LabVIEW
Deploying stand alone real time application with CRIO 9056
Hi aysenurkurt,
@aysenurkurt wrote:
We developed an application using our tensorflow model with the vision development module. This .vi application was developed with labview 64-bit. We want to deploy this application stand alone and real time on CRIO 9056.
First (most important) question: why do you want to deploy a full-blown LabVIEW-64bit application (using TensorFlow and VISION) to a cRIO target?
@aysenurkurt wrote:
I don't know much about Labview. How can we do this?
- First you should learn more about LabVIEW.
- Then learn about using VISION and its supported targets.
- Then learn about your TensorFlow package and its supported targets.
- Then learn about LabVIEW-RT/-FPGA to fully faciliate your cRIO target.
Once you learned all that you are able to answer your question on your own - and you will be able to fully support your own application.
To answer your question: LabVIEW-64bit and cRIO targets don't play together (so far)!
The first issue you will run into is the fact that NI hasn't fully supported 64 bit for the RT targets. There is a Real-Time Module toolkit, and it does support 64 bit LabVIEW. But many of the supporting toolkits on the cRIO targets aren't 64 bit compatible yet. What this means is you can deploy and run VIs on a cRIO using LabVIEW 64-bit, but you can't run VIs that use DAQmx, VISA, XNet, or other common tools.
So I think you will need to download and install LabVIEW 32-bit. Both 32 and 64 bit versions of LabVIEW can be installed on the same machine with no problem. From there you'll need the Real-Time Module, and any drivers to recognize your embedded device. I have a hard time knowing what ones I need, and usually end up download and installing a bunch of extra stuff. A cRIO has two components. A Real-Time target which is usually running an ARM, or x86 based processor, and an FPGA. If you intend on writing code that needs to be deployed to the FPGA you also need the FPGA Module toolkit. This isn't required but allows for a lot more low level control.
Once things are installed you can launch LabVIEW, create a new project, add a remote target, and have it find you device. From there if you run a VI it will first compile it, then deploy it to your target and run it. You can make builds which can also run on startup of the cRIO, and then they don't need to be tethered to a PC.
I don't know what training is available for RT development, but I found this that I think works if you have an active SSP. Other than that I'd just recommend looking up examples, and maybe youtube videos. If you don't know much about LabVIEW in general, checkout these resources on the Wiki.
