segfault in diamond example

[max-radin]

Hi, I'm getting a segfault on the CNBSE step of the diamond example. I'm running ocean 3.0.0 in an ubuntu container with 16gb of memory. I tried decreasing ecut and other parameters but got the same result. Any advice?

Here is the last part of CNBSE/ocean.log:

 Initialization complete           0
   mult =  0.44089472E-02
   1.0000000000000000              100

Program received signal SIGSEGV: Segmentation fault - invalid memory reference.

Backtrace for this error:
#0  0xffffbc4ad08b in ???
#1  0xffffbc4ac047 in ???
#2  0xffffbcd3c7bf in ???
#3  0xffffbc2a7c40 in ???
#4  0xffffba1bfd17 in ???
#5  0xffffba1d251f in ???
#6  0xffffba1d264f in ???
#7  0xffffba148193 in ???
#8  0xffffba19f307 in ???
#9  0xffffba19f55b in ???
#10  0xffffba1ad3f3 in ???
#11  0xffffba1adb8b in ???
#12  0xffffba1aec43 in ???
#13  0xffffba0467fb in MPI_Reduce
#14  0xffffbc66fc97 in mpi_reduce__
#15  0xaaaad31a737b in __ocean_psi_MOD_ocean_psi_send_buffer
#16  0xaaaad31c02c7 in __ocean_action_MOD_ocean_xact.constprop.0
#17  0xaaaad31aac7b in MAIN__
#18  0xaaaad31a38e3 in main

===================================================================================
=   BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
=   PID 7124 RUNNING AT 743548cd03b8
=   EXIT CODE: 9
=   CLEANING UP REMAINING PROCESSES
=   YOU CAN IGNORE THE BELOW CLEANUP MESSAGES
===================================================================================
YOUR APPLICATION TERMINATED WITH THE EXIT STRING: Segmentation fault (signal 11)
This typically refers to a problem with your application.
Please see the FAQ page for debugging suggestions

[jtv3]

1. Is this unique to the diamond example?
2. What compilers are you using (including version) and what MPI library?
3. Please recompile with some debugging flags (on gfortran this would be "-g -0g -fbacktrace"

[max-radin]

The backtrace with debugging flags is below. I'm using gfortran 11.2.0 and mpich.

Backtrace for this error:
#0  0xffff8572d08b in ???
#1  0xffff8572c047 in ???
#2  0xffff85fbe7bf in ???
#3  0xffff85527c40 in ???
#4  0xffff8343fd17 in ???
#5  0xffff8345251f in ???
#6  0xffff8345264f in ???
#7  0xffff833c8193 in ???
#8  0xffff8341f307 in ???
#9  0xffff8341f55b in ???
#10  0xffff8342d3f3 in ???
#11  0xffff8342db8b in ???
#12  0xffff8342ec43 in ???
#13  0xffff832c67fb in ???
#14  0xffff858efc97 in ???
#15  0xaaaacef574b3 in core_reduce_send
    at /app/OCEAN/OCEAN2/OCEAN_psi.f90:825
#16  0xaaaacef574b3 in __ocean_psi_MOD_ocean_psi_send_buffer
    at /app/OCEAN/OCEAN2/OCEAN_psi.f90:465
#17  0xaaaacef574b3 in __ocean_psi_MOD_ocean_psi_send_buffer
    at /app/OCEAN/OCEAN2/OCEAN_psi.f90:434
#18  0xaaaacef71307 in __ocean_action_MOD_ocean_xact.constprop.0
    at /app/OCEAN/OCEAN2/OCEAN_action.f90:514
#19  0xaaaacef5acbb in ocean_haydock_herm_do
    at /app/OCEAN/OCEAN2/OCEAN_haydock.f90:258
#20  0xaaaacef5acbb in __ocean_haydock_MOD_ocean_haydock_do
    at /app/OCEAN/OCEAN2/OCEAN_haydock.f90:186
#21  0xaaaacef5acbb in __ocean_driver_MOD_ocean_driver_run
    at /app/OCEAN/OCEAN2/OCEAN_driver.f90:40
#22  0xaaaacef5acbb in ocean
    at /app/OCEAN/OCEAN2/OCEAN.f90:65
#23  0xaaaacef53923 in main
    at /app/OCEAN/OCEAN2/OCEAN.f90:9

I've found that the diamond example does work in my environment when the number of MPI processes is reduced to 1. So for the time being I plan to continue my project using just a single process.

I haven't gotten any other example to work yet, although the errors seem unrelated. For example, I get the following error from the STO example:

mpirun -n 4 /app/bin/pw.x  -npool 4 -inp nscf.in > nscf.out 2>&1
QE62!!
5.55651  8.30299  6.92975
0.204198  0.305129  0.254664
Fractional nuber of electrons, but fixed occupations
DFT Stage Failed

[max-radin]

Ok it looks like the failure in the STO example is because the QE parser is not expecting a + in the exponent of the nelec tag. If no one is already working on this, I can make a PR to handle this case.

[jtv3]

go ahead and open a PR against the develop branch