While running our unit tests, I came across an unexpected result. I
assumed that the Standalone Player would be more efficient than any of
the browser implementations, however I was shocked when the Standalone
Player tests easily ran 2-3X faster.
Inexact test results running our tests several times using Flash 10.x
players on Mac OS X 10.6.x:
Opera: ~55s
Safari: ~45s
Firefox: ~44s
Flash Player Standalone: ~20s
I am trying to figure out whether such a great disparity is expected.
Does anyone have any ideas why this might be happening? If you have
traditionally run your unit tests in the browser, this might save you
some time.
Some thoughts:
1. This is an OS-specific issue
2. We have a memory leak or a memory signature that is handled better
by the Standalone Player. Perhaps less garbage collection.
3. The Standalone has some performance optimizations that haven't
made it into the browser implementations
4. The Browser Flash Players are inferior implementations. There
appears to be some finger pointing already:
http://tech.groups.yahoo.com/group/flexcoders/message/148899
<http://tech.groups.yahoo.com/group/flexcoders/message/148899>
