[FLASH-USERS] Problems with RadBlastWave, ReinickeMeyer, and CriticalRadShock

Tue Apr 22 15:36:54 EDT 2025

(My sincere apologies if this is the second instance of this email, I got a bounce back from the first time and I don't know if it actually went through)

Dear FLASH Team,

I have been running into various problems getting some of the sample simulations running, namely radflaHD/RadBlastWave, the ReinickeMeyer simulation, and CriticalRadShock.  I have successfully compiled and have run both Sedov and magnetoHD/BlastBS.

I can't get the other three to compile unless I remove the option "+parallelio".  I have tried installing HDF5 and HYPRE with many permutations, such as installing HDF5 with cmake and using command "CC=/home/me/intel/oneapi/2025.1/bin/mpicc ./configure --enable-parallel", and I've also tried using the last version of HYPRE (v2.28, whereas current = 2.33) that Adam R. said he's tested in a recent email.   If I remove  the "+parallelio" option they all compile successfully, but then they fail when run.  It's my understanding that these should be able to run successfully 'out of the box' so I am concerned I've got myself a mess (but HDF5 works great for me with Sedov and BlastBS).

==> For RadBlastWave (setup with +serialio), compiled successfully but failed when run:

....
     342 1.7511E-06 1.4637E-12  ( 1.660E+12,   0.00    ,   0.00    ) |  1.173E-04 1.907E+49 1.714E-09
     343 1.7511E-06 1.5368E-12  ( 1.855E+12,   0.00    ,   0.00    ) |  1.173E-04 1.907E+49 3.559E-12
     344 1.7511E-06 1.6137E-12  ( 1.660E+12,   0.00    ,   0.00    ) |  1.129E-04 1.907E+49 1.714E-09
     345 1.7511E-06 1.6944E-12  ( 1.855E+12,   0.00    ,   0.00    ) |  1.129E-04 1.907E+49 4.379E-12
 DRIVER_ABORT: [hy_uhd_unsplitUpdateCastroLike] Negative 3T internal energy, CHECK LOG
Abort(1) on node 1 (rank 1 in comm 0): application called MPI_Abort(MPI_COMM_WORLD, 1) - process 1

==> For RadBlastWave (setup with +parellelio), compilation error:

.....
age.o workspace.o -L /home/adamkowalski/Desktop/Projects/FLASH/hdf5-1.14.6/hdf5/lib/ -lhdf5 -lz -L/home/adamkowalski/Desktop/Projects/FLASH/FLASH4.8/hypre/hypre/src/hypre/lib -lHYPRE  -L/home/adamkowalski/.local/lapack -llapack
/usr/bin/ld: io_h5file_interface.o: in function `io_h5init_file_':
/home/adamkowalski/Desktop/Projects/FLASH/FLASH4.8/object/io_h5file_interface.c:165:(.text+0xa3): undefined reference to `H5Pset_fapl_mpio'
/usr/bin/ld: io_h5file_interface.o: in function `io_h5open_file_for_read_':
/home/adamkowalski/Desktop/Projects/FLASH/FLASH4.8/object/io_h5file_interface.c:273:(.text+0x1d0): undefined reference to `H5Pset_fapl_mpio'
/usr/bin/ld: io_h5read_unknowns.o: in function `io_h5read_unknowns_':
/home/adamkowalski/Desktop/Projects/FLASH/FLASH4.8/object/io_h5read_unknowns.c:55:(.text+0x38b): undefined reference to `H5Pset_dxpl_mpio'
/usr/bin/ld: io_h5write_unknowns.o: in function `io_h5write_unknowns_':
/home/adamkowalski/Desktop/Projects/FLASH/FLASH4.8/object/io_h5write_unknowns.c:166:(.text+0x524): undefined reference to `H5Pset_dxpl_mpio'
/usr/bin/ld: io_h5write_unknowns_sp.o: in function `io_h5write_unknowns_sp_':
/home/adamkowalski/Desktop/Projects/FLASH/FLASH4.8/object/io_h5write_unknowns_sp.c:251:(.text+0x503): undefined reference to `H5Pset_dxpl_mpio'
collect2: error: ld returned 1 exit status
make: *** [Makefile:384: flash4] Error 1

==>     For ReinickeMeyer (setup with default serial io), compiled successfully but failed when run:
....
    317 1.9368E-10 2.7894E-22  ( 4.581E-01,  0.000E+00,  0.000E+00) |  1.922E-13 2.789E-22
     318 1.9368E-10 6.3350E-27  ( 4.586E-01,  0.000E+00,  0.000E+00) |  1.073E-13 6.335E-27
     319 1.9368E-10 1.2288E-31  ( 4.591E-01,  0.000E+00,  0.000E+00) |  4.673E-14 1.229E-31
     320 1.9368E-10 1.0000E-31  ( 4.596E-01,  0.000E+00,  0.000E+00) |  1.996E-14 1.890E-36
[gr_hypreSolve]: Nonconv./failure ierr=256, descr=[Method did not converge], component=0, converged=F, num_iterations=290, final_res_norm=1.3457733013687052E+023
     321 1.9368E-10 1.0000E-31  ( 4.601E-01,  0.000E+00,  0.000E+00) |  1.085E-19 6.745-105
[gr_hypreSolve]: Nonconv./failure ierr=1, descr=[Generic error], component=0, converged=F, num_iterations=0, final_res_norm=0.0000000000000000
 DRIVER_ABORT: [Diffuse]: computed dt is not positive! Aborting!
Abort(1) on node 5 (rank 5 in comm 0): application called MPI_Abort(MPI_COMM_WORLD, 1) - process 5

==> CritRadShock, compiled successfully with serialio, but failed to run (fails if try to setup with +parallelio):

 set: Can not change name with constant attribute:useHeatexchange

 Driver_abort called. See log file for details.
 Error message is ERROR: unable to change constant name
 Calling MPI_Abort() for immediate shutdown!

 Logfile_open:  Error: cannot open log file; io_status =           29
 Logfile_open:  Error: cannot open log file; io_status =           29
Abort(1) on node 0 (rank 0 in comm 0): application called MPI_Abort(MPI_COMM_WORLD, 1) - process 0

Thank you very much for any guidance you can offer.

Adam

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://flash.rochester.edu/pipermail/flash-users/attachments/20250422/4673799a/attachment.htm>