labview 7.1 and Dual Core Processors

[task]

Hi,
I was wondering if anyone has encountered any problems using Labview 7.1 and the newer dual-core processors. I have a Dell Dimension DXPO51, 2.8 GHz Pentium D with 1 GB ram. There is NO virus software or any other non-essential programs running in the background. The computer is totally dedicated to controlling a machine.
Other NI boards installed: PCI-6221 for analog inputs, PCI-6602 for reading encoders, PCI-7352 Motion Control board to control 2 motors.
The problem:
Periodically when I start my program, it locks up and I get the "Tell Microsoft about this problem..." screen. After it exits out of Labview and I restart, the Run arrow is not broken but the program will not run until I make a change on the block diagram. Then I can run again. Doesn't happen all the time, doesn't happen after running after a certain amount of time, doesn't happen after a particular change is made (i.e. adding case statements, adding an event, etc.). In other words, totally random.

I have tried totally uninstalling ALL National Instruments software, including cleaning the registry, and re-installing everything. Didn't help.

The machine will be shipping with the executable file only but I'm still leary of problems arising due to this issue.

Any help or suggestions of things to try would be appreciated.

Thanks,
Teri

[MattP]

Hi Teri,

I've seen scattered reports of users having trouble with dual-processor machines with earlier versions of LabVIEW, though I'm not aware of verified solutions in those cases.  The dual-processor operation should be completely transparent to LabVIEW, although I know XP has had some difficulties in providing the abstraction correctly.  A workaround that has been successful in the past is to set the processor affinity for LabVIEW to a single processor.  This should eliminate the dual-processor nature of the system from consideration.  If the problem still recurs after that, then the issue stems from another source and we can investigate further.  Does it show this behavior with any program (or just by opening the LabVIEW environment)?  How often does this occur (10%, 50% of the time)?

[task]

Thanks for the response. The problem with setting the affinity is you have to do it manually everytime the machine is reset...there probably is an automatic way but I haven't had time to investigate it. I can't determine if it happens with any other program since I only work in Labview on that machine...it's dedicated to run a test machine. As for frequency, in an 8-hour day it will probably happen 5 or 6 times during debug. The lockup occurs after I make some changes, hit Ctrl-S, Ctrl-E, Ctrl-R to run the program. When I re-enter, I have to make a change on the block diagram to run again...the run arrow is not broken, but doesn't run the program. My next machine is also a dual-core machine along with a PCI-7356 so I'll see if that one has a problem.
Thanks again,
Teri

[MattP]

Hi Teri,

There's a pretty decent article on methods by which you can automate the affinity-setting process located on Informit.  I recommend giving one of those methods a try - it may provide us with a very easy workaround.  Failing that, we'd need to know if your application shows the same behavior on another system to determine if the cause is some issue with the VI's, the operating system, or LabVIEW itself.  Does LabVIEW give any indication of the errors its experiencing before or after the crashes?

[viksiv]

Hi,

I have recently deployed stand-alone Vision application that runs 2X 1.4MPixels IEEE1394 cameras on HP Dual Core 3.2Gh -1GB computer equipped with 2X Firewire Cards. Application was built in LV7.1 and Vision 7.1. Computer itself does not have much of NI software except IMAQ1394 v2.0.2 (comes with Acquisition package together with IMAQ and MAX), LV Run-Tiime Engine v7.1 and corresponding licenses. This rather simple application is used for monitoring of optical systems aligning processs and has been running for about a month now without any problems.

[task]

Hi Matt,
I will look into the affinity discussion. Unfortunately I do not have time to swap systems...my machine is shipping out of the country in 2 days and I have to finish clean-up items. Labview gives no indication before or after the crash. If it happens today, I'll do a screen shot and forward it. I did have a lockup the other day even after setting the affinity to one processor. For some reason I'm leaning toward the problem being involved with the motion control board. This is the third system I've used these boards and I seem to remember this problem occuring on one with a Xeon processor. My next system has the dual core also along with a PCI-7356 motion control board...I'll keep tabs on that one as I start digging into it.
Thanks again,
Teri

[task]

Hi Matt,
I think I figured out was is happening to cause the lockup but I don't know if it has anything to do with the dual-core processor (probably not). The lockup occurs when I hit Ctrl-S, Ctrl-E, Ctrl-R too fast. It's a large program so the save takes a few seconds, so I think the combination of Ctrl-S and Ctrl-anything until the save occurs causes the problem. If I wait until the hourglass goes away, it doesn't happen. So...I wait!
Thanks again,
Teri