Comment 7 for bug 941889

Okay I have re-run the test mentioned in comment #4. My results are:
== 5.3 ==

OLTP test statistics:
    queries performed:
        read: 1553035
        write: 0
        other: 0
        total: 1553035
    transactions: 0 (0.00 per sec.)
    deadlocks: 0 (0.00 per sec.)
    read/write requests: 1553035 (8627.93 per sec.)
    other operations: 0 (0.00 per sec.)

General statistics:
    total time: 180.0008s
    total number of events: 1553035
    total time taken by event execution: 1797.9183s
    response time:
         min: 0.10ms
         avg: 1.16ms
         max: 71.55ms
         approx. 95 percentile: 2.63ms

Threads fairness:
    events (avg/stddev): 155303.5000/542.04
    execution time (avg/stddev): 179.7918/0.00

== 5.2 ==
OLTP test statistics:
    queries performed:
        read: 1467139
        write: 0
        other: 0
        total: 1467139
    transactions: 0 (0.00 per sec.)
    deadlocks: 0 (0.00 per sec.)
    read/write requests: 1467139 (8150.74 per sec.)
    other operations: 0 (0.00 per sec.)

General statistics:
    total time: 180.0007s
    total number of events: 1467139
    total time taken by event execution: 1797.8794s
    response time:
         min: 0.09ms
         avg: 1.23ms
         max: 17.57ms
         approx. 95 percentile: 2.70ms

Threads fairness:
    events (avg/stddev): 146713.9000/1128.34
    execution time (avg/stddev): 179.7879/0.00

== Comparison ==
5.3 is a bit faster on AVG response time (1.16 vs 1.23), and on 95 %ile
(2.63 vs 2.70 ms).
5.3 managed to run a bit more queries: 1,553,035 vs 1,467,139. But this is only a 5%
difference.