On Mon, Apr 12, 2010 at 07:59:37PM -0400, Roland Haas wrote:
Test suites which I can make run:
Great! Thanks!
Test suites which I could not make run (successfully):
- ADMConstraints_Mask in ADMConstraints
- ADMConstraints_Physical in ADMConstraints
- ADMConstraints_Shift in ADMConstraints
- test_ADM in ADMConstraints
They are all in ADMConstraints which IMHO has a very limited lifetime itself anyway, for production at least. We could probably run them including ADM, make sure they pass and include another testsuite using McLachlan. I would use this procedure for all teststuites which only fail because of ADM missing.
- checkpoint2 in AHFinderDirect
- needs LegoExcision
LegoExcision will be part of the ET. I am not sure if it is generally available from the AEI, the relevant URL is
https://svn.aei.mpg.de/numrel/Excision/LegoExcision/trunk/
- Kerr-Cartoon in AHFinderDirect
- needs Cartoon2D
Cartoon will also be part of the ET. It currently still in the Cactus BetaThorns.
- de_Sitter in Exact
- runs with ML+MoL but produces completely different results in
- eg. the lapse (which is set by Exact) (AEIThorns/Exact/test/de_Sitter.par) Missing: ADM_BSSN
As far as I know ADM_BSSN is in AEI_Archive. It is probably time to archive this testsuite as well and replace it with the corresponding McLachlan version. It would be interesting to know if the testsuites which (only) need BSSN_MoL pass if that thorn is included. Can you please test that as well?
- bam in TwoPunctures
- needs bam_elliptic (AEIThorns/TwoPunctures/test/bam.par) Missing: BAM_ELLIPTIC
I wouldn't just remove this and there is no real way to make it work publicly, so we may just have to accept that this doesn't run within the ET.
Missing: CARPETMASK
This thorn is included in the ET thornlist. Maybe it got added after you build your configuration?
Frank
Hello all,
It would be interesting to know if the testsuites which (only) need BSSN_MoL pass if that thorn is included. Can you please test that as well?
Well I have gone ahead and added all the thorns that I could lay my hands on to the executable (still missing are a *thorn* ADM_BSSN, InterpToArray [in AEIDevelopment],BAM_ELLIPTIC, IDLINEARWAVES, NOISE, NORM, TATELLIPTIC, TAT*). These prevent
GaugeWave-type0.01-rho1 in BSSN_MoL (GaugeWave, variant 0.01, with 50 points) Missing: IDLINEARWAVES NOISE NORMS
de_Sitter in Exact (AEIThorns/Exact/test/de_Sitter.par) Missing: ADM_BSSN
bam in TwoPunctures (AEIThorns/TwoPunctures/test/bam.par) Missing: BAM_ELLIPTIC
idpuncture in TwoPunctures (AEIThorns/TwoPunctures/test/idpuncture.par) Missing: IDPUNCTURE TATELLIPTIC TATJACOBI TATMG TATPETSC
test_legoexcision in LegoExcision (LegoExcision with AHFinder on schwarzschild BH) Missing: ADM_BSSN
outer-buffers in Carpet (Carpet/Carpet/test/outer-buffers.par) Missing: INTERPTOARRAY
waveinterp-1p in CarpetInterp (Carpet/CarpetInterp/test/waveinterp-1p.par) Missing: INTERPTOARRAY
waveinterp-2p in CarpetInterp (Carpet/CarpetInterp/test/waveinterp-2p.par) Missing: INTERPTOARRAY
With the new thorns some tests work (62 out of 101 runnable, 39 failed). The ones that failed are:
