upperwal / EntangledMPI

Fault Tolerance framework for High Performance Computing [Supports ULFM, replication and checkpointing]
MIT License
2 stars 1 forks source link

MPI_Comm_dup is collective call #30

Open upperwal opened 6 years ago

upperwal commented 6 years ago

It is used in MPI_Send and MPI_Recv to create a new comm and differentiate from MPI_COMM_WORLD if rep_error_handler is invoked. But MPI_Comm_dup is collective hence if only some processes are calling MPI_Send and MPI_Recv program will stuck.

Find some other alternate to differentiate comms.

upperwal commented 6 years ago

Reverted to pointer MPI_Comm