[petsc-users] stiff support in src/ts/examples/tutorials/ex9.c

[Jed Brown]

Geoffrey Irving <irving at naml.us> writes:

> On Oct 7, 2013, at 12:11 PM, Geoffrey Irving <irving at naml.us> wrote:
>
>> Does src/ts/examples/tutorials/ex9.c support implicit methods for
>> stiff problems?  It doesn't explicitly create any linear solvers, but
>> maybe these are created automatically if you pick the appropriate TS?
>> That would presumably require derivative information though, which I
>> don't see.

The example assembles a block-diagonal matrix based on a cell-centered
characteristic decomposition for preconditioning.  I don't recall if it
was ever tested.  Better to just use coloring (really easy and not very
wasteful 1D problems).  Here is an example of a shallow water dam-break
using an implicit time integrator.

./ex9 -da_grid_x 100 -initial 0 -xmin -2 -xmax 2 -limit mc -physics shallow -bc_type outflow -ts_final_time 1 -ts_monitor_draw_solution -ts_type arkimex -ts_arkimex_fully_implicit -ts_adapt_type none -snes_converged_reason -ts_monitor -snes_fd_color -ts_max_snes_failures -1


Note that the problems in this example are non-stiff so implicit methods
won't be more efficient.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 835 bytes
Desc: not available
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20131007/6b2582e6/attachment.pgp>