November 2013 - firedrake - imperial.ac.uk

Re: [firedrake] Informal finite element course.
by David Ham 06 Dec '13

06 Dec '13

Hi Rod, That might be tricky. I had anticipated standing at a whiteboard and winging it, and I am not aware of us having a room with the right gear to webcast that, although if someone wants to prove me wrong I'm all ears. David On 29 October 2013 18:36, M R Tohid <mraste2(a)tigers.lsu.edu> wrote: > Hi, > I am also interested and would like to participate if it is not too much > of a hassle to webcast the lectures. > > Cheers, > Rod > > > On Tue, Oct 29, 2013 at 10:52 AM, David Ham <David.Ham(a)imperial.ac.uk>wrote: > >> Hi All, >> >> Having been extensively lobbied over many months, and also being aware >> that I am explaining basic finite element repeatedly to different people, I >> have decided to cave in to the pressure and actually do something a bit >> more organised. >> >> Following very loosely the model of Colin's FEEC course, I think the >> plan would be to do a session of 90 minutes or so perhaps once a week for a >> few weeks until we have got far enough. The target market is anyone who >> doesn't yet know how to derive the finite element method for systems like >> Poisson, the wave equation, advection diffusion etc. We would cover the >> basic theory of finite element, and matters such as quadrature, boundary >> conditions, Galerkin orthogonality and so on. >> >> At this stage I am just gauging interest, so please reply (to list) >> indicating if you would be interested. This will enable me to work out >> whether its worth doing and where I should do it. >> >> Regards, >> >> David >> >> >> >> -- >> Dr David Ham >> Departments of Mathematics and Computing >> Imperial College London >> >> http://www.imperial.ac.uk/people/david.ham >> >> _______________________________________________ >> 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

3 3

PDESoft 2014
by David Ham 28 Nov '13

28 Nov '13

Hi all, I have just accepted an invitation to give a keynote at PDESoft 2014 in Heidelberg. http://pdesoft.uni-hd.de/index.html This looks like a really interesting conference for us: the PC includes people from FEniCS, Dune and ClawPack and previous keynotes include people like Tim Warburton and Anders Logg. I think we should seriously consider submitting more abstracts to this once the call for papers comes out, as it looks like many of the people we would like to see as our peers will be involved. Regards, David -- Dr David Ham Departments of Mathematics and Computing Imperial College London http://www.imperial.ac.uk/people/david.ham

1 0

UK Manycore workshop Dec 16
by Kelly, Paul H J 28 Nov '13

28 Nov '13

If you wish to attend the UK Manycore Workshop please just register: http://www.oerc.ox.ac.uk/research/ccoe/ukmac2013 Attendance for the dinner is compulsory. Paul

1 0

Finite element course
by David Ham 26 Nov '13

26 Nov '13

First, thanks to all who turned up for the first lecture yesterday. The recording of the first lecture is now up. The Silverlight Panopto version is at: http://bit.ly/1ifxZxD while the mp4 podcast version is at: http://bit.ly/1j955g5 The podcast version has higher video quality and works on Linux, but has some audio sync issues about which I have filed a bug. The textbook I plan to use in future weeks is: The Mathematical Theory of Finite Element Methods by Brenner and Scott. Imperial has bought the PDF rights to this book so you can access it at: http://link.springer.com/book/10.1007%2F978-0-387-75934-0 I recommend you print chapter 0 and bring it to the next lecture. The notes I used yesterday are attached, for the benefit of anyone who didn't get a copy. The version below has a couple of very minor errors fixed, but not enough to really concern anyone. Finally, WPL is almost solidly booked next week, however it is free on Friday morning. PLEASE TELL ME NOW IF FRIDAY MORNING IS NOT A GOOD TIME. Regards, David -- Dr David Ham Departments of Mathematics and Computing Imperial College London http://www.imperial.ac.uk/people/david.ham

1 2

Re: [firedrake] Finite element course
by David Ham 18 Nov '13

18 Nov '13

Dear All, This week's finite element lecture is now scheduled for 1030-1200 on Friday in WPL 3.01. Regards, David On 15 November 2013 12:12, Ham, David A <david.ham(a)imperial.ac.uk> wrote: > First, thanks to all who turned up for the first lecture yesterday. > > The recording of the first lecture is now up. The Silverlight Panopto > version is at: > http://bit.ly/1ifxZxD while the mp4 podcast version is at: > http://bit.ly/1j955g5 > > The podcast version has higher video quality and works on Linux, but > has some audio sync issues about which I have filed a bug. > > The textbook I plan to use in future weeks is: The Mathematical Theory > of Finite Element Methods by Brenner and Scott. Imperial has bought the PDF > rights to this book so you can access it at: > http://link.springer.com/book/10.1007%2F978-0-387-75934-0 > > I recommend you print chapter 0 and bring it to the next lecture. > > The notes I used yesterday are attached, for the benefit of anyone who > didn't get a copy. The version below has a couple of very minor errors > fixed, but not enough to really concern anyone. > > Finally, WPL is almost solidly booked next week, however it is free on > Friday morning. PLEASE TELL ME NOW IF FRIDAY MORNING IS NOT A GOOD TIME. > > Regards, > > David > > -- > Dr David Ham > Departments of Mathematics and Computing > Imperial College London > > http://www.imperial.ac.uk/people/david.ham > -- Dr David Ham Departments of Mathematics and Computing Imperial College London http://www.imperial.ac.uk/people/david.ham

