Hi Roland,
We have software which uses the mellanox shared receive queue and we
are seeing a substantial performance penalty (on the order of 4-5
usec on 0 byte messages). I don't know all the technical details but
if you could keep this in mind when working on the shared receive
queue perhaps the performance penalty could be much less in gen2.
Thanks,
Galen
On Jul 25, 2005, at 4:05 PM, Roland Dreier wrote:
Hal> A proposed API was written up a while ago (mid June). Not
Hal> sure how far along the implementation is.
I've not made much progress since then but I'll start working
seriously soon.
- R.
_______________________________________________
openib-general mailing list
[email protected]
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
