On Saturday 18 April 2015 01:56:37 Karlsson & Wang wrote: > I am running linxucnc via ethernet and have studied the CPU and > network load. The minimum servo period seems to be around 200µs > despite both CPU load and network load is low. I have not yet > investigated if this is because network packets supposed to be > received are missing or if process is not executed in time. It however > seems to make small difference if CPU is heavily loaded with other > tasks for example capture packets on the ethernet interface which is > good. > > Is there any possibilities in Linux RT to investigate which processes > are scheduled to be run in real time? Is there to investigate what > might be locking a scheduled process from execution? > > Nicklas Karlsson
This is somewhat of a strange question, and it prompts me to ask why you feel the need to do this. I did do some experimentation, down to 250 u-s for the servo thread, nearly a year ago, trying to get a stable behavior out of my spindle servo. However I did find that it helped some, but was not a universal panacea. The eyeopener to the real problem I was seeing was doing a rockhopper chart on the setup, after I had given up on software stepping and put a mesa 5i25 card in the machine, killing the need for a BASE_THREAD totally, and discovering that my addf statements in the .hal file were a quite a bit out of order versus the signal flows, such that it took several servo periods for the error to propagate thru the system from the encoder inputs to the final pwm output that drove the servo amp. With the correct sequence of addf's (they execute in the order in the .hal file) all the feedback lags disappeared. So my SERVO_THREAD is now set back at the default 1 millisecond and everything Just Works(TM). This does not answer your question directly, and I trust one of the developers will chime in, perhaps with the info you seek. LinuxCNC does have the ability to set DEBUG_LEVEL in the .ini file, so it may be possible to extract what you want from the info output that can give. I believe that while the debug level is stated in a 32 bit hexidecimal format, that the various outputs are triggered by the binary value of the individual bits of the 32 bit value given. However, the last time I looked, I wasn't sure which bits I needed to set in order to track what I wanted to track, so after about half a days playing, I set it back to $00000000 and have not played with it since. It is possible that debugging is better explained today at: <http://wiki.linuxcnc.org> than what I was reading most of a decade ago, so that would be my first stop. Sorry I can't be more helpful. Cheers, Gene Heskett -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author) Genes Web page <http://geneslinuxbox.net:6309/gene> ------------------------------------------------------------------------------ BPM Camp - Free Virtual Workshop May 6th at 10am PDT/1PM EDT Develop your own process in accordance with the BPMN 2 standard Learn Process modeling best practices with Bonita BPM through live exercises http://www.bonitasoft.com/be-part-of-it/events/bpm-camp-virtual- event?utm_ source=Sourceforge_BPM_Camp_5_6_15&utm_medium=email&utm_campaign=VA_SF _______________________________________________ Emc-developers mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-developers
