NI Scan Engine: Windows EXE vs Real-Time

[bkb2005]

This is something I can't find information for and is driving me crazy. I have a project where I'm accessing a cRIO with the NI Scan Engine. If I compile this to a Windows EXE file, what exactly is happening under the hood? Is this Windows EXE deploying real-time code to the cRIO and acting as a remote panel? Or is the actual logic running locally?

[jgcode]

This is something I can't find information for and is driving me crazy. I have a project where I'm accessing a cRIO with the NI Scan Engine. If I compile this to a Windows EXE file, what exactly is happening under the hood? Is this Windows EXE deploying real-time code to the cRIO and acting as a remote panel? Or is the actual logic running locally?

I am pretty sure its along the following lines:

Once the Scan Engine software it deployed to the RT brick, the scan engine will be contently updating the memory map even if you aren't explicitly calling the variables.

That is why they say do not install the SE software unless you use it as you will get a performance hit for no reason.

If network publishing is enabled, these variables will be like any other Network Accessed Shared Variable - you can get a hook into them via the SV Engine and you have use them in your Windows PC application.

Here are some resources you may like to check out:

http://zone.ni.com/devzone/cda/tut/p/id/7338

http://zone.ni.com/devzone/cda/tut/p/id/7693

http://zone.ni.com/devzone/cda/tut/p/id/4679

[Christian_L]

I am pretty sure its along the following lines:

Once the Scan Engine software it deployed to the RT brick, the scan engine will be contently updating the memory map even if you aren't explicitly calling the variables.

That is why they say do not install the SE software unless you use it as you will get a performance hit for no reason.

If network publishing is enabled, these variables will be like any other Network Accessed Shared Variable - you can get a hook into them via the SV Engine and you have use them in your Windows PC application.

Here are some resources you may like to check out:

http://zone.ni.com/d...a/tut/p/id/7338

http://zone.ni.com/d...a/tut/p/id/7693

http://zone.ni.com/d...a/tut/p/id/4679

Correct. Figure 5 in the first link above shows it best. The Scan Engine, once installed on cRIO, always runs regardless of anything else going on on the cRIO. Once you have deployed I/O variables to the cRIO target, they will be hosted by the Scan Engine, even after rebooting the cRIO system. You have to explicitly undeploy or remove them from the Scan Engine using the LV project or Distributed System Manager to get rid of them.

Your host application using the variable nodes connects to the Scan Engine on the cRIO using the PSP protocol and read/write I/O variable value. This happens the same way whether the VI is running in the LV IDE or in a compiled EXE.

[bkb2005]

So then what is the point of using LabVIEW real-time target? I can just run a Windows EXE on an Intel i7-980x six core processor and outperform it.

[jgcode]

So then what is the point of using LabVIEW real-time target? I can just run a Windows EXE on an Intel i7-980x six core processor and outperform it.

Out-perform? Well it depends in what area:

The point is that the Windows target is not deterministic - if you have system requirements where you need to minimize jitter then you are going to want to use a Real Time target.

Additionally you may have environmental considerations whereby a PC will not cut it, and you require the ruggedness and toughness that a PLC/PAC will deliver.

You may need to run a system headless.

You may also want to make use of the FPGA for re-configurable IO.

The list goes on...

At the end of the day the choice will be application specific.