multi simulations should report that they are multi simulations
The .log files for an 8-simulation system on a 32-core machine running 4 OpenMP threads per MPI process report
Using 1 MPI process Using 4 OpenMP threads
which is fine, but there should also be a diagnostic that -multi is working the way the user intended it.
#3 Updated by Mark Abraham over 6 years ago
Also, non-REMD multi simulations do not check for .tpr consistency, e.g. with number of simulation steps. Not sure if that's a feature or a bug. md_support.c tries to check the number of steps, but the code is broken (e.g. use of MASTER, not SIMMASTER; and/or subsequent use of gmx_bcast, which only functions within a simulation)