[c-nsp] Netflow CPU usage

[mack]

We have a cisco 6513 with Sup720-3BXLs and DFCs on the line cards.
When attempting to turn on mls nde sender the CPU usage goes through the roof (70+%).
We have sufficient traffic to saturate the netflow tcam no matter what our settings are.
It works with ip flow export only but activating mls nde sender pushes CPU up.
Obviously we are only seeing a partial picture with this configuration.

Current cache size is 235520 (recommended maximum).

Current settings are:

mls aging fast time 8 threshold 1
mls aging long 300
mls aging normal 60
mls flow ip interface-full
no mls flow ipv6
no mls nde interface

#show mls nde
Netflow Data Export is Disabled

 Netflow Aggregation Enabled
   destination-prefix aggregation exported 2688160 packets, 88178724 records

My current thoughts are moving a portion of the traffic to a different chassis to reduce the load and flow counts.
This will only reduce the problem not solve it.  More likely it will just move the problem.
Any recommendations to reduce the CPU load would be appreciated.

Mack