The calculation repeatedly terminates.

[prerna]

Dear all,

I'm currently utilizing EPW version 5.7 and QE version 7.2 to investigate superconductivity in a material. Initially, everything worked smoothly when I employed a k-point mesh of 12x12x1 in my 'nscf.in' file. However, upon increasing the mesh to 18x18x1 or 24x24x1, I encountered a persistent issue where 'epw.x' unexpectedly terminated before wannerization stage without any error message. I attempted to address this problem by switching to EPW version 5.4.1. This time, the termination occurred after completing the calculation of kgmap, without any error.

Notably, my material consists of only three atoms and I've verified that the disentanglement convergence criteria is satisfied, and the spread is also reasonable.

I'm keen to understand the potential root causes of this behavior. Could this possibly be attributed to a memory issue?

Code: Select all

-------------------------------------------------------------------
     WANNIER      :    153.14s CPU    155.73s WALL (       1 calls)
     -------------------------------------------------------------------

     Calculating kgmap

     Progress kgmap: ########################################
     kmaps        :      0.43s CPU      0.80s WALL (       1 calls)
     Symmetries of Bravais lattice:  24
     Symmetries of crystal:          12


     ===================================================================
     irreducible q point #    1
     ===================================================================

     Symmetries of small group of q: 12
          in addition sym. q -> -q+G:

     Number of q in the star =    1
     List of q in the star:
          1   0.000000000   0.000000000   0.000000000
     Imposing acoustic sum rule on the dynamical matrix

        q(    1 ) = (   0.0000000   0.0000000   0.0000000 )

Many thanks in advance for any advice.

Regards
Prerna
IIT Ropar

[stiwari]

Hi Prerna,

How are you exactly running epw? What kind of machine you are using ? If you are using an hpc with SLURM scheduler or PBS scheduler then it must have a <JOB_ID>.error file (you can ask your scheduler to write one when you submit the job, e.g., #PBS -o err.error) which might contain the error. Also, how many cores are you using for parallelization? To me, it seems like a machine related issue such as memory overflow.

Please provide more details such as your input and output files. If you find the .err file please attach that too.

Best regards,
Sabya.