The HACC WG will convene at the usual time today. Today's agenda:

(1) Continue reviewing feedback and revising allocator kind info proposal

  + Continue with existing asserted/provided approach

  + Or, create new requested/provided approach

  + Please give this some thought and come with an opinion. I feel we have not heard from enough members of the WG to commit to a specific direction.

(2) Discuss MPI_THREAD_MULTIPLE clarification

  + Concern is that doing nothing leaves it ambiguous

  + If we add an advice to users that it's ambiguous, then need to assume the weaker semantic

  + Should we just specify the weaker semantic and add a new threading level in the future?

(3) Goal setting for accelerator synchronous communication

Cheers,

 ~Jim.