error in automatic detection of NTAUPAR

[kdoblhoff]

This is a continuation of "error? in automatic determination of ntaupar in rpar calculations" (posted in the "Using Vasp" part of the forum forum/viewtopic.php?f=4&t=18449#p21697) as proper bug report.

The full description of the (potential) bug can be found in the post cited above. In short, in the OUTCAR, the memory requirement is reported to increase as ntaupar decreases and the code ends up using a combination of ntaupar and nomegapar that - according to the outcar - may lead to memory issues, while manually setting NTAUPAR to 2 and NOMEGAPAR to 4 works without getting a warning.

I attach the full input and output as zip.
Code: vasp.6.2.1 16May21 (build Oct 14 2021 10:26:40) complex
Compilation info as far as I can access it (vasp is preinstalled on the national supercomputer that I used for these calculations): foss-2021a-CUDA-11.3.1
Executed on: 1 node with 128 cpu and 1TB of memory in total, using 32 cpus.

[merzuk.kaltak]

Hello,

This problem should should be fixed in vasp 6.3.0 and newer.