pspw_md QA tests returning NaN when running on more than 24 processors

[edoapra]

Describe the bug
The pspw_md QA tests returning NaN when using more than 24 processors on tahoma using the current master branch ccc2e41

Describe settings used
ARMCI_NETWORK=MPI-TS
FC=gfortran
Singularity image
oras://ghcr.io/edoapra/nwchem/nwchem-dev.mpi-ts.ompi41x:20230710_233358

The hotfix/release-7-2-0 branch is not affected by this bug
pspw_md.output.239354.txt

Attach log files

diff -u pspw_md.ok.out.nwparse pspw_md.out.nwparse
--- pspw_md.ok.out.nwparse	2023-07-11 03:28:56.000000000 -0700
+++ pspw_md.out.nwparse	2023-07-11 03:28:56.000000000 -0700
@@ -1,13 +1,13 @@
 Effective nuclear repulsion energy (a.u.) 85.91
-Total PSPW energy : -13.56964
+Total PSPW energy : nan
 == 0.000 0.000 0.000 
-Total PSPW energy : -13.74176
+Total PSPW energy : nan
 == 0.000 0.000 0.000 
-Total PSPW energy : -13.74176
+Total PSPW energy : nan
 == 0.000 0.000 0.000 
-Total PSPW energy : -14.32395
+Total PSPW energy : nan
 Effective nuclear repulsion energy (a.u.) 85.91
-Total PSPW energy : -14.27678
-Total PSPW energy : -14.33590
-Total PSPW energy : -14.25267
-Total PSPW energy : -14.38971
+Total PSPW energy : nan
+Total PSPW energy : nan
+Total PSPW energy : nan
+Total PSPW energy : nan

[ebylaska]

Ok, I'll try to figure out a fix.

[edoapra]

Commit 3355e82 fixes this issue

[edoapra]

Unfortunately, the bug is still present in the master branch at this time

[ebylaska]

I fixed another bug I found while working on this...Hopefully, have a solution soon.