[ofiwg] definition of a completion in OFI
sean.hefty at intel.com
Thu Mar 5 10:50:47 PST 2015
> In IB terms:
> For reliable services, a WQE is not considered complete until all
> necessary responses have arrived. (compliance statement C9-60).
> For unreliable services, the requester shall consider a Message Send or
> RDMA WRITE complete when the last byte has been committed to the wire (or
> an unrecoverable error occurs). (Compliance statement C9-180).
> For bonus credit, can you point to the IB spec location where this is
> [PG] See above. How shall I collect my bonus credit?
Excellent - thanks! I'll check the iwarp specs again for similar wording.
I'll save you an extra dessert in Monterrey.
> [PG] I don't think that's a reasonable definition of completion, but the
> point is arguable. The definitions that IB uses for the requester side
> are the ones given above. The responder side may be a different story. I
> think we should adopt the same definitions as used by IB.
IMO, this ends up being too strict, which may limit optimizations. I can see a use for supporting the existing definition, that simply means that the local buffers can be reused, in order to allow for queueing locally or in the fabric. A stricter completion model meets this requirement.
More information about the ofiwg