06-19-2012 01:46 PM
Hi all,
I am an experienced LV programmer, but not with the embedded systems (RT, FPGA yes).
I have a friend who has been doing some work with a pandaboard (ARM9 dual core) and I am going to start working with him on a project, so I will have my own board.
I am just wondering if I can use the pandaboard with the LV for ARM system (I have a developer's subscription). I really don't know anything about embedded LV, and I can't seem to find good resources for getting started.
So...
1) Can I use it with the pandaboard?
1b) If yes, is it so convoluted for a beginner that I am better off with one of the NI-sold eval kits?
2) Can you point me to a good resource for getting started? Not just "jump-right-in" but maybe something even higher level so I get a feeling for what systems the kit will run on and how it all works?
Thanks!
Jed
06-19-2012 11:16 PM
Hi Jed,
It would be great if NI made it easy for us to target your Pandaboard, but alas their marketing strategy is to whet your appetite with underperforming, 6 year old Cortex-M3 or ARM7 evaluation boards (which actually run LabVIEW in an (intentionally?) inefficient manner - see here) and then convince you that their own H/W is really the solution to your problem. While it might be a solution, you end up paying an arm and a leg for it compared to the Pandaboard for example.
The only possible option I am left to try that can be coded up in LabVIEW which is reasonably priced, low power, customisable, capable, small form-factored, soft real time and OEM friendly is a Computer On Module containing an Intel x86 Atom CPU that runs Linux. The Qseven form factor suits my needs and hopefully my evaluation of it is favourable.
If you still want to use your Pandaboard, .NET MicroFramework is worth looking at and is growing in popularity. I plan on doing this (or even going to plain embedded C) should my evaulation with LabVIEW for Linux on a Q7 not prove satisfactory.
06-20-2012 12:18 AM
Well, thanks for the honest opinion. It seemed like those eval boards were a few generations behind.
I'll keep my ear to the ground for new info, but this doesn't seem all that promising.
Is it possible that keeping LV for ARM updated with the changes in HW is too much work for the payoff?
06-20-2012 12:21 AM
> Is it possible that keeping LV for ARM updated with the changes in HW is too much work for the payoff?
In my opinion it is. Since you would need to write all your own peripheral drivers in C for the tier 2 board you planned on targetting which must be able to run the Keil RTOS BTW (I'm unsure if the Pandaboard can do that).
06-20-2012 12:24 AM
Looks like Pandaboard can't. No biggie. I might look into those other Eval boards, but it might just be worth waiting a little longer for the next iteration (if it comes).
06-20-2012 12:39 PM
Hi Jed and Peter,
Peter answered correctly. You can't directly target that board without at least Tier 2 development. I wish it was easier for you.
We are actively developing labview embedded and realize that people want more targets. As has been noted by developers before on this forum before though any development is not a simple undertaking and takes time. Just by reading the embedded outlook for 2012 it should be clear NI isn't ignoring the development or needs of the embedded industry.
06-20-2012 11:30 PM
>> Is it possible that keeping LV for ARM updated with the changes in HW is too much work for the payoff?
>In my opinion it is. Since you would need to write all your own peripheral drivers in C for the tier 2 board you planned on targetting which must be able to run the Keil RTOS BTW (I'm unsure if the Pandaboard can do that)
Jed, I misread what you wrote. For one individual to turn a Tier 2 board into a Tier 1 board it wouldn't be worth the effort, however I maintain that it is certainly worth NI's effort to do this for a couple of new and capable boards. It does not take them long at all, but they have intentionally been holding back. In fact we received a quote from NI to convert a Tier 2 board of our choosing to a Tier 1 board and for the major peripherals. NI told us it would only take them a couple of person-months work to do. Lets say it took them 3 months to code all the peripherals for a new Tier 1 board. I'm even then happy to triple that time out to 9 months for them to debug, qualify and document it, heck lets take it out to 1 year for the inevitable project overruns encountered !
1 person year is a drop in the ocean of the total amount of R&D resources NI has available. NI has at least 1000 R&D engineers working for them.[1] So clearly they are choosing not to do it, and my suggestion as to why - well, read my first reply to your question again.
[1] Last year NI spent $210.2M on R&D. Assume labour and loading rates to be approx US$90/hr with a 40 hour work week.
06-21-2012 01:57 AM
Hi Jesse,
This is great news. I've been eagerly awaiting some enhancements for LabVIEW Embedded for ARM. (Great product once upon a time that has been allowed to lapse, but can have new life breathed into it.) Do you have any supporting details?
You mentioned that NI realize that people want more targets. Will we get a new Tier 1 board?
I've read cover to cover the Embedded Systems Outlook 2012. I could find nothing that gives me an outlook. What did you find?
Regards,
Vito
06-22-2012 12:43 PM
Hi vitoi,
For immediate product development I recommend looking at post ten of this thread, where Sam, the ARM PSE explains where we are at. I double checked with the Sam that the stance he outlines is still accurate.
06-24-2012 11:58 PM
Hi Jesse,
I'm confused. You mentioned "We are actively developing labview embedded and realize that people want more targets." You then refer us to this thread post 10, which states "At this point we are still supporting LabVIEW embedded for ARM but we do not have a defined roadmap for future developemnt. ... the future of this product is uncertain ..."
Sam, the NI ARM PSE goes on to state (page 2 post 2):
"Since the future of the product is uncertain at this point we tyically offer the following guidelines:
If the current features of the module meets your needs and you are working on a relativly short term project (~1 year) we recommed using LabVIEW Embedded for ARM.
If you are working on a long term, multi year project, or the current features of the toolkit do not meet your needs we can suggest alternatives (hardware and/or software)."
We know LabVIEW Embedded for ARM has been effectively abandoned and Sam's post even states that if your project goes beyong a year we should look at alternative (hardware and/or software), presumably within and without NI.
I'd like to think that LabVIEW embedded for ARM will be developed after 6 years of falling behind, but I think it's wishful thinking. If you have some information that turns my wishful thinking into hope, I'd like to hear it.
Regards,
Vito