Edgar Gabriel

@StaticObserver thank you for the clarification. At this point here are the options that we have - UCX master has a new flag that allows you to disable the caching...

> In our case the buffers also changes constantly. What's worse is that in ROCM case they don't have the switch to turn off the IPC caching... you can compile...

we are still investigating the best path forward. The truth is that this concept seemed to work for many years without issues (and still works for many application), it looks...

@andymwood thank you for the report, I will have a look over the weekend

@wenduwan was that fix back ported to the 5.0 branch as well?

@ggouaillardet Thank you, I think you are on the right track. ```mca_common_ompio_set_aggregator_props()``` is only invoked once per ```MPI_File_write_all()``` per process (or at least it should be). However, I think you...

@ggouaillardet you are technically correct, and I am fundamentally ok with the ```realloc()``` solution that you suggest. Note, that e.g. a similar parameter on the file handle (```fh->procs_in_group```) is freed...

@ggouaillardet would you like to file pr with the fix, or shall I go ahead and do that?

@devreal since we talked about this issue offline, I will assign it to you. Please don't hesitate to ping me if you need any help!

@devreal ok, thank you! I think we should probably file a ticket with the HDF5 folks in that case, will ping you in a bit about that.