bosilca

Looking at the last 2 years of updates in the DPM related code, many of us (you/ICM/LANL/Amazon/UTK) tried to do so. Smaller steps but it got to a point where...

1. It does not need to be OMPI_DECLSPEC, we don't expose our internal output streams. 2. `libmca_common_monitoring.la` is the common support for monitoring in other frameworks. It is included in...

The stubs point to a libcuda.so that allows linking CUDA applications using the driver API (such as OMPI) on platforms without GPUs. This is different from what other libraries require,...

The current approach of packing/unpacking device memory issues one cudaMemcpy per item in the indexed datatype. The performance will be terrible. This is something I wanted to take a look...

You are stracing yout mpirun process, not your 3 application processes, so it is difficult to see why it deadlock. The last line is a blocking `poll`, it is a...

This is far from complete, because `MPI_APPNUM` is a required predefined attribute on `MPI_COMM_WORLD`. It shall not matter under which batch scheduler the application is started, the attribute must exists....

It does not matter what SLURM provides or not, they are not an MPI implementation, so they cannot care less of what the MPI standard requires. OMPI is an MPI...

> Passing cleaned parameters seems sensible. But what about allreduce and alltoall? Neither Allreduce or Alltoall provide the opportunity for the user to provide a buffer that shall not be...

The _stringtag_ is not about identifying the group, because otherwise the MPI standard would be standardizing how external software would have to name their groups. Instead, the stringtag is an...

> Thank goodness we retained the verbiage about restrictions on the group argument concerning order of processes in the group. What would have been the order of processes in the...