Allow for Stacked Feedback Nodes

[PaulRB]

When using feedback nodes, you can store data like you can with shift registers. However, you cannot get multiple data points from previous iterations like you can with stacked shift registers. I think this is best shown with a picture, so here is my idea:

 

[Darren]

With my increased involvement in LabVIEW FPGA programming over the past couple of years, I can definitely see the need for this.  Kudos.

[Dragis]

i definitely like the concept. we'll have to work on the look and feel a bit. i'll try to mock up some alternatives and post them at some point.

[Norbert_B]

I like the idea, but there is something, i would do otherwise. Using your suggested layout, the feedback node would "increase delay" from bottom to top of the node, which is just the other way round than shift registers. Without proper symbolics, this would be irritating.

On a second note, i am wondering how initializing of the node is properly done....

Nevertheless, a good suggestion and i hope R&D can figure out a decent way to address the two points i mentioned....

[Darin.K]

Initialization would probably be an array as we do now. I usually wish that the enable terminal was movable, if it were on top then you could pull down to increase the delay. The alternative I have wished for is to set the delay and then have an array output of all the stored values, a simple buffer.

[vitoi]

Feedback nodes have been mentioned elsewhere in the exchange. I personally liek the shift register as it is clearer to understand and part of "basic" LabVIEW.

[Getwonder]

This would be useful but how the repetition would be done. As the feedback is done only through the repetition of the same thing.

[Darren]

---

@vitoi wrote:

Feedback nodes have been mentioned elsewhere in the exchange. I personally liek the shift register as it is clearer to understand and part of "basic" LabVIEW.

---

I used to say the same thing, and I still feel mostly the same way...for desktop LabVIEW programming.  For LabVIEW FPGA programming, however, feedback nodes represent a unique semantic...particularly when coding inside single-cycle timed loops. I see this idea as beneficial for this programming use case.

[AristosQueue (NI)]

Chaining feedback nodes together gives the same effect as adding the stacked syntax. If we're talking FPGA, is the stacked syntax something that mimics existing iconography? Or would those diagrams be clearer to FPGA developers if they continued to use the chain of feedback nodes, showing each one chaining to the next one?

[Darren]

Chained feedback nodes are currently the norm...I contend that with the stacked feedback node idea, diagrams like the following would become easier to understand to all LabVIEW FPGA developers:

[AristosQueue (NI)]

Um, Darren, I hate to tell you this, but I don't think that stacked feedback nodes can help your example program at all. You're doing operations on the interim values. A stacked feedback node wouldn't allow that -- you'd be able to get the set of historical values and nothing more. The constant toggling that you're doing here would still require a set of feedback nodes.