Hello! I’m making a game and I’m testing Google App Engine for how fast it responses. It’s nice, but not perfect. I made a benchmark:
http://2.latest.lagstest.appspot.com/
source code is here:
http://2.latest.lagstest.appspot.com/lagtest.zip
The code is simple, it fires 1 request every 200 ms to GAE server and takes current timestamp. When receiving response, it takes another timestamp and computes how much time has passed, showing it by bar height.
Now, I would expect false benchmark results if I would compute delta time in ENTER_FRAME event, but why if I do it in a callback function results are still dependent on CPU usage? Is that just flash or am I doing something very wrong ?
Can You tell me what the lowest average lag can I achieve? I get ~50 ms now when connecting with localhost! Is GAE’s ~200 ms a good or bad result? Have anyone of You made similar benchmarks?
In the app linked above I used URLLoader. I also tried pyAMF before:
http://0.latest.lagstest.appspot.com/lagtest.zip (old pyAMF v.3)
http://1.latest.lagstest.appspot.com/lagtest.zip (pyAMF v.6)
These are based on Aral Balkan’s GAE SWF PROJECT.
Should I use pyAMF or URLLoader if I see no difference?
I asked a question on GAE discussion group if You’re interested to read something more:
http://groups.google.com/group/google-appengine/browse_thread/thread/ed5c2e60529890ff
If You can’t answer me, where else should I ask?
