Hi Dominik Scanned through your slides and thought they looked good. Would definitely be nice to create an MPI-level abstraction for dynamic ops! Only comment I had was to perhaps provide the equivalent of PMIx attributes to the MPI API as a
ZjQcmQRYFpfptBannerStart
This Message Is From an External Sender
This message came from outside your organization.
ZjQcmQRYFpfptBannerEnd
Hi Dominik
Scanned through your slides and thought they looked good. Would definitely be nice to create an MPI-level abstraction for dynamic ops!
Only comment I had was to perhaps provide the equivalent of PMIx attributes to the MPI API as a means of making it easier to extend in the future. I suspect the community at-large will discover additional use-cases that require new inputs, and it would be unfortunate if we had to create new APIs every time that happened.
BTW: I have the resource usage code working in both PMIx and PRRTE now - getting ready to commit those changes. Just need to clean up a couple of things first.
Ralph
On Aug 18, 2025, at 12:12 PM, Dominik Huber via mpiwg-sessions <mpiwg-sessions@lists.mpi-forum.org> wrote:
This Message Is From an External Sender
This message came from outside your organization.
Hi all,
As discussed in our last meeting, we’re considering a presentation on the current state of our Dynamic Processes discussions at a tentative non-voting meeting in early September.
Since many of us will be on vacation or at EuroPar in August, it may be difficult to hold WG meetings during this period.
I won’t be able to attend next week’s WG meeting (if one is planned), but the usual Zoom link should still work.
Best regards,
Dominik