The Chromium Benchmarking Extension is a quick-and-dirty way to test page load time performance within Chrome.
- Can clear the cache between each page load
- Can clear existing connections between each page load
- Works with both SPDY and HTTP
- Measures time-to-first-paint, overall page load time KB read/written, and several other metrics
- Can compare performance between SPDY and HTTP
- Chrome version 7 or later
- Command-line flag "--enable-benchmarking" should be used to start chrome
Option 1: Use the benchmark from the chromium source code
The benchmark is part of the chromium source code. You'll find it here:
To run Chrome with the benchmark, use the following command line: chrome.exe --enable-benchmarking --load-extension=<path-to-chrome-tree>\src\chrome\common\extensions\docs\examples\extensions\benchmark Option 2: Install from Chrome Extension Gallery Install the extension from Chrome Extension Gallery following this link. Then you'll need to restart the browser to use it. When you run chrome, use: chrome.exe --enable-benchmarking As of M34, the method that seems to work is to load the source code as an unpacked extension after enabling developer mode and use the flag --enable-net-benchmarking in addition to --enable-benchmarking
- Iterations: How many times the test page should be loaded to collect performance date.
- Clear Results: Clear result table.
- Clear connectons: Reset the internal socket pool, oherwise chrome will reuse the socket connections if possible.
- Clear cache: Reset the cache, otherwise chrome could load the page from cache if possible.
- Enable Spdy: Use spdy to load a page. When enabled, if the web server does not support spdy, an error is reported and test will stop.
- URL to load: A comma separated list of urls to collect performance date. Note, the urls listed here should be the final url. E.g.: if url1 will be redirect to url2, url2 should be used. Otherwise, an error will be reported and test will stop. You can use "Load URLs From File" to load the comma separated url list from file if your test set is large.