[Developers] Steerable parameter of AHFinder (Erik)

Yosef Zlochower yosef at astro.rit.edu
Fri Jul 20 06:00:08 CDT 2007


Hi,

  I see this behavior with the 'Erik' tagged version of AHFinderDirect. 
Apparently,
at least for some architectures, if horizon n was never found, then on 
recovery,
the grid function ah_radius[n] will be filled with zeros. It is these 
zeros that trigger
the abort (although I don't understand why they are ever used ... ).
 As a workaround what I do is modify the function
void BH_diagnostics::load(CCTK_ARGUMENTS,  const int horizon_number) in
AHFinderDirect/src/driver/BH_diagnostics.cc
so that it changes a radius of zero to a non-zero value.

Yosef

Luca Baiotti wrote:
> Hi,
>
> I have steered the parameters AHFinderDirect::find_after_individual[n]
> at recovery (to an earlier time, so that at the recovery time it should
> havfe already searched for the horizon) and this caused a crash in
> AHFinder. Does anyone have experience with this?
>
> The stderr/out says:
>
> INFO (AHFinderDirect): proc 0: searching for horizon 1/3
> cactus_belladonna_NoDebug_allThorns:
> /data1/baiotti/Cactus/configs/belladonna_NoDebug_allT
> horns/build/AHFinderDirect/patch/coords.cc:313: Assertion
> `jtutil::fuzzy<fp>::NE(r, 0.0)'
> failed.
> Abort (core dumped)
>
>
>
>
>
> The backtrace says:
>
> #0  0x00002aaaab69143a in raise () from /lib64/tls/libc.so.6
> #1  0x00002aaaab692870 in abort () from /lib64/tls/libc.so.6
> #2  0x00002aaaab68aaeb in __assert_fail () from /lib64/tls/libc.so.6
> #3 0x000000000124a2b8 in
> partial_xyz_wrt_r_mu_nu__Q2_14AHFinderDirect12local_coordsFdN21RdN84 ()
> #4 0x000000000121f05a in
> partial_xyz_wrt_r_rho_sigma__Q2_14AHFinderDirect7z_patchCFdN21RdN84 ()
> #5 0x000000000121c57b in
> rho_sigma_metric__Q2_14AHFinderDirect5patchCFdN91T1RdRdRd ()
> #6 0x000000000121d355 in
>
> __CPR94__integrate_gridfn__Q2_14AHFinderDirect5patchCFiN71Q3_J21JJ37J18integration_method
> ()
> #7 0x000000000122c42e in
>
> __CPR106__integrate_gridfn__Q2_14AHFinderDirect12patch_systemCFibN22N71Q3_J21J5patch18integration_method
> ()
> #8 0x000000000120cad6 in
>
> __CPR435__Newton__14AHFinderDirectFPC4_cGHiN22RQ2_J8J16horizon_sequencePCPQ2_J8J7AH_dataRCQ2_J8J16cactus_grid_infoRCQ2_J8J13geometry_infoRCQ2_J8J13JJacobian_infoRCQ2_J8J11solver_infoRCQ2_J8J7IO_infoRCQ2_J8J19BH_diagnostics_infobRCQ2_J8J10error_infoRCQ2_J8J12verbose_infoRQ2_J8J24iteration_status_buffers
> ()
> #9  0x00000000012089fa in AHFinderDirect_find_horizons ()
> #10 0x0000000000409345 in CCTK_CallFunction ()
> #11 0x000000000134edb9 in Carpet::CallFunction ()
> #12 0x000000000040b4e0 in CCTKi_ScheduleCallFunction ()
> #13 0x000000000040e1b1 in ScheduleTraverseFunction ()
> #14 0x000000000040df9b in ScheduleTraverseGroup ()
> #15 0x000000000040ddf2 in CCTKi_DoScheduleTraverse ()
> #16 0x000000000040a328 in ScheduleTraverse ()
> #17 0x0000000000409aa6 in CCTK_ScheduleTraverse ()
> #18 0x00000000012f3daa in Carpet::ScheduleTraverse ()
> #19 0x00000000012f0f9e in Carpet::CallPostRecoverVariables ()
> #20 0x00000000012f08ba in Carpet::Initialise ()
> #21 0x000000000040566b in main ()
>
> _______________________________________________
> Developers mailing list
> Developers at cactuscode.org
> http://www.cactuscode.org/mailman/listinfo/developers
>   



More information about the Developers mailing list