Home AMX User Forum AMXForums Archive Threads AMX Hardware

NetLinx NX Integrated Controller discussion

Chat about testing the product to its limit


  • CPU load testing

    Well i have had great success in a small cpu load test.
    I had a project that had 14 rooms on the one controller 4 of them were vc rooms.
    The normal ni controller could not handle this and it would just freak out and crash (unable to connect to the device ).
    luckily i had a DGX frame to split the load up .

    I did a test with the same code on a nx1200 and it ran with only a 56% load , i then added more rooms to see what it could do and ended up with 21 rooms on the one controller and to my surprised it ran fine with a 76% load .

    This test was not with the rooms connected only one touch panel to see if the gui was smooth .
    That fact that the code loaded up on a nx and not on a ni proves that its got allot of potential.

    Thought i might share this small test to everyone in hope that people will start using because of there processing power.


    Code BIG /Test BIG / Produce BIG
  • John NagyJohn Nagy Posts: 1,728
    While I agree with your conclusions, I have to wonder about your code that crashes a standard NI with 14 rooms and 1 panel.

    Good code should not do that. Curing bad code with more power just kicks the problem down the road, and hardly provides a good measuring stick.

    By comparison, we can support 50 rooms with 50 touch panels on a single standard (not even memory enhanced) NI3100, with nearly no user-perceivable slowdowns. That can include intercoms, thermostats, 50 sources, 50 TVs, And no crashes. As others here can also attest, with good code, it can be done. Most systems including ours will benefit from the new speed, of course.

    You really didn't need to start a new topic as this is being discussed in pretty good detail in another topic for weeks now.
  • understand , but this was just a test and not a working solution .

    also couldn't find the threat before
Sign In or Register to comment.