Non-realtime Simulations

[Larry Brackney]

I've been experimenting with using LabVIEW to run simulations. I've
been using the point-by-point integral as a basis for constructing my
simulations within a metronomed while structure to set the step size,
and things work fine as long as the step size is achievable on my
machine. This is a problem for stiffer systems that require small
step sizes or for complex simulations that are computationally
challenging.

Has anyone built a "simulation framework" with a synthetic timing
mechanism that can run in non-realtime? I've poked around the LV
on-line examples, and haven't found anything quite like what I'm
looking for. I could use the ODE solvers as some other simulation
examples have done, but building up the simulation using free integra
tors
and a block diagram paradigm is important for this exercise.

Thanks in advance.

- larry

[Labviewguru]

Larry,

The problem with trying to simulate realtime is that you have to rely on the operating system.

If you don't have, or don't want to get LabVIEW Realtime, probably because you want to run in a software environment, then I would suggest that you get a realtime version of Linux and install LabVIEW for Linux on top of it. You may be able to get your local sales rep to get you an evalutation copy of LabVIEW for Linux, but I doubt he or she can help beyond that. However; there seems to be extensive support for this on NI's website, and with their engineers.

I'm sorry I can't be of any more assistance.

I am in the process of setting up a linux workstation server, and haven't yet started developing in Linux extensively, but from what I have seen, t
here isn't really much of a transition.

Good luck, and let us know how you do.

[Larry Brackney]

Thanks for the reply.

> The problem with trying to simulate realtime is that you have to rely
> on the operating system.

Actually, I DON'T require a realtime simulation. I want to have
the structure containing free integrators to run at whatever speed
the OS and host can manage. The structure has to be set up to keep
track of iterations and create properly scaled time values for plotting
results. I can replace my system clock-based structures with a bit of
effort, but I was hoping someone might have already done this.

> If you don't have, or don't want to get LabVIEW Realtime, probably
> because you want to run in a software environment, then I would
> suggest that you get a realtime version of Linux and install LabVIEW
> for Linux on top of
it. You may be able to get your local sales rep
> to get you an evalutation copy of LabVIEW for Linux, but I doubt he or
> she can help beyond that. However; there seems to be extensive
> support for this on NI's website, and with their engineers.

I'm running LabVIEW for Linux, but on top of a non-realtime distro.
Hadn't really occured to me to run realtime simulations under Linux.
If I need to do so, I'll keep that in mind. Thanks for the tip!

> I am in the process of setting up a linux workstation server, and
> haven't yet started developing in Linux extensively, but from what I
> have seen, there isn't really much of a transition.

THAT I can help you with. 🙂

- larry

[Labviewguru]

Any comments on Distributions of Linux?

I am trying to setup a linux system as both a workstation and a server (file server, and connection port to internet).

I chose Suse Linux 7.3 on the recommendation of my brother (thanks for flying to England before installation, ya creep!) but believe that Suse may not be the ideal choice (no support except installation, and they offer "enterprise connectivity" that would "solve all of my internet connection problems" for a mere $1,000.)

Any advice? I can only choose RedHat if I am going to maintain this system as a server/workstation.

Please feel free to reply to this via email, as this probably isn't the proper forum for this.

labviewguru@nethere.net is your best shot.

Thanks!