test_axibrill (from IDAxiBrillBH) test_axibrill_nostagger (from IDAxiBrillBH) test_ADM_2 (from ADM) schw-0050 (from Cartoon2D) test_cartoon_2 (from Cartoon2D) test_cartoon_3 (from Cartoon2D) onestep (from BSSN_MoL) test_BSSN_MoL_Carpet_1_onelevel (from BSSN_MoL) test_BSSN_MoL_Carpet_1_threelevel (from BSSN_MoL) test_BSSN_MoL_Carpet_1_twolevel (from BSSN_MoL) timerefinement (from BSSN_MoL) checkpoint (from AHFinderDirect) checkpoint2 (from AHFinderDirect) Kerr-definition-expansion-product (from AHFinderDirect) Kerr-modification-radius2 (from AHFinderDirect) Kerr-selection-areal-radius-definition-expansion-product (from AHFinderDirect) recover (from AHFinderDirect) recover2 (from AHFinderDirect) test_ah (from Dissipation) test_ob (from Dissipation) KS-tilted (from Exact) Schwarzschild_EF (from Exact) Kerr-rotating-180-staggered (from RotatingSymmetry180) Kerr-rotating-180 (from RotatingSymmetry180) Kerr-staggered (from RotatingSymmetry180) Kerr (from RotatingSymmetry180) Kerr-rotating-90-staggered (from RotatingSymmetry90) Kerr-rotating-90 (from RotatingSymmetry90) qlm-bl (from QuasiLocalMeasures) qlm-ks-boosted (from QuasiLocalMeasures) qlm-ks-shifted (from QuasiLocalMeasures) qlm-ks-tilted (from QuasiLocalMeasures) qlm-ks (from QuasiLocalMeasures) qlm-minkowski (from QuasiLocalMeasures) whisky_test_tov_ppm (from Whisky) whisky_test_tov_ppm_no_trp (from Whisky) test_BSSN_MoL_Carpet (from CarpetRegrid) test_BSSN_MoL_Carpet_keep (from CarpetRegrid) test_BSSN_MoL_Carpet (from CarpetRegrid2)
test_axibrill fails with a level 0 warnng:
WARNING level 0 in thorn IDAxiBrillBH processor 0 host phys44230.physics.gatech.edu (line 317 of shmgp.F77): -> Trying to find out whether this line is reached
so I assume this was never used...
the Cartoon tests run but fail
AHFinderDirect still fails presumable for the same reason Eloisa in investigating
KS-Tilted needs thorns "CoordBase and CartGrid3D" to work (test fails in 146 out of 253 files once it runs).
Schwarzschild_EF fails in 105 of 112 files
all RotatingSymmetry tests run and produce files but do not compare.
the QuasiLocalMeasures still run but fails as in the first time I tried.
whisky_test_tov_ppm and whisky_test_tov_ppm_no_trp differ in one file each (using BSSN_MoL)
the CarpetRegrid ones using BSSN_MoL run but fail in 37 out of 56 files (each)
I have no idea how to fix the ones using BSSN_MoL given that I have never used it since it is a private thorn of AEI :-)
I have attached the enlarged thornlist should anyone from LSU/AEI who uses BSSN_MoL want to look into this.
Yours, Roland
On Apr 14, 2010, at 11:19 , Roland Haas wrote:
Hello all,
It would be interesting to know if the testsuites which (only) need BSSN_MoL pass if that thorn is included. Can you please test that as well?
Well I have gone ahead and added all the thorns that I could lay my hands on to the executable (still missing are a *thorn* ADM_BSSN, InterpToArray [in AEIDevelopment],BAM_ELLIPTIC, IDLINEARWAVES, NOISE, NORM, TATELLIPTIC, TAT*).
InterpToArray should be made public, and should probably go into CactusUtils.
IDLinearWaves should go into EinsteinInitial, and Noise/Norm into CactusNumeric.
TATelliptic is probably not used -- and if so, it makes sense only together with TATPETSc. This is a non-linear interface to PETSc, whereas EllPETSc is for linear equations only.
Are these on our list?
-erik
On Wed, Apr 14, 2010 at 12:19:12PM -0400, Roland Haas wrote:
With the new thorns some tests work (62 out of 101 runnable, 39 failed). The ones that failed are:
whisky_test_tov_ppm (from Whisky) whisky_test_tov_ppm_no_trp (from Whisky)
Interesting. They worked for me last time I checked which was not more than a week ago. I'll retest.
Frank
users@lists.einsteintoolkit.org
-
Erik Schnetter -
Frank Loeffler -
Roland Haas