Cached and non-cached runs question

[avellayappan]

Peter,

Have a question on the report generated by sitespeed.io? Because it is selenium based, I think - the numbers reported are non-cache based. However, just wanted to see how we calculate the first and repeat runs? In the report that are two parts:

1. Weights and Number of files - reported by sitespeed.io
2. Timing metrics reported by browsertime How can we measure the above ones for cached and non-cached runs using sitespeed.io?

Thanks!

[tobli]

Hi! For Browsertime, the numbers are indeed collected using Selenium. As best possible, we try to instruct each browser to disregard caching (by adding browser specific start flags, and restarting the browser between each run). I'm practice it doesn't work perfectly, and some caching is hard to avoid (such as OS level caching of DNS queries). Collecting data with caching, separate from non cached, is certainly possible. It would be great if you could open a feature request for that, so we can detail what kind of setup would be useful to you.

[soulgalore]

Hi @avellayappan,

1. Weights and Number of files - reported by sitespeed.io The reported are non cached first run. Would you need to have it in first/repeated? The information is there, but we don't show it today. However, one important thing about the size: The reported size are not the compressed size for text based assets, that's an issue inherited from YSlow/PhantomJS. Will look into that later.

Best Peter