Hi,
I used to visualize the positions of NaNs simply by using contour plot of visit. But I failed to visualize the data of NaNmask.h5 generated by recent ET and GRHydro. The output says NaNChecker found NaNs and terminated the run. I tweaked the following parameters but got no difference:
NaNChecker::ignore_restricted_points = "no"
IOHDF5::output_symmetry_points = no IOHDF5::out3D_ghosts = no
I tested even the logging level dependency. The crash time depends on it but no difference for NaN visualization. Actually the above hdf5 setup makes very weird Carpet grids but might do only on Visit.
I'm using the latest version of stable ET. My visit has no problem in displaying NaNmask.h5 made by old ET.
Thanks for your help in advance,
Hee Il
Hee Il
Are you setting NaNChecker::action_if_found = "terminate"? If you set it to "abort", then NaNChecker will abort before the nan mask can be output.
-erik
On Apr 29, 2014, at 11:30 , Hee Il Kim heeilkim@gmail.com wrote:
Hi,
I used to visualize the positions of NaNs simply by using contour plot of visit. But I failed to visualize the data of NaNmask.h5 generated by recent ET and GRHydro. The output says NaNChecker found NaNs and terminated the run. I tweaked the following parameters but got no difference:
NaNChecker::ignore_restricted_points = "no"
IOHDF5::output_symmetry_points = no IOHDF5::out3D_ghosts = no
I tested even the logging level dependency. The crash time depends on it but no difference for NaN visualization. Actually the above hdf5 setup makes very weird Carpet grids but might do only on Visit.
I'm using the latest version of stable ET. My visit has no problem in displaying NaNmask.h5 made by old ET.
Thanks for your help in advance,
Hee Il
Users mailing list Users@einsteintoolkit.org http://lists.einsteintoolkit.org/mailman/listinfo/users
Yes, I chose the action "terminate" to make the nan mask. But if Con2Prim of GRHydro finds NaN first, then the run is terminated before making the output of the nan mask.
Hee Il
2014-04-29 18:35 GMT+09:00 Erik Schnetter schnetter@cct.lsu.edu:
Hee Il
Are you setting NaNChecker::action_if_found = "terminate"? If you set it to "abort", then NaNChecker will abort before the nan mask can be output.
-erik
On Apr 29, 2014, at 11:30 , Hee Il Kim heeilkim@gmail.com wrote:
Hi,
I used to visualize the positions of NaNs simply by using contour plot
of visit. But I failed to visualize the data of NaNmask.h5 generated by recent ET and GRHydro. The output says NaNChecker found NaNs and terminated the run. I tweaked the following parameters but got no difference:
NaNChecker::ignore_restricted_points = "no"
IOHDF5::output_symmetry_points = no IOHDF5::out3D_ghosts = no
I tested even the logging level dependency. The crash time depends on it
but no difference for NaN visualization. Actually the above hdf5 setup makes very weird Carpet grids but might do only on Visit.
I'm using the latest version of stable ET. My visit has no problem in
displaying NaNmask.h5 made by old ET.
Thanks for your help in advance,
Hee Il
Users mailing list Users@einsteintoolkit.org http://lists.einsteintoolkit.org/mailman/listinfo/users
-- Erik Schnetter schnetter@cct.lsu.edu http://www.perimeterinstitute.ca/personal/eschnetter/
My email is as private as my paper mail. I therefore support encrypting and signing email messages. Get my PGP key from http://pgp.mit.edu/.
On Tue, Apr 29, 2014 at 06:40:19PM +0900, Hee Il Kim wrote:
Yes, I chose the action "terminate" to make the nan mask. But if Con2Prim of GRHydro finds NaN first, then the run is terminated before making the output of the nan mask.
Hi,
Can you try setting GRHydro::GRHydro_c2p_failed_action = "terminate"? "abort" is default, and does what you described. "terminate" should allow the mask to be written.
Frank
Hi
2014-04-30 1:15 GMT+09:00 Frank Loeffler knarf@cct.lsu.edu:
On Tue, Apr 29, 2014 at 06:40:19PM +0900, Hee Il Kim wrote:
Yes, I chose the action "terminate" to make the nan mask. But if Con2Prim of GRHydro finds NaN first, then the run is terminated before making the output of the nan mask.
