which operating system is best for computer control using LabVIEW

[topaz]

I am using LabVIEW 7.1 and want to write a program to control a system of pumps, heaters, and  compressors.  I am wondering if using Windows XP is a bad thing, given how many background processes computers tend to use these days.  I don't want my computer to slow down for intermittant checking for updates, or any other sort of routine maintenance computers are set up to do.  I have heard mixed things about windows reliability and computer control, most of which is people saying windows could crash and completely ruin my tests.  Does anyone have any advice on the best operating system to use for continuous computer control of my system? 

[Wayne.C]

Look at LabView Real Time running on XP.  My experience has been that this is stable and deterministic.  You will still have to tweek XP a bit to control the background tasks.

How many I/O are you talking about?  You could use an executable deployed on a Compact Field Point.  No windows issues to affect operation.  You could then use a remote panel or another program for the user interface.  I've had good results with this configuration as well.

[PaulG.]

I'm very happy with XP. Since the days of win 3.1 (?) this is probably the most stable, reliable and easy to use version of Windows yet. It's so good this whole "Vista" business makes me very, very nervous.

[Ray.R]

You may get many differing opinions on this one..

It's probably related to what good / bad experiences people have had in the past with the OS.

My prefered OS has always been Win-2000.  Maybe since I had no trouble working with it & Labview.

XP?  Hum... do a search on Win-XP and do a bit of reading.  However, I must say that since SP2 and recent updates, it has improved.. and is Okay..

And yes... way too many processes running in the background.  I've been removing them.. one by one 😉

[tst]

If you want guranteed timing, you have to use a Real-Time system with RT programming techniques. We do have control systems running on XP for months at a time without crashing and without being too slow (mind you, their normal response frequency is only a few Hz). The key is to make the computer a dedicated one. If people don't install junk on XP, then it's stable and runs perfectly well for as long as you want it to. If you disconnect it from the network, get rid of the virus scanner, get rid of the automatic updates and remove unnecessary services (that one might not be necessary) it will probably react fast enough for most stuff you would try to throw at it, especially if you set your process to have high priority.

All this, of course, is only relevant if your program doesn't require high performance (speed, memory, etc.). If it does, then you would have to do some special things no matter which OS you're using.

[Jarrod_S.]

If you plan on using a desktop PC for control type applications, then your best option in terms of performance will be with a Real-Time operating system. Windows is not deterministic for many reasons. Determinism involves the ability to respond to an event within a specified amount of time with an acceptable amount of error (jitter). First of all, Windows is not deterministic because it has to service mouse and keyboard events, as well as anti-virus software, automatic updates, etc.

You can, however, convert an existing desktop PC to a Real-Time target with a Real-Time OS and use it for your control type application with the deterministic behavior of PXI, Compact FieldPoint, and other Real-Time targets. Check out this link for more information.

On a side note to a previous reply, using the LabVIEW Real-Time module on Windows XP won't make your application deterministic. Your best bet, as others have noted, is to use the LabVIEW RT Module on an RT OS with good RT programming practices (avoiding memory allocations, etc.).