Hi Alon, I am on the master branch. I probably don't understand the starting mechanisms. On loading, I do not want the program (main) to be called, so I set shouldRunNow and noInitialRun to false, without truly understanding what it means. Because I have no setStatus, doRun gets called but not main, so for the intended run later I have to set calledRun to false because main was actually never called.
The performance bug is probably fixing itself in the future.I should have given you a more detailed description earlier, sorry for that. Checking the console log in FF for the second run I found the following line that probably explains everything: -- [15:54:06.745] asm.js link error: As a temporary limitation, modules cannot be linked more than once. This limitation should be removed in a future release. To work around this, compile a second module (e.g., using the Function constructor). @ http://gaia.respawned.com/gaia.js:98973 Not necessary anymore, but here are the steps to reproduce this. The fluid simulation is using the time-steps as main loop. - goto http://gaia.respawned.com wait until you see the gnuplot graph. - switch to the "Run" tab and hit Start - wait for two time steps, performance output always comes after TS 0. - now hit "Pause", wait a few seconds (time step needs to finish) - click "Reset", this will call the encapsulated compiled version again in the worker. - click "Run" again and wait again for the first two time steps to finish. (it will take a lot longer now) Cheers, C > > -- You received this message because you are subscribed to the Google Groups "emscripten-discuss" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
