Zelimir Ikovic ha scritto: > I wanted to know how predictable and reliable GAMBAS is when a reaction to > real-world events is required. > > Let say that Java RealTimeSystem is intended only for suitable operating > systems, which means that only a > real-time operating system, such as QNX, is appropriate for implementing the > JVM. > So that means it depends on underlynig OS version ... > > Simply installing an RTS extension > and renaming java.lang.Thread instances to > javax.realtime.RealtimeThread will not turn the > application into a real-time app. > > I just wanted to know what is Gamas reaction to real world compared under > different underlyng OS. > FreeBSD, OpenBSD and Linux are not he same. > I understand but, let me say, your sample code had nothing to do with realtime... it was simply "fast as much as possible", without any mechanism to ensure some kind of real-time response to some event. To only read a counter is not enough. I can run 4 stepper motors directly (no step+direction, but driving the phases directly instead) with a CPU more than 100 times slower than a x86 CPU. But this is possible using interrupts. Practically every OS can, in some way, use interrupts - gambas can not. So, the only way in gambas to do precise timing is to let it run alone, with maximum priority and privileges. Or, to delegate high timing precision to some external piece of software (a libray or a driver) which can take advantage of interrupts, or callbacks, or signals.
On the other hand I think that gambas can, with tight loops, do delays like you did in the original source. Surely it is predictable enough to let you calculate in some way the correct amount of cycling to obtain the needed delay - the fact is that gambas will never be the only thing running on the system, and you have little control about that. The same problem of windows xp - if you want precise timing, you must obtain privileges, either for the program or using a driver. I must also add that, without particular precautions, linux is more predictable about sleep() than win2000 or xp. Regards, -- Doriano Blengino "Listen twice before you speak. This is why we have two ears, but only one mouth." ------------------------------------------------------------------------------ Download Intel® Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev _______________________________________________ Gambas-user mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/gambas-user