Hi,
Can you try setting GRHydro::GRHydro_c2p_failed_action = "terminate"? "abort" is default, and does what you described. "terminate" should allow the mask to be written.
Frank
The output could be made by GRHydro::GRHydro_c2p_failed_action = "terminate" but it also has no information on the location of NaNs. I found an older NaNChecker works to locate NaNs. I couldn't find any version info, so I attached below a diff file.
Hee Il
#### diff -r NaNChecker.ET NaNChecker.old ####
Only in NaNChecker.ET/doc: .svn diff -r NaNChecker.ET/interface.ccl NaNChecker.old/interface.ccl 31,36d30 < CCTK_INT FUNCTION \ < GetRefinementLevel \ < (CCTK_POINTER_TO_CONST IN cctkGH) < USES FUNCTION GetRefinementLevel < < diff -r NaNChecker.ET/param.ccl NaNChecker.old/param.ccl 54,64d53 < < BOOLEAN ignore_restricted_points "do not check grid points whose values will be restricted away" STEERABLE = ALWAYS < { < } "no" < < STRING restriction_mask "grid function to use to decide which points are restricted away, points where the mask is zero are ignored" STEERABLE = ALWAYS < { < "CarpetReduce[:][:]weight" :: "Carpet's reduction mask" < "CarpetEvolutionMask[:][:]evolution_mask" :: "takes prolongation stencil into account" < ".*[:][:].*" :: "any grid function with points masked out set to zero" < } "CarpetReduce::weight" diff -r NaNChecker.ET/schedule.ccl NaNChecker.old/schedule.ccl 8c8 < schedule NaNChecker_ResetCounter at BASEGRID ---
schedule NaNChecker_ResetCounter at INITIAL
20,30c20 < schedule NaNChecker_NaNCheck_Prepare IN NaNChecker_NaNCheck < { < LANG: C < OPTIONS: level < } "Prepare data structures to check for NaNs" < < schedule GROUP NaNChecker_NaNCheck as zzz_NaNChecker_NaNCheck at POSTSTEP < { < } "Check for NaNs and count them in NaNChecker::NaNsFound" < < schedule NaNChecker_NaNCheck_Check IN NaNChecker_NaNCheck AFTER NaNChecker_NaNCheck_Prepare ---
schedule NaNChecker_NaNCheck at POSTSTEP
34,49d23 < } "Check for NaNs" < < schedule NaNChecker_NaNCheck_Finish IN NaNChecker_NaNCheck AFTER NaNChecker_NaNCheck_Check < { < LANG: C < OPTIONS: level < } "Count NaNs in NaNChecker::NaNsFound" < < schedule NaNChecker_TakeAction at POSTSTEP after zzz_NaNChecker_NaNCheck < { < LANG: C < OPTIONS: global loop-level < } "Output NaNChecker::NaNmask and take action according to NaNChecker::action_if_found" < < schedule GROUP NaNChecker_NaNCheck as zzz_NaNChecker_NaNCheck at POST_RECOVER_VARIABLES < { 52c26 < schedule NaNChecker_TakeAction at POST_RECOVER_VARIABLES after zzz_NaNChecker_NaNCheck ---
schedule NaNChecker_TakeAction at POSTSTEP after NaNChecker_NaNCheck
diff -r NaNChecker.ET/src/make.code.defn NaNChecker.old/src/make.code.defn 5c5 < SRCS = NaNCheck.cc ---
SRCS = NaNCheck.c
Only in NaNChecker.old/src: NaNCheck.c Only in NaNChecker.ET/src: NaNCheck.cc diff -r NaNChecker.ET/src/NaNCheck.h NaNChecker.old/src/NaNCheck.h 16d15 < namespace NaNChecker { 32d30 < } // end namespace NaNChecker Only in NaNChecker.ET/src: .svn Only in NaNChecker.ET: .svn Only in NaNChecker.ET: test
users@lists.einsteintoolkit.org