While updating the cpp driver from 0edf3d2eff88d7594ed7dc070f889b1bcacb677f to master I found out a very big regressing into our internal benchmarks.
You don't have to add any data inside. The difference is noticeable even with empty results.
I did a small c++ app to show this regression (please find attach):
0edf3d2eff88d7594ed7dc070f889b1bcacb677f - test completes in 1082 ms
master - in over 3800 ms (never seen less than 3.4 seconds)
I performed the tests on a MacBook Pro (Retina, 15-inch, Mid 2014) 2.5 GHz Intel Core i7, 16 GB 1600 MHz DDR3
Basically the test is measuring how fast you can send 100000 prepared queries and receive their responses.
Mac OS X, Linux