> пт, 13 груд. 2019 о 18:38 Dewey Garrett пише: > > > Current Docs: > > > > > > http://buildbot.linuxcnc.org/doc/scratch/v2.9.0~pre0~dgarr-dryrun~6eca6ac/html/integrator/dryrun.html > > > > > > http://buildbot.linuxcnc.org/doc/scratch/v2.9.0~pre0~dgarr-dryrun~6eca6ac/html/man/man9/dryrun_helper.9.html > > > > Branch ref: https://github.com/LinuxCNC/linuxcnc/commits/dgarr/dryrun > > > > This branch is force-pushed when there are changes. > > > > Thanks for the great job! > > Is there a way to speed it up to estimate the program run time? I.e. > simulation mode.
Have done similar for other programs. It was implemented as function called periodically and for simulation simply run in loop as fast as computer calculated. Assuming parameters for motion planner is correct so that there is no following error there should be no need to run feedback loop, trajectory planner should be enough. I came to the conclusion program could either stop waiting for something like a periodic timer or be called then something happen for example triggered from a periodic timer. Stop waiting need support from operating system so that other threads could continue executing! Called then something happen and program must be implemented as a state machine! It is often more work to implement as a state machine though control loops usually have very few states and i have an advantage it also could be called from interrupt which is particularly useful then timers with nested interrupts with priority could be used to call function. Regards Nicklas Karlsson _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
