SPECFEM / specfem3d

SPECFEM3D_Cartesian simulates acoustic (fluid), elastic (solid), coupled acoustic/elastic, poroelastic or seismic wave propagation in any type of conforming mesh of hexahedra (structured or not).
GNU General Public License v3.0
390 stars 223 forks source link

'adios2_debug_mode_*' make error #1596

Closed brioglade closed 1 year ago

brioglade commented 1 year ago

hi, can't make completely with asdf and adios2 configure. How to fix it?

the version of gcc g++ gfortran is (Ubuntu 11.3.0-1ubuntu1~22.04) 11.3.0.

Error: Symbol ‘adios2_debug_mode_off’ at (1) has no IMPLICIT type src/shared/adios_manager.F90:243:67:

243 | call adios2_init(myadios2_obj, comm_adios, adios2_debug_mode_on, ier) | 1

Error: Symbol ‘adios2_debug_mode_on’ at (1) has no IMPLICIT type make: *** [src/shared/rules.mk:194: obj/adios_manager.shared_adios_module.o] Error 1

danielpeter commented 1 year ago

unfortunately, the ADIOS2 library decided to go with a non-backward compatible init routine in its newest version.

until we submit a fix for the newest ADIOS2 version intialization routine call, please downgrade ADIOS2 to a version <= 2.8.3

danielpeter commented 1 year ago

looks like you'll have to remove the installed ADIOS2 binaries and libraries by hand, i.e., sudo rm -f ...

brioglade commented 1 year ago

The solver can't stop itself. There is an error with "ASDF_FORMAT = .true." . However, the solver result is OK.

running database generation on 32 processors...

running solver on 32 processors...

^C[mpiexec@LEGION-IAH7H] Sending Ctrl-C to processes as requested [mpiexec@LEGION-IAH7H] Press Ctrl-C again to force abort HDF5: infinite loop closing library L,T_top,F,P,P,Z,FD,VL,FD,VL,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD HDF5: infinite loop closing library L,T_top,F,P,P,Z,FD,VL,FD,VL,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD HDF5: infinite loop closing library L,T_top,F,P,P,Z,FD,VL,FD,VL,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD,FD HDF5: infinite loop closing library

danielpeter commented 1 year ago

thanks for pointing out!

ASDF and the underlying (parallel) HDF library might show some hiccups depending on the versions installed on your system. If you encounter a stall in the solver where in the output_solver.txt nothing is added after:

..
 creating ASDF file: ./OUTPUT_FILES/synthetic.h5
 writing waveforms in parallel...

then try to turn on the flag WRITE_SEISMOGRAMS_BY_MAIN in the simulation Par_file:

# decide if main process writes all the seismograms or if all processes do it in parallel
WRITE_SEISMOGRAMS_BY_MAIN       = .true.

this can help with saving the seismograms in the ASDF format.