When running the treeUpdate benchmark in Chrome 45, benchpress reports scriptTimes and pureScriptTimes that are higher than what was reported in previous versions of chrome:

Chrome 45: ~19ms
Chrome 44: ~9ms

Looking at the browser timeline in chrome 45, the script time should be around 11ms. I am assuming that benchpress is over-reporting in chrome 45.

Note: Chrome Team changed a bunch of event names in Chrome 45, which benchpress depends on. Benchpress has been adapted to those new names (see #3411), but maybe more than just names changed. Further investigation is required.