Skip to content

Deadlock problems with the Send/Irecv approach... use now Isend/Irecv +...

Anthony Royer requested to merge openmpi_deadlock into master

Deadlock problems with the Send/Irecv approach... use now Isend/Irecv + MPI_Wait /right after/ each exchange (WIP: seems ok when the number of subdomains is equal to the number of MPI processes, but reaches a deadlock otherwise...)

Merge request reports