Code Processing Sequence During Boot of NX / NI

tholt0831

I am having a problem with code that ran on an NI-3100 running on an NX-3200. I believe it is related to the processing sequence of code during a boot. I thought I understood the order, but maybe I don't. Can some one provide a complete list of when each section is executed?

DEFINE_START in main-program and modules

DEFINE_EVENTS - ONLINE: in main-program and modules
(How do virtual-devices fit in here, both for the local to the NI/NX and for a remote NI/NX)

cmatkin

Hi tholt0831,

There shouldn't be any differences.

We upgrade NI's to NX's fairly often and the code only requires a few changes with respect to devices depending on model upgrading.

These do load faster.

Maybe post some code for us to check?

Auser

Check whether the DEVICE HOLDOFF setting on the 3200 is the same as on the 3100(s) you've been running it on too.

vincen

Be careful that traditional DEFINE_START and DEFINE_PROGRAM*are deprecated and should no more be used with NX series In all matters if you program fully in Netlinx these sections should be already empty or removed !

J Donachiue

Why would you not want to use DEFINE_START?

I use it to initialize Variables on Boot.

Do we have a better place to get a Boot Status?

ericmedley

J Donachiue wrote: »

Why would you not want to use DEFINE_START?

I use it to initialize Variables on Boot.

Do we have a better place to get a Boot Status?

It has to do with under the hood stuff and the way Java is working.

You can make your own Define_Start like this:

DEFINE_DEVICE

dv_My_Master = 0:1:0

DEFINE_EVENT

data_event[dv_My_Master]{
  online:{
    // put define_start stuff here.  Now it's an event.
    } // online
  }//D-e

DHawthorne

vincen wrote: »

Be careful that traditional DEFINE_START and DEFINE_PROGRAM*are deprecated and should no more be used with NX series In all matters if you program fully in Netlinx these sections should be already empty or removed !

That's not exactly true. Technically, every DEFINE_FUNCTION and DEFINE_MODULE are part of the DEFINE_START section. If you leave it out, you are going to get a compiler error. DEFINE_PROGRAM, however, you can do without.

NZRob

Also CREATE BUFFER can only be in the DEFINE_START section otherwise there is a compile error too ( just tried it)

ericmedley

NZRob wrote: »

Also CREATE BUFFER can only be in the DEFINE_START section otherwise there is a compile error too ( just tried it)

Yeah, this knda points to an old AMX credo of never do somethng that breaks old code. We've been doing DEFINE_START for ever. I'm not sure what exactly is the issue with the discussion on define_start with the new NX series and how it interfaces with JAVA under the hood. I do know that using define program on them results in a significant performance hit. Perhaps it's the same issue.

Define_start has always struck me as odd in regard to what all goes there, like for example, CREATE_BUFFER and why you can't create code with any stack_vars but put the same code in a function or call and call it in DEFNE start and it compiles and works fine. I know it all has to do with the order of the runtime build. But it manifests itself in puzzling ways sometimes.