November 2016 - firedrake - imperial.ac.uk

Getting centre mass of the cell
by George Ovchinnikov 18 Nov '16

18 Nov '16

Hello, There is a nice function, which gives me cell by its coordinates "mesh.locate_cell(x)" I wonder, does it have some kind of inverse analogue, so I could get coordinates for a cell (given by id of cell)? Sincerely, George

1 0

matplotlib query
by William Booker 18 Nov '16

18 Nov '16

Dear Firedrakers, Probably a simple question but how do I get matplotlib to use fixed axes? I've tried plot(all_u, axes= [0.0,1.0,0.0,1.0]), but this doesn't seem to be correct. Thanks, Will

2 2

Re: [firedrake] par_loop error
by Andrew McRae 17 Nov '16

17 Nov '16

Yes, I found it slightly frustrating that I had to enumerate an extra dozen Firedrake/UFL objects that I'd made use of but weren't yet in the namespace. However, I was writing an example file, which is equivalent to "user code". For core Gusto modules, I agree that 'from firedrake import *' is Bad. On 17 November 2016 at 15:49, Miklós Homolya <m.homolya14(a)imperial.ac.uk> wrote: > I don't like that because then flake8 can no longer whether a name has > been defined or not, so it cannot report mistyped names. > > If I don't want to type pkg.foo every time, then I rather do > > > from pkg import (foo, bar, spam, ..., egg, ham) > > > i.e. listing all the stuff I use from the package. > > On 17/11/16 14:53, Shipton, Jemma wrote: > > heh, yes, Andrew and I were talking about this yesterday. I'm not entirely > sure why we don't use from firedrake import * in gusto - would anyone like > to convince me that we should/shouldn't? > ------------------------------ > *From:* firedrake-bounces(a)imperial.ac.uk <firedrake-bounces(a)imperial.ac > .uk> <firedrake-bounces(a)imperial.ac.uk> on behalf of Andrew McRae > <A.T.T.McRae(a)bath.ac.uk> <A.T.T.McRae(a)bath.ac.uk> > *Sent:* 17 November 2016 14:49:09 > *To:* firedrake > *Subject:* Re: [firedrake] par_loop error > > Where? > > https://github.com/firedrakeproject/gusto/blob/master/gusto/advection.py > > On 17 November 2016 at 14:46, T. M. Bendall <t.bendall15(a)imperial.ac.uk> > wrote: > >> It's imported all things from firedrake as * >> >> So I guess so? >> >> Tom >> On 17/11/16 14:40, Lawrence Mitchell wrote: >> >> On 17/11/16 14:37, T. M. Bendall wrote: >> >> ../gusto/advection.py:244: in apply >> self.limit_slope(self.D1) >> >> Has this advection.py file imported the par_loop name? >> >> Lawrence >> >> >> >> >> _______________________________________________ >> firedrake mailing listfiredrake@imperial.ac.ukhttps://mailman.ic.ac.uk/mailman/listinfo/firedrake >> >> _______________________________________________ > firedrake mailing listfiredrake@imperial.ac.ukhttps://mailman.ic.ac.uk/mailman/listinfo/firedrake > >

1 0

Re: [firedrake] par_loop error
by Andrew McRae 17 Nov '16

17 Nov '16

Where? https://github.com/firedrakeproject/gusto/blob/master/gusto/advection.py On 17 November 2016 at 14:46, T. M. Bendall <t.bendall15(a)imperial.ac.uk> wrote: > It's imported all things from firedrake as * > > So I guess so? > > Tom > On 17/11/16 14:40, Lawrence Mitchell wrote: > > > On 17/11/16 14:37, T. M. Bendall wrote: > > ../gusto/advection.py:244: in apply > self.limit_slope(self.D1) > > Has this advection.py file imported the par_loop name? > > Lawrence > > > > > _______________________________________________ > firedrake mailing listfiredrake@imperial.ac.ukhttps://mailman.ic.ac.uk/mailman/listinfo/firedrake > > >

