|
From: Jon S. B. <js...@ha...> - 2001-06-12 04:46:29
|
VS Renganathan wrote: > > Hi David, > > From: "David Megginson" <da...@me...> > > Eventually, it might be a good idea to externalize the FDM as well, so > > that the integration issues with JSBSim (and any others) become > > simpler, but it still sounds like a great start. > > Ok, Once I get started I'll check with John Berndt. Huh? I must've missed something. Note that this is the first message I am posting from the new machine. Boy, am *I* a happy camper!!! Between setting up the machine and burn-in, and "The Flood" I have been a little out of it. Is this related to threading and distributed processing? Sounds like it... > > Next, you need to decide how the processes are going to communicate. > > Personally, I prefer sockets so that you're not limited to running on > > a single SMP machine (you could run the FDM somewhere else on your > > subnet, for example); a lot of other posters seem to prefer shared > > memory. > > When I first learnt socket programming I remember having read that if two > processes communicate using sockets on the same host the results could be > *unpredictable*. What do you say about this. Since we have to be able to run > FlightGear on a single processor machine as well, would it not be better to > distinguish the two as suggested by Alex Perry. Too bad reflective memory is so expensive. That's one of the methods we are using to communicate between a VME machine and our SGI. Pretty cool stuff. Jon |