Hui Zhou
Hui Zhou
@roblatham00 Could you take care of this (old) ticket?
I determined that the reported issue no longer exist. The check of `default_striping_factor` has been removed already. This commit -- d2cb66671016753efd22f3293ed8984c97b91441 -- fixes it.
I have no clue what this is about. Searching `MPIU_INSTR_` digs up this page: https://github.com/pmodels/mpich/blob/main/doc/wiki/design/Internal_Instrumentation.md
> > I have no clue what this is about. Searching `MPIU_INSTR_` digs up this page: https://github.com/pmodels/mpich/blob/main/doc/wiki/design/Internal_Instrumentation.md > > It looks like the old instrumentation code was removed in [606f9b6](https://github.com/pmodels/mpich/commit/606f9b687230dcd3537e538e6983f8e7b573cd76)....
Confirmed. Lock request should return failure (via active message ACK) if there it is received inside a pending fence.
There currently is no plan to add Hadoop yarn support and it is not clear how needed is this feature. I am closing this issue now. Reopen with more details...
test:mpich/ch4/ofi
Currently, that will need to overwrite `MPIR_CVAR_ENABLE_GPU`, and it can get messy (modifying cvar at runtime in general). We probably will need a separate global variable, to replace the CVAR...
I think overwriting `MPIR_CVAR_ENABLE_GPU` should be okay since we are doing it during init before any other GPU functions.
For my understanding, is `ADIO` one of the backend and different back end will have different representation for `MPI_File`?