Benchmark the effect of enabling lookups in QML

As we implement inline caching for binding expressions, signal handlers and generally any JS code run with a QML context, we must evaluate what the impact of these changes is in terms of

(1) run-time performance (qmlbench)
(2) start-up performance
(3) memory consumption

Early data and experience suggests that this is the right direction in any case, but the numbers should help us identify areas of compromise and avoid accidental regressions.

The acceptance criteria for this task is to collect information here about these three areas before/after landing the series.