Parallel Processing Failure: Boundary condition

[Ananth Narayan S am24d038]

Hello NEK5000 Users and developers

I am facing an issue on parallelizing the code and have found that when boundary condition is set to 'I  ' or 'SYM' instead of 't  ' the code works like charm. Also the code is smooth with 'nek UnipolarEHD' but not with 'nekmpi UnipolarEHD 4'

But with 't  ' the code struggles to solve Passive scalar transport. I am attaching the relevant files (please refer userbc and useric subroutines). I have tried certain combinations but, ultimately I want the a Dirichlet condition on bottom wall and zero gradient on top wall for PS 2. With 'solver = none' for passive scalar 2, I can get the solver run in parallel.

Please note that the geometry is 2D rectangle with the order of boundary id: Left, Top, Bottom, Right.

The screen freezes after, it cannot solve PS 2
          1  Hmholtz PS 1     198   2.7336E+01   1.0000E+00   1.0000E-06   1.0000E+01   F
          1  Hmholtz PS 1     199   2.0748E+01   1.0000E+00   1.0000E-06   1.0000E+01   F
          1  Hmholtz PS 1     200   1.1441E+01   1.0000E+00   1.0000E-06   1.0000E+01   F
          1  Error Hmholtz PS 1     200   1.1441E+01   1.5369E+05   1.0000E-06

Please find attachments for reproducibility.

Thanks

Ananth Narayan