Hello Curtis,
according to
http://www.cisco.com/web/partners/downloads/765/tools/quickreference/routerperformance.pdf
a c3660 is capable of a max of 51-61 Mbps of traffic.
Other C36xy are less capable.
Traffic is forwarded using CEF but still the device uses its main cpu to move packets.
I think what you see is compatible with expected performance of the platform.
this should be the meaning of no cpu gain removing WCCP.
Hope to help
Giuseppe