Scalibity issue with 500 users on 2 node RAC

We are doing performance testing on 2 node , 11gR2 cluster on AIX 7.1. This is for just 500 users. These the AWR reports for intance1 and 2. Could you please shed some light :-

a) why there is such a huge difference in top 5 timed events?

b) what is this "Global cache log flushes for cr blocks served" ?

Instance-1

Top 5 Timed Foreground Events
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
                                                           Avg
                                                          wait   % DB
Event                                 Waits     Time(s)   (ms)   time Wait Class
------------------------------ ------------ ----------- ------ ------ ----------
DB CPU                                           18,513          12.3
latch: row cache objects             49,674       5,378    108    3.6 Concurrenc
latch: cache buffers chains         125,336       4,604     37    3.1 Concurrenc
cursor: pin S                       127,775       2,261     18    1.5 Concurrenc
gc cr block busy                    277,289       2,204      8    1.5 Cluster

Global Cache and Enqueue Services - Workload Characteristics~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~