4 4

par_loop error
by T. M. Bendall 17 Nov '16

17 Nov '16

Dear Firedrakers, Just wondering if anyone can help me with an error that I'm coming up again as I can't think of a way to solve it myself! I'm editing the DG_advection class for gusto by trying to implement a limiter but I'm running up against this error (see below). I guess I was expecting to be able to call 'par_loop' without any problems? Cheers, Tom _____________________________________________ test_dgadvection[False-False] _____________________________________________ tmpdir = local('/tmp/pytest-of-tmb15/pytest-16/test_dgadvection_False_False_0'), vector = False, continuity = False @pytest.mark.parametrize("vector", [False, True]) @pytest.mark.parametrize("continuity", [False, True]) def test_dgadvection(tmpdir, vector, continuity): dirname = str(tmpdir) > f_err = run(dirname, vector, continuity) test_DG_advection.py:85: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ test_DG_advection.py:70: in run f_advection.apply(f, fp1) ../gusto/advection.py:244: in apply self.limit_slope(self.D1) _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ self = <gusto.advection.DGAdvection object at 0x7f7c4649c810> D = Coefficient(WithGeometry(FunctionSpace(<firedrake.mesh.MeshTopology object at ...ctorElement(FiniteElement('Lagrange', Cell('triangle', 3), 1), dim=3), 1)), 38) def limit_slope(self, D): self.Dmax.assign(-1.0e10) self.Dmin.assign(1.0e10) self.Dbar.project(D) self.deltaD.project(D-self.Dbar) > par_loop(self.max_kernel, dx, {"maxq":(self.Dmax,RW), "q":(self.Dbar,READ)}) E NameError: global name 'par_loop' is not defined

2 3

Linux integration testing is now Drone only, please merge master
by David Ham 16 Nov '16

16 Nov '16

Dear Firedrakers, We have just transitioned build testing on Linux completely to Drone. Mac testing remains on Travis because Drone doesn't yet do Mac. If you are working on a Firedrake branch, please merge from origin/master (or rebase). This will cause your branch to stop triggering Linux builds on Travis. Many thanks, David

1 0

Firedrake user and developer workshop -- 27 & 28 March 2017
by David Ham 15 Nov '16

15 Nov '16

Dear Firedrakers, The first Firedrake user and developer workshop will be held on 27 & 28 March 2017 at Imperial College London. Please save the dates! The Firedrake workshop will provide the opportunity for Firedrake users and developers to engage with each other to communicate the ways that Firedrake can be used in simulation science, the latest developments in the project, and the future developments anticipated. The workshop will incorporate a half day Firedrake tutorial for interested new users, and will present more experienced users with the opportunity to spend face to face time with developers and other users. A call for contributions and registration information will be published in a few weeks. I hope to see many of you there. Regards, David

2 1

Automatic developer upgrades - risk of failure.
by David Ham 10 Nov '16

10 Nov '16

Dear Firedrakers, Executive summary: I have just merged a change to firedrake-upgrade which might, in a minority of cases, cause an upgrade fail. If this happens to you then first try running the upgrade again. If that fails please report a bug. Full version: Python packages can be installed in editable or non-editable mode. In editable mode, the python files which are run when you import a module are the ones in the source directory. This means that any edits you make apply automatically, without re-installing the package. Firedrake used to have developer mode and non-developer mode installs. These corresponded to editable and non-editable Python packages respectively. However the non-developer mode caused confusion among users and some bugs, it also has no advantages. We therefore some time ago removed the --developer install option so that all new installs are in developer mode. The downside of this is that there are older, non-developer mode installs out there on peoples computers, and this is now an untested and unsupported route. To fix this, latest change to Firedrake will cause firedrake-install to upgrade all installs into developer mode. Once this occurs, it has only advantages. However there are a minority of cases where uninstalling the old packages will fail, so this email is there to provide some explanation about what to do if your update fails. Regards, David

1 0

Re: [firedrake] Mixed system
by David Ham 07 Nov '16

07 Nov '16

A brief status update on this: the real-function-spaces branch of Firedrake taken together with the globals_petsc_changes branch of PyOP2 no longer exhibits the direct crash observed. However I am also yet to actually successfully solve using a block preconditioner. I need to consult with Lawrence to establish whether the problem is another bug or pilot error. Regards, David On Mon, 24 Oct 2016 at 11:37 Mitchell, Lawrence < lawrence.mitchell(a)imperial.ac.uk> wrote: > > > On 24 Oct 2016, at 11:13, Anna Kalogirou <A.Kalogirou(a)leeds.ac.uk> > wrote: > > > > Hi Lawrence, > > > > Any update regarding the bug reported below? > > This appears to be more involved than I had first thought. So no luck > yet, sorry! David and I need to find some time together to work out the > differences/corner cases. > > Cheers, > > Lawrence > _______________________________________________ > firedrake mailing list > firedrake(a)imperial.ac.uk > https://mailman.ic.ac.uk/mailman/listinfo/firedrake >

2 1

Consultation on dates: Firedrake users and developers workshop.
by David Ham 04 Nov '16

04 Nov '16

Dear Firedrakers, We're planning on holding a 2 day meeting in the spring to enable the users and developers in the Firedrake community to exchange ideas about how Firedrake is and can be used, the latest developments and future directions of the project. I need to look into booking rooms and setting this up, so I need to gauge when works. It needs to be outside Imperial term times (so we can get rooms) and it's obviously better if it doesn't clash with too many people's teaching or childcare constraints - so outside school terms is good. With that set of constraints in mind, my first choice is the week beginning 27 March 2017. I believe that outside term for most UK universities, and before school Easter holidays. Is anyone aware of any reasons why that week would be a bad idea? Regards, David

1 0