Insufficient system resources error

[napview]

I have not looked at those init VIs yet but the error is repeatable but it does not mean where it stopped is the real problem regarding to Visa issues. The GPIB buffer must be overflowed with repeating open and close visa improperly. To troubleshoot it, if you have time, try to copy at least two loops (start from outer or inner) from the diagram into a new vi and run it a few hundred times (no long time delay in between). Then increase to 3 loops based on the same program and repeat it. This way will take some time (a few hours most) but it sure will point to the real problem.

Troubleshooting with only one loop without a visa close is not exactly what you have in the program. This error ONLY occurs in more 1 loop structure either with the same instrument or not.

[REash]

One loop takes twenty minutes, so doing it a few hundred times is not happening.

Dennis, it is now doing the group of 6 for the second time around, and has not received an error yet.  I'll report back on Monday on how the remainder of the test went.

Message Edited by REash on 12-08-2006 03:25 PM

[Dennis_Knutson]

Can you post an image of the block diagram. I don't have 8.2 but I don't see how the number of loops could affect anything at all. I also don't know how the GPIB buffer is affected by either a VISA Open or VISA Close.

[napview]

The number of loops does matter because the visa open creates a different reference everytime. Then one loop has only one reference but two would make it different regarding to the way the GPIB port works. One thing I did not tell you was when the error happened, put a probe on the reference out of the first error in the visa chain, you should see something like "GPIB0::6::INSTR:x" or "(x)" where x is an integer, 6 is the GPIB address. This number to me indicated that the visa resourse is not closed properly and it would increase the index of the that resource until it ran out of buffer memory or the like. This was seen only at more than one loop structure.

The bottom line is there is not a good explanation of why this error happens but that is what the failing pattern seen so far from older to the latest version of labview.

[napview]

You can simply create a simplified diagram of the loop structure without the actual testing functions , just open visa, init, take one short reading of the instruments, close. It would only take several minutes to do this, and the total time for about 4 loops (average number of nested loops) for one reading takes about one minute, run it overnight or for about couple hours for a few hundred times total: a x b x c x d times (a,b,c,d are number of iteration per loop a,b,c,d).

[Dennis_Knutson]

You're going to have to prove that. The session id is exactly the same for the three separate Open VISA I have here.

[napview]

Of course they are the same from your diagram because of the same visa resource. What I meant by more than one loop will matter because they are used by different resources therefore causing the buffer overflow. It is misunderstood that more than one loop will cause the error, NO if only one same resource (only one instrument) is used. Then there is  no need to even close the visa for it because there is only one reference existing in the buffer but more than two sure causes some headache.

[Dennis_Knutson]

You are confusing. You just said that every time a VISA Open is called, it creates a new reference. Of course you get a different reference for a different instrument but that is not what the problem is. He was repeatedly opening a reference to the same instrument and not calling a VISA Close on that reference. Go back and read the entire thread. If the problem is fixed, it would seem that the issue is with the single reference and nothing at all with the the other references because he made no changes at all there. That's why the problem is so weird to me.

And what buffer are you talking about? A memory buffer?

[REash]

napview, you are very confusing to me as well.  I'm thinking exactly the same thing as Dennis about those last few posts of yours.

Dennis, I am at home now so I can't get an image of the block diagram to you, but it is very large.  It takes up WELL over 2 19" screens set at a decent resolution.  I will try to get an image on Monday with SnagIt.

[napview]

Actually if you read the whole thread again then you are the one who got confused and lost track of the raised issue. Only when I pointed it out the visa open/close issue then you jumped on that but before you shotgun at driver, IVI and even the size of the test program which all were not related at all. Anyway I used to have this problem randomely until that was how I got rid of it.

Another thing if you did not know is that this could happen to ANY resources including active X, .dll (SW only), not just the HW side. Solutions were found based on real life experience!