I know this is probably little used these days, but I am in the process of writing a small modification to create another test and wanted to make sure that the old brill initial data test will run. So I tried:
./simfactory/bin/sim submit brill_initialdata --parfile=par/test_brilldata_1.par --procs=1 --walltime=1:0:0
It did not get started really. Here is the output which indicates where it had trouble:
Activating thorn Cactus...Success -> active implementation Cactus Activation requested for --->bam_elliptic LocalReduce admbase staticconformal ellbase boundary pugh pughslab pughreduce CartGrid3D CoordBase SymBase idbrilldata ioascii iobasic ioutil time <--- Thorn pugh requests automatic activation of MPI Thorn MPI requests automatic activation of hwloc Thorn hwloc requests automatic activation of zlib Error: Thorn bam_elliptic not found
Is bam_elliptic existent and if so does it work with the current flavors of ET?
Comer
On Wed, May 27, 2015 at 11:30:46AM -0400, Comer Duncan wrote:
Error: Thorn bam_elliptic not found
Is bam_elliptic existent and if so does it work with the current flavors of ET?
BAM_Elliptic probably still exists somewhere, but I have not heard of someone using it in quite a while and would not count on it still working. BAM is used as an elliptic solver here, but there are alternatives (sor or petsc). There doesn't seem to be a test case for either (there should at least be one for sor), but there is a par file in the par/ directory of IDBrillData. If what you are interested in is a test, and you prepare one using sor (it might not take a lot given how fast that one runs), we would be interested in it to add to the toolkit.
Frank
Hi Frank,
While I could use sor, it seems quite a dip into the past as it was used in the really old days of NR. Now one would hope that more modern and efficient and faster tools are available.
Which brings me to ask: what elliptic solvers are part and parcel of ET nowadays? I seem to remember but can not confirm that Erik had created some meta thing for solving elliptics of the for m delsqr(phi) + M*phi + N = 0. What can you tell me about this topic?
Thanks for your help.
Comer
On Wed, May 27, 2015 at 12:10 PM, Frank Loeffler knarf@cct.lsu.edu wrote:
On Wed, May 27, 2015 at 11:30:46AM -0400, Comer Duncan wrote:
Error: Thorn bam_elliptic not found
Is bam_elliptic existent and if so does it work with the current flavors
of
ET?
BAM_Elliptic probably still exists somewhere, but I have not heard of someone using it in quite a while and would not count on it still working. BAM is used as an elliptic solver here, but there are alternatives (sor or petsc). There doesn't seem to be a test case for either (there should at least be one for sor), but there is a par file in the par/ directory of IDBrillData. If what you are interested in is a test, and you prepare one using sor (it might not take a lot given how fast that one runs), we would be interested in it to add to the toolkit.
Frank
On Wed, May 27, 2015 at 12:50:20PM -0400, Comer Duncan wrote:
Which brings me to ask: what elliptic solvers are part and parcel of ET nowadays? I seem to remember but can not confirm that Erik had created some meta thing for solving elliptics of the for m delsqr(phi) + M*phi + N = 0. What can you tell me about this topic?
I am not the best person to talk about using an elliptic solver - Erik might be. The ET does have an interface to Petsc, and this thorn at least seems to have the option to use that. I personally have never used that though.
Frank
users@lists.einsteintoolkit.org
-
Comer Duncan -
Frank Loeffler