[QE-users] QE crashing right after "Starting wfcs"

[Pietro Davide Delugas]

Most likely the processes are killed because they try to allocate more memory per node than allowed in this SKU. This should reported on stderr.

Il 14 ago 2023 2:09 PM, Luis Cebamanos <lcebamanos at lenovo.com> ha scritto:
Hello,

I am trying to find out why QE crashes right after printing out the following:

     number of k points=     1
                       cart. coord. in units 2pi/alat
        k(    1) = (   0.0000000   0.0000000   0.0000000), wk =   2.0000000

     Dense  grid: 27637241 G-vectors     FFT dimensions: ( 540, 540, 540)

     Smooth grid:  9771712 G-vectors     FFT dimensions: ( 375, 375, 375)

     Estimated max dynamical RAM per process >       4.42 GB

     Estimated total dynamical RAM >    1125.78 GB

     Initial potential from superposition of free atoms

     starting charge    5231.9496, renormalised to    5232.0000

     negative rho (up, down):  6.436E-01 0.000E+00
     Starting wfcs are 5168 randomized atomic wfcs

===================================================================================
=   BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
=   RANK 13 PID 220321 RUNNING AT maryland
=   KILLED BY SIGNAL: 9 (Killed)
===================================================================================

I cannot see anything in the logs and I can run well on a different SKU with less cores/node. How could I get a bit more info about this crash? Is there a way to find out what the problem could be?

Regards,

Chris


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20230814/fade935d/attachment.html>