Home AMX User Forum AMX General Discussion
Options

CPU usage on NI3100

I have been trying to clean up this quite decent size project for a house: 2 x NI3100, 4 x NI900, 4 x MVP8400i, 9 x NXD430, 1 x iPad, 4 x R4, Alarm, Pool, Lights, Thermostats, iWeather, IP phone (AMX), KScape, Request, XM Antex, Autonomics Apple TV, 15 zones, TVs, Sat and BluRays...
I had first to complete the job and then ended up by re-writing a good 90% of the project.
Solved tons of errors and problems along the way.

Now I'm facing this issue where the main master NI3100 is freezing after a few days (sometimes 3 days sometimes a bit more): can't connect to it with Netlinx Studio, have to power it off and on.
It has been working good for a couple of month until the owners experienced a major strom and loss of power for 12 hours.
I started by applying the most recent Master firmware 3.60.453 and dDevice firmware 1.30.8
Among other checks, I ran 'cpu usage' and got 99.68%

So my questions are:
1. What exactly does the cpu usage[b/] command return and how do you interpret it?
2. What else to look for if that 99.68% looks a bit high?

Thanks for the help.

Comments

  • Options
    John NagyJohn Nagy Posts: 1,734
    There are several older threads with much detailed discussion of CPU usage, a quick search will find it and not require all the repeats here.
  • Options
    jgreer9514jgreer9514 Posts: 23
    Route Mode Direct, Master Circular Reference?

    My first thoughts on a larger AMX system and the cpu usage maxing out is the master to master communication is maxing the system out. I'm not sure how your system is programmed but I always use "route mode direct" on the NI's. Even in my large systems I would expect the cpu usage to be around 10%.

    I use event based programming for feedback.
    No code in define_program.
    Make sure your url list is not containing any circular references to other masters.
    Probably worth rebooting all masters and making sure another master is not hanging up another master.

    John
Sign In or Register to comment.