WebPagetest Forums

Full Version: Site with "minify HTML" is slower as without
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
I tested my page with HTML minify. And is slower! Why?
Tests:
without minify: http://www.webpagetest.org/result/100420_7QCE/
with minify: http://www.webpagetest.org/result/100420_7QDQ/

page: http://www.anatomic.sk]
page with html minify: http://www.anatomic.sk/min-html.html
Try doing at least 5 runs (10 would be even better) to make sure the timings are stable. There tends to b ea fair amount of variation in results so don't assume the results from a single run are representative.
I did a lot of measurements, and it's still the same. Strange. Html minify may not always be the best choice.
And strange is a division loading JS in minify html.

Thank's for reply and the best tool for site optimizing!
I have 1 idea. if possible, it would be a good to add a LOAD TIME to TEST HISTORY, because of faster search of good and bad results.
Are you minifying on the fly or is the output pre-minified?

I can't think of a good reason why it should be slower to parse the smaller version of the file since the code doesn't care about comments and whitespace. If it's being minified on the fly I could see the extra overhead negating any benefits (and if it's a page that could be flushed early the processing could be blocking that as well).

The way the test history is done right now it can't include the load time (it's just pulled from flat files that are written when the test is initiated) but if and when I hook it into a database that should be more doable. Was trying to avoid database integration because it makes stand-alone installs a LOT easier to not have to deal with it but maybe have it as an optional enhancement.
(04-21-2010 01:17 AM)j_____ Wrote: [ -> ]I did a lot of measurements, and it's still the same. Strange. Html minify may not always be the best choice.
And strange is a division loading JS in minify html.

Thank's for reply and the best tool for site optimizing!
I have 1 idea. if possible, it would be a good to add a LOAD TIME to TEST HISTORY, because of faster search of good and bad results.

Your minified Version loads the Basepage, JS and CSS first, thus its Time to Start Render is earlier, than the non-minified Version.

Your non-minified Version loads the Basepage, CSS, , 2 PNG files, 1 JPG and JS, so Time to Start Render is later. However, as you start parallelizing earlier, the Time until Document Complete is lower.

From a Customer perception view I would choose the Minified Version, getting a 0.5 Seconds gain in Time to Start Render, which trades off against a 0.2 increase in Document Complete Time.

Though minifying itself doesn't seem to be the culprit, rather the different pattern in fetching the objects, as mentioned above. A difference in ~350 Bytes in the size of the Basepage shouldn't make much of a timing difference.

Just my 2 Cents,
Markus
Reference URL's