I believe the test Sharath Sulochana mentions shows a final number of the avg CPU utilization regardless of processes. (cpu_utilization_rate)
Looking deeper into the test results, I see that goxdcr CPU utilization rate stays relatively low and same across tests.
(Old test)

(New test)

ns_server CPU utilization also stayed the relatively same between tests.
However, memcached CPU increased between the tests:
(Old test):

(New test):

This could explain why the overall CPU went up.
I also noted that compaction progress looks a bit different between the tests as well, but cannot directly conclude that it has affect on overall CPU usage. But that does show how KV behaviors between the two tests are not identical between them.
Sharath SulochanaIs there any regression on DCP?