 2AZ Tel: +44 (0)20 7594 9666 Fax: +44 (0)20 7594 9668 Email: g.gayer(a)imperial.ac.uk www.imperial.ac.uk/climatechange *From:* Gayer, Gosia *Sent:* 08 April 2014 14:13 *To:* Toumi, Ralf; Czaja, Arnaud (a.czaja(a)imperial.ac.uk) Voulgarakis, Apostolos (a.voulgarakis(a)imperial.ac.uk) Ham, David A ( david.ham(a)imperial.ac.uk) *Cc:* Nissan, Hannah; Hoskins, Brian J; Buckle, Simon J *Subject:* RE: Tim Palmer, Grantham Seminar Speaker - 9 April Apologies, the programme is now attached. Gosia *From:* Gayer, Gosia *Sent:* 08 April 2014 14:11 *To:* Toumi, Ralf; Czaja, Arnaud (a.czaja(a)imperial.ac.uk) Voulgarakis, Apostolos (a.voulgarakis(a)imperial.ac.uk) Ham, David A ( david.ham(a)imperial.ac.uk) *Cc:* Nissan, Hannah; Hoskins, Brian J; Buckle, Simon J *Subject:* RE: Tim Palmer, Grantham Seminar Speaker - 9 April Dear All Thank you for making yourselves available to meet Professor Tim Palmer. Please find attached Tim’s schedule for tomorrow. The meetings will be in the small office at the Grantham Institute. Best wishes Gosia ------------------------------ *From:* Gayer, Gosia *Sent:* 04 April 2014 11:19 *To:* Czaja, Arnaud; Brindley, Helen E; Graven, Heather; Toumi, Ralf; Piggott, Matthew D; Cotter, Colin J; Ham, David A *Subject:* Tim Palmer, Grantham Seminar Speaker - 9 April Dear All I am in the process of organising a programme for Professor Tim Palmer, who is our Grantham Seminar speaker on 9 April 2014. Title – “Greater accuracy with less precision: a new paradigm for weather and climate prediction”. *Speaker – **Professor Tim Palmer, Royal Society Research Professor in Climate Physics, University of Oxford* http://www3.imperial.ac.uk/newsandeventspggrp/imperialcollege/naturalscienc… We know that our speakers value the opportunity to talk to Imperial’s academics and I normally organise 3 x 30 minute sessions starting at 14.15 ie: 14.15-14.45 14.45-15.15 15.15-15.45 Your names were suggested by Professor Brian Hoskins or you may have expressed interest to meet Professor Palmer yourself. Pease let me know if you are willing and available to meet Professor Palmer. It would be very helpful to have your answer asap. Many thanks Gosia Gosia Gayer Grantham Institute for Climate Change Imperial College London South Kensington Campus London SW7 2AZ Tel: +44 (0)20 7594 9666 Fax: +44 (0)20 7594 9668 Email: g.gayer(a)imperial.ac.uk www.imperial.ac.uk/climatechange -- Dr David Ham Departments of Mathematics and Computing Imperial College London http://www.imperial.ac.uk/people/david.ham

1 0

New meeting time
by Lawrence Mitchell 03 Apr '14

03 Apr '14

Dear all, the firedrake status meeting has moved from Thursday at 3pm to Friday at 1pm in WPL 301. Hopefully this is still convenient for everyone. Lawrence

2 1

Re: [firedrake] R space
by David Ham 02 Apr '14

02 Apr '14

Hi All, Lawrence and I had a productive lunch in which we think we worked out how to do R efficiently. Conversation on this subject will now move to the issue (https://github.com/firedrakeproject/firedrake/issues/226) and all of your input is actively sought. David On 12 March 2014 09:21, Farrell, Patrick <patrick.farrell06(a)imperial.ac.uk>wrote: > On 12/03/14 09:17, Patrick Farrell wrote: > > R is useful for things other than boundary conditions, too. If you have > a constrained > > optimisation problem with a single constraint (an integral constraint, > say, like [0]), > > P.S. For working code where you need this, see > > > https://bitbucket.org/dolfin-adjoint/da-applications/src/3907be10e58ab0f61e… > > which solves the KKT equations associated with example 4 of [0] by means > of a combined continuation on log barrier terms and casting the problem > as a mixed complementarity problem to the PETSc VI solver. > > I'd be dead impressed if firedrake could do this (even more so if it could > do this on GPUs). > > Patrick > > _______________________________________________ > 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

1 0