Hi, We have recently started to use silk performer 15.0 for performance testing. I had a few questions: Does silk performer do real browser emulations for doing performance testing? If yes can i rely on the page load time as the load time observed by an actual user? I am seeing differences between the loadeventend fired by the browser and load time reported for a single user test in silk performer. How can How does silk performer achieve browser parsing, performance improvements launched by browsers (like look ahead parsing), browser parallelism (loading many scripts/css/images at the same time), browser rules (block parsing while download javascript) and other nuances of different browsers. The reason why i am asking this is because when we evaluated silk performer against jmeter we realised that jmeter was sending requests sequentially which was totally off compared to the page load which we were observing. However when we ran silk performer we saw that it did achieve some degree of parallelism in execution of the page but the results of page load are still not tallying with what we see when compare the time against navigation timing apis given by the browser (navigationstart - loadeventend) time for the page load. If silk does not use real browser emulation then how do i rely on the performance numbers given out by silk performer? I am trying to understand when silk is reporting out time for pageload is it real the real page load time for the page.
↧