This isn't a CASTEP error, this is an error from the MPI library. Have you successfully run other CASTEP calculations on this machine? Have you checked whether there are any error or log files for this calculation?

Philip James Hasnip Thanks for responding.

Actually 3 calculations ran successfully in CASTEP (Material Studio 2020) but after that I am getting this message after the job failure

[mpiexec@Rahat-Saqfi] ..\hydra\pm\pmiserv\pmiserv_cb.c (863): connection to proxy 0 at host Rahat-Saqfi failed

[mpiexec@Rahat-Saqfi] ..\hydra\tools\demux\demux_select.c (103): callback returned error status

[mpiexec@Rahat-Saqfi] ..\hydra\pm\pmiserv\pmiserv_pmci.c (520): error waiting for event

[mpiexec@Rahat-Saqfi] ..\hydra\ui\mpich\mpiexec.c (1149): process manager error waiting for completion

I checked in Project log, it only says that log file has been created and nothing else but I am getting job failures.

I don't have much experience with Intel's MPI, unfortunately. It's probably best to contact BIOVIA Support, I'm sure they'll have seen this before. Alternatively, you could ask Intel directly, but they aren't always very good at responding.

In the mean time, you could check the hydra service is actually running, by using:

hydra_service.exe -status

And you can test the Gateway etc using the Licence Administrator tools.

If all else fails, you can get the free academic CASTEP licence and just compile and run CASTEP yourself, with the input files generated by Materials Studio.

may i ask have you resolved this problem? i met the same question and now i have no idea to fix it. i think this comes from some faults about parameters. when i chose BFGS, it could be solved sometimes, and it appears when i use TPSD.

Jierui Zhang this appears to be an issue with the MPI itself, and is probably not related to CASTEP at all. I suppose it's possible that it's triggered if CASTEP aborts with an error message -- you could look on the server and see whether there are any CASTEP error files; if your main CASTEP inputs were called mycalc.param and mycalc.cell, then error files would be called mycalc.0001.err, mycalc.0002.err etc.

The only other thing I can think of is whether your calculation is running out of RAM, and that is triggering some problems in the MPI system. You can check the CASTEP output file for its estimate of how much RAM it needs, it isn't 100% accurate but it should give you a rough idea.