1 0

Re: [firedrake] Finite element course
by Fabio Luporini 15 Nov '13

15 Nov '13

Thanks for all this, David. It's OK for me next Friday morning. -- Fabio 2013/11/15 Ham, David A <david.ham(a)imperial.ac.uk> > First, thanks to all who turned up for the first lecture yesterday. > > The recording of the first lecture is now up. The Silverlight Panopto > version is at: > http://bit.ly/1ifxZxD while the mp4 podcast version is at: > http://bit.ly/1j955g5 > > The podcast version has higher video quality and works on Linux, but > has some audio sync issues about which I have filed a bug. > > The textbook I plan to use in future weeks is: The Mathematical Theory > of Finite Element Methods by Brenner and Scott. Imperial has bought the PDF > rights to this book so you can access it at: > http://link.springer.com/book/10.1007%2F978-0-387-75934-0 > > I recommend you print chapter 0 and bring it to the next lecture. > > The notes I used yesterday are attached, for the benefit of anyone who > didn't get a copy. The version below has a couple of very minor errors > fixed, but not enough to really concern anyone. > > Finally, WPL is almost solidly booked next week, however it is free on > Friday morning. PLEASE TELL ME NOW IF FRIDAY MORNING IS NOT A GOOD TIME. > > Regards, > > David > > -- > Dr David Ham > Departments of Mathematics and Computing > Imperial College London > > http://www.imperial.ac.uk/people/david.ham >

1 0

Supporting flattened Args
by Lawrence Mitchell 13 Nov '13

13 Nov '13

Dear all, we currently have two supported ways of passing things into kernels, depending on whether or not the kernel expects "flattened" arguments or not. This is best explained with an example: In the flattened case our kernel computes a scalar value, in the non-flattened case it does not. The code generation takes care of looping sufficiently over all the pieces. The flattened case is only applicable to indirect dats accessed with an iteration-space index (or mats, which I'll sweep under the carpet for now), so imagine we write a par_loop accessing a dat through a map with arity 3, but the dat is vector-valued. void flat_kernel(double A[1], ...) { A[0] += ...; } void kernel(double A[2], ...) { A[0] += ...; A[1] += ...; } In the flattened case, the inner loop does: // loop over elements (indexed by ele) for ( i = 0; i < 6; i++ ) { flat_kernel(A + A_map[ele * 3 + i % 3] * 2 + i / 3, ...); } In the non-flattened case: for ( i = 0; i < 3; i++ ) { kernel(A + A_map[ele * 3 + i] * 2, ...); } that is in the flat case we explicitly point to every data element, in the non-flat case we just point to the start of the block. FFC expects the former. Maintaining both code generation paths is becoming somewhat involved, particularly in the extruded case (note the current issues with vector function spaces on extruded meshes are an example of this). Do we want to continue supporting both cases? In firedrake we currently do rely on both. Flattened for all ffc-generated kernels, non-flattened in, at least, interpolation. Lawrence

1 0

James Maddison's seminar tomorrow 16:10
by David Ham 11 Nov '13

11 Nov '13

Hi All, Our friend James Maddison, formerly of Oxford, now a lecturer at Edinburgh, is giving the applied maths seminar tomorrow afternoon at 16:10. This is work he has been doing with FEniCS and might possibly do with Firedrake in the future, so I suggest we are very interested indeed. http://www3.imperial.ac.uk/newsandeventspggrp/imperialcollege/naturalscienc… Regards, David -- Dr David Ham Departments of Mathematics and Computing Imperial College London http://www.imperial.ac.uk/people/david.ham

2 1

Informal finite element course.
by David Ham 11 Nov '13

11 Nov '13

Hi All, Having been extensively lobbied over many months, and also being aware that I am explaining basic finite element repeatedly to different people, I have decided to cave in to the pressure and actually do something a bit more organised. Following very loosely the model of Colin's FEEC course, I think the plan would be to do a session of 90 minutes or so perhaps once a week for a few weeks until we have got far enough. The target market is anyone who doesn't yet know how to derive the finite element method for systems like Poisson, the wave equation, advection diffusion etc. We would cover the basic theory of finite element, and matters such as quadrature, boundary conditions, Galerkin orthogonality and so on. At this stage I am just gauging interest, so please reply (to list) indicating if you would be interested. This will enable me to work out whether its worth doing and where I should do it. Regards, David -- Dr David Ham Departments of Mathematics and Computing Imperial College London http://www.imperial.ac.uk/people/david.ham

6 10

Re: [firedrake] Informal finite element course.
by Fabio Luporini 11 Nov '13

11 Nov '13

That's perfect for me. Thanks -- Fabio 2013/11/11 Ham, David A <david.ham(a)imperial.ac.uk> > On 11 November 2013 14:08, Rathgeber, Florian < > f.rathgeber10(a)imperial.ac.uk> wrote: > >> On 11/11/13 13:41, Lawrence Mitchell wrote: >> > >> > On 11 Nov 2013, at 13:38, David Ham <David.Ham(a)imperial.ac.uk> wrote: >> > >> >> Dear all, >> >> >> >> This morning we managed to record a short section of audio and video >> from the access grid room onto Panoptico, which is college's lecture >> recording system. I think we're therefore ready to go with the informal >> finite element course. This means we need to do it in WPL 3.01, which has >> the disadvantage that the whiteboards are less than amazing, but it will do. >> >> >> >> The second disadvantage of WPL 3.01 is that it's booked a hell of a >> lot of the time. At this stage I could book the room from 1030 until 1200 >> this Thursday. Please ANSWER NOW whether this is a good time for the first >> lecture. >> > >> > Those timings WFM. >> >> WFM too. Have we concluded that the video is good enough and we don't >> need to faff with tethered recording on my DSLR? >> >> Florian >> >> > I think so. The resolution was OK-ish and I plan to use extra fat > whiteboard markers (of which I have a set) to help things along. > > >> > Lawrence >> >> > > > -- > Dr David Ham > Departments of Mathematics and Computing > Imperial College London > > http://www.imperial.ac.uk/people/david.ham >

1 0