MD3220i MPIO Performance Issues

Hi all,

Looking for some clarification on whether the performance we are getting through MPIO is expected or not quite right. 

We have a 3 node Hyper-V/Storage Cluster using Dell R710's attached to an MD3220i array.  Each server has 2 iSCSI NIC's and various others for cluster, management and vm traffic.   Each iSCSI NIC is on its own subnet and has one connection to each of the MD3220i controllers.  The iSCSI traffic goes through 2 identical switches for failover and is contained to its own VLAN, jumbo frames are enabled accross all NIC's/VLANs/Switches.

The problem:

Running storage benchmarks we only see a set of iSCSI NIC's running at arround 40-60% utilization, never higher.  each are load balancing properly but seem uncapable of reaching near 100% each.  All the benchmarks suggest that we are only seeing 130-140MB/s max throughput.  The latency and IOP/s are as expected.

Is this how a dual MPIO team should behave?  Surely we should be able max out both those cards to the storage array?  The array has 24x 146 15k SAS drives in a raid 10 so should be more than capable of churning out the data.

Any sugestions on where we should look/what might be causing this problem?