⬆️ ⬇️

Javascript performance? What about page repainting performance?

Performance tests are now becoming more aggressive, while it seems that the emphasis is solely on artificial tests that can show us the performance of only some components of browsers, in particular, the performance of JavaScript functions selected for these specific tests.



It is surprising that so little attention is paid to other, more relevant aspects of the performance of modern pages, or not at all. For example, how quickly do browsers with dynamically changing page style (CSS) work?



At Velocity 2009, Lindsay Simon from Google talked about this. And on the slides from the presentation there are a couple of interesting graphs .

opera css change perfomance

It turned out that the Opera engine works very quickly with dynamically changing page content, and this is exactly what is used everywhere on the Internet. What if in the future browser performance will be measured more comprehensively, as opposed to a simple JavaScript test?

')

Another nuance that I noted on another day is insufficient and highly specialized memory tests , which do not take into account that different browsers work with memory differently. For example, they ignore Chrome's shared memory, and it does not seem at all that Opera's advanced cache would be taken into account.



In addition, Opera clears the cache of unused items every 10 minutes, but it is not visible that this would be reflected in the test report. We could clear the cache all the time, but then we would have to sacrifice performance, as Firefox apparently does.



In the context of complex testing of JavaScript and memory, Jarakan seems to increase the performance of artificial JavaScript tests, but perhaps by using more memory.

Source: https://habr.com/ru/post/63409/



All Articles