Anatase PAW relaxation

[ilukacevic]

Dear all,

I'm trying to relax anatase phase of TiO2 using paw datasets. However, when it starts the 1st iteration of 1st broyden step, the calculation stops and gives a lots of these lines:

=== Connectors / Fallbacks ===
Connectors on : yes
Fallbacks on : yes
DFT flavor : none
FFT flavor : none
LINALG flavor : netlib-fallback
[0] 16 at [0x00000000054bfe28], cart_sub.c[169]
[0] 16 at [0x00000000054c0cc8], cart_sub.c[167]
[0] 16 at [0x00000000054c0878], cart_sub.c[165]
[0] 56 at [0x00000000054c0a08], cart_sub.c[158]
[0] 32 at [0x00000000054c0ef8], mpid_vc.c[93]
[0] 40 at [0x00000000054c0e28],lmpid_vc.c[93]
[0] 8 at [0x00000000054c0c18], lolocal_proc.c[90]
[3] 40 at [0x00000000055b30c8], mpid_vc.c[93]
[3] 8 at [0x00000000055b2e88], cart_sub.c[169]
[3] 8 at [0x00000000055b2dd8], cart_sub.c[167]
[3] 8 at [0x00000000055b2988], cart_sub.c[165]
[3] 56 at [0x00000000055b2b18], cart_sub.c[158]
[3] 40 at [0x00000000055b2f38], mpid_vc.c[93]
[3] 8 at [0x00000000055b2d28], local_proc.c[91]
[3] 8 at [0x00000000055b2c78], local_proc.c[90]
[3] 40 at [0x00000000055b2a48], mpid_vc.c[93]
[3] 8 at [0x00000000055b2548], cart_sub.c[169]
[3] 8 at [0x00000000055b2498], cart_sub.c[167]
[3] 8 at [0x00000000055b2308], cart_sub.c[165]
[3] 56 at [0x00000000055b1bb8], cart_sub.c[158]
[3] 40 at [0x00000000055b28b8], mpid_vc.c[93]
[3] 8 at [0x00000000055b26a8], local_proc.c[91]
[3] 8 at [0x00000000055b25f8], local_proc.c[90]
[3] 40 at [0x00000000055b23c8], mpid_vc.c[93]
[3] 8 at [0x00000000055b1b08], cart_sub.c[169]
[3] 8 at [0x00000000055b1e68], cart_sub.c[167]
[3] 8 at [0x00000000055b1db8], cart_sub.c[165]
[3] 56 at [0x00000000055b20b8], cart_sub.c[158]
8 at [0x00000000054bf468], cart_sub.c[167]
[0] 8 at [0x00000000054bf3b8], cart_sub.c[165]
...
[1] 40 at [0x0000000005c5a388], [3] 56 at [0x0[1] 8 at [0x0000000005c599f8], cart_sub.c[169]
[1] 8 at [0x0000000005c59d58], cart_sub.c[167]
[1] 8 at [0x0000000005c59ca8], cart_[3] 56 at [0x00000000055b1cd8], mpid_vc.c[93]
[3] 8 at [0x00000000055b1a58], cart_sub.c[169]
[3] 8 at [0x00000000055b15c8], cart_sub.c[167]
[3] 8 at [0x00000000055b1518], cart_sub.c[165]
[3] 56 at [0x00000000055b1798], cart_sub.c[158]
[3] 8 at [0x00000000055b19a8], local_proc.c[91]
[3] 8 at [0x00000000055b18f8], local_proc.c[90]
[3] 32 at [0x00000000055b16c8], mpid_vc.c[93]
[node014.srce.cro-ngi:mpispawn_0][child_handler[3] 8 at [0x00000000055b1258], cart_sub.c[167]
[3] 8 at [0x00000000055b10c8], [node014.srce.cr[3] 56 at [0x00000000055b0978], cart_sub.c[158]
[3] 8 at [0x00000000055b1468], local_proc.c[91]
[3] 8 at [0x00000000055b13b8], local_proc.c[90]
[3] 32 at [0x00000000055b1188], [node014.srce.[3] 8 at [0x00000000055b08c8], cart_sub.c[169]
...


Does anyone know what this means and how it could be solved? I'm using abinit version 7.4.3 on a linux cluster.

Thanks in advance.

Igor Lukacevic

[sponce]

Dear Igor,

It look to me that the issue is not related to the input file but rather to the way Abinit is compiled. Does Abinit passes all the tests ("make tests") on that machine?

If not, could you provide the option you used for the compilation? (the "machine_name.ac" file)

Regards,

Samuel Poncé

[gmatteo]

Does anyone know what this means and how it could be solved? I'm using abinit version 7.4.3 on a linux cluster.

It seems that Abinit is crashing inside one of the MPI routines. Could you post the entire log file of the calculation?
What happens if you use paral_kgb=0?