Issues running VMEC from default install in raven
Closed this issue · 3 comments
I updated my local STELLOPT repository to the latest commit, and recompiled VMEC2000 by running as per usual ./build_all VMEC2000 in the STELLOPT directory (after setting MACHINE and STELLOPT_PATH) and loading the usual modules module load intel/21.4.0 impi/2021.4 mkl hdf5-mpi netcdf-mpi fftw-mpi.
After compiling, I run a usual VMEC input file (attached), and a segmentation error results:
forrtl: severe (174): SIGSEGV, segmentation fault occurred
Image PC Routine Line Source
xvmec2000 000000000170584A for__signal_handl Unknown Unknown
libpthread-2.31.s 00001515BE8DE910 Unknown Unknown Unknown
libmpi.so.12.0.0 00001515BAB081E0 MPI_Sendrecv Unknown Unknown
libmpifort.so.12. 00001515BBCF2515 pmpi_sendrecv Unknown Unknown
xvmec2000 000000000062EC24 Unknown Unknown Unknown
xvmec2000 000000000050BE59 Unknown Unknown Unknown
xvmec2000 0000000000518F05 Unknown Unknown Unknown
xvmec2000 00000000005573C5 Unknown Unknown Unknown
xvmec2000 0000000000617FC3 Unknown Unknown Unknown
xvmec2000 00000000006211B0 Unknown Unknown Unknown
xvmec2000 00000000006175A9 Unknown Unknown Unknown
xvmec2000 0000000000622922 Unknown Unknown Unknown
xvmec2000 0000000000410112 Unknown Unknown Unknown
libc-2.31.so 00001515B9E3E24D __libc_start_main Unknown Unknown
xvmec2000 000000000041002A Unknown Unknown Unknown
The run is done with (within a batch script):
module purge
module load intel/21.4.0 impi/2021.4 mkl hdf5-mpi netcdf-mpi fftw-mpi
export LD_LIBRARY_PATH=${LD_LIBRARY_PATH}:${NETCDF_HOME}/lib:${MKL_HOME}/lib/intel64
srun /u/rodre/bin/xvmec2000 input.temp
Any idea of what the issue could be? Previous compilations did work.
@kudav More detailed error output:
forrtl: error (65): floating invalid
Image PC Routine Line Source
xvmec2000 00000000020C627B Unknown Unknown Unknown
libpthread-2.31.s 000014D1753B3910 Unknown Unknown Unknown
xvmec2000 00000000009D7D7D funct3d_par_ 173 funct3d.f
xvmec2000 0000000000CAC300 evolve_ 134 evolve.f
xvmec2000 0000000000CC3E87 eqsolve_ 91 eqsolve.f
xvmec2000 0000000000CAA1B5 runvmec_ 365 runvmec.f
xvmec2000 0000000000CC73DD MAIN__ 328 vmec.f
xvmec2000 0000000000410312 Unknown Unknown Unknown
libc-2.31.so 000014D17083E24D __libc_start_main Unknown Unknown
xvmec2000 000000000041022A Unknown Unknown Unknown
@SebereX This should now be corrected. Please try again with the latest develop and report back.
This is now running without the previous parallelisation issues appearing.