"Talpey, Thomas" <[EMAIL PROTECTED]> wrote on 05/09/2006 06:58:09 PM:
> Of course, the NFS/RDMA copies do not need to compute the checksum,
> so they are more efficient than the socket atop IPoIB. But I am not
> sure that the payload per WQE is important. We are nowhere near the
> op rate of the adapter. I think the more important factor is the interrupt
> rate. NFS/RDMA allows the client to take a single interrupt (the server
> reply) after all RDMA has occurred. Also, the client uses unsignalled
> completion on as many sends as possible. I believe I measured 0.7
> interrupts per NFS op in my tests.
>
> Well, I have been very pleased with the results so far! We'll have more
> detail as we go.
>
> Tom.
Computing the checksum is expensive on receiver, on sender it's free.
Yes, high interrupt rate is another killer of the performance. But I think
payload is important, with large MTU supports in IPoIB RC mode, the
performance would be much better.
Have you tried to send payload smaller than 2044? Any difference?
Thanks
Shirley Ma
IBM Linux Technology Center
15300 SW Koll Parkway
Beaverton, OR 97006-6063
Phone(Fax): (503) 578-7638
_______________________________________________ openib-general mailing list [email protected] http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
