[mpiwg-rma] [EXTERNAL] ambiguity in MPI_WIN_FLUSH_LOCAL and completion of bidirectional operations

Jeff Hammond jeff.science at gmail.com
Mon Oct 21 12:57:12 CDT 2013

On Mon, Oct 21, 2013 at 12:48 PM, Underwood, Keith D
<keith.d.underwood at intel.com> wrote:
>> I'm not sure I agree with your hardware crystal ball.  NIC-to-NIC reliability
>> would still permit separation of local and remote completion since the former
>> occurs as soon as the data reaches the NIC.  While on-chip, coherent
>> networks may become common at the high-end of HPC, do you believe that
>> distinct-NIC-based systems will completely disappear soon?
> I wouldn't think that higher bandwidth NICs would be able to afford to buffer all of the data on the NIC for the general case...

Sure, MPI_GET_ACCUMULATE is presumably not going to show any benefit
to this for larger messages, but I am interested in the
latency-sensitive regime.


Jeff Hammond
jeff.science at gmail.com

More information about the mpiwg-rma mailing list