I decided to test:
Mozilla Firefox 3.6.6
Google Chrome 5.0.375.86 and
All the browsers used for test were clean installs on Ubuntu Maverick Meerkat Alpha.
Mozilla Firefox 3.6.6:
There were 5246 total tests. Firefox failed in 262 of them. Firefox used 80-95% CPU the whole time with sometimes dropping to as low as 20% for few seconds. Memory usage was 190mb to 220mb.
It even gave a prompt to stop the script because it was memory intensive. This was a good thing. Here’s a screenshot:
Once the tests were over, memory usage dropped to 48mb indicating that the memory was released.
Google Chrome 5.0.375.86
Google Chrome 5.0.375.86 failed in 222 of them (Yes, 3 times would’ve made it devil’s number :P). Chrome used 70-85% CPU and used around 220-240mb memory. The memory was freed once the test was over. However, failing in 222 on a test made by its own company is kinda shameful. This also however proves that Google is impartial.
Opera 10.60 had quite astonishing results. It failed on only 78 tests. That was a big margin win from other two. Opera is really doing some work on its engine.
Just like the other two, it used 80-90% of CPU during testing and consumed around 200-220mb memory. However, it didn’t seem to release the memory once the test was complete. Even after closing that tab, memory used still showed 205mb. Now that could really become one big negative point. Maybe its just on linux (windows user please comment). But they must really sort out this issue.
With such a huge margin, Opera is clearly the winner. I am starting to think ‘Was this even needed?’
Its become customary now – cheers!