On Donnerstag 26 Mai 2005 13:17, Gerard ROBIN wrote:
>   I was in fact the first to ask for: Why that message and that Freeze ?
This message is because of the fact that this is the case.
Even if you can see ground there is a posibility that in floating point 
arithmetics the test for that can fail.
Also there could be a serious problem in this area.
And why that freeze: It is somehow paranoid but it shows that there is 
something wrong. This could either be a roundoff problem, which should be 
handled in some way or it could be some other problem which should be handled 
in a different way.
As long as it is not clear to me what is going on here. I am searching for 
that reason ...

>   I wonder, and i ask is it in relationship with an other problem which
> makes the flight uncontrolled and the CPU out of capacity ? ( that is 
> CPU speed independent, low or high speed machine same problem ==>I have
> several working machines )
>   I mean :
>   Probably during tile loading and or dynamics sceneries, the system try
> to answer to:
>       that loading,
>       the control of the aircraft
>       the dynamic display of everything.
>   Because we are supposed to be in real time, one function decision,
> goes probably against an other function decision, correct it and so
> on ..... the system becomes unstable going into a devil loop.
In short: you are talking about race conditions.
I don't think that this is a race condition since scenery loading is done in a 
seperate thread, but hanging that scenery into the scenegraph is done in the 
render thread which also does the FDM computations.

>  I get less than one fps display.
So if you can recover, you are talking about a completly different problem.

>   To me , under linux i can observe the cpu loading.It becomes crazy
> with hight variations, without any hope of stability again.
Sorry, under Linux I have always 100% cpu usage when running flightgear ...

>   I save the situation with the activation of Pause during a very short
> delay, and again the system become stable coming up in an 'heaven'
> situation.
See above.

>   Could we conclude with that question:
>   Does management of sequence and priority of modules activations in a
> complex situation could be the cause of theses problems ? (Strange
> accelerations with...  FDM freeze......
I don't understand what you mean.



Mathias Fröhlich, email: [EMAIL PROTECTED]

Flightgear-devel mailing list

Reply via email to