Seems like we essentially agree that telemetry is the way to decide what to measure, and microbenchmarks is a handy way to take measurements?
I've updated the wiki with some results and two proposals for summarizing them. I'm not happy with any of the summaries - input is appreciated. I also see (also noted by Geoff) that some more statistics should be gathered for the results in order to have an idea of how stable they are. I'll try to gather necessary data to calculate std-deviation. Input is also appreciated here. However, from the current results on wiki it seems pretty clear that that clearing the cache on a Galaxy S is a bad idea. It's also slow on the Nexus, although not that bad. _______________________________________________ dev-tech-network mailing list [email protected] https://lists.mozilla.org/listinfo/dev-tech-network
