<br><font size=2><tt>"Talpey, Thomas" <Thomas.Talpey@netapp.com>
wrote on 05/10/2006 03:53:04 AM:<br>
<br>
> At 11:13 PM 5/9/2006, Shirley Ma wrote:<br>
> >Have you tried to send payload smaller than 2044? Any difference?<br>
> <br>
> <br>
> You mean MTU or ULP payload? The default NFS reads and writes are<br>
> 32KB, and in the addressing mode used in these tests they were<br>
> broken into 8 page-sized RDMA ops. So, there were 9 ops from the<br>
> server, per NFS read. I used the default MTU so these were probably<br>
> 19 messages on the wire. I don't expect much difference with smaller<br>
> MTU, but smaller NFS ops would be noticeable.<br>
> <br>
> Tom.<br>
> </tt></font>
<br>
<br><font size=2 face="sans-serif">I meant payload less than or equal to
2044, not IB MTU. IPoIB can only </font>
<br><font size=2 face="sans-serif">send <=2044 payload per ib_send_post().
NFS/RDMA in this case send </font>
<br><font size=2 face="sans-serif">32KB per ib_post_send(). It would be
nice to know the performance </font>
<br><font size=2 face="sans-serif">difference under same payload for IPoIB
over UD and NFS/RDMA. Is that</font>
<br><font size=2 face="sans-serif">possible? <br>
</font>
<br><font size=2 face="sans-serif">Thanks<br>
Shirley Ma<br>
IBM Linux Technology Center<br>
15300 SW Koll Parkway<br>
Beaverton, OR 97006-6063<br>
Phone(Fax): (503) 578-7638</font><font size=2><tt><br>
</tt></font>