Actor Launch problem

[MikaelH]

Hi

I have a very strange Actor Launch problem. I start an Actor to do some calculations for every attached DUT. And that has been working great, 1,2,3 Actors(DUTs) no problem.

But now I’m starting to get problem when I have around 5 or more Actors that I start of the same type.

The screenshot below shows my problem.

The queue used to get the Actor’s Enqueuer back after the Asynchronous start of the Actor, didn’t receive anything.

I debugged the code, and could see that the Queue-Reference was the same for the Enqueuer and the Dequeuer. Also no errors were found on the Enqueuer Error in.

Has anybody seen something similar?

The code only fails if there are too many started actors at the same time!

Cheers,

Mike

[drjdpowell]

Are you overriding “Pre-launch Init.vi” such that it might produce an error (thus preventing the sending of the enqueuer back to the Launch VI)?

[MikaelH]

Are you overriding “Pre-launch Init.vi” such that it might produce an error (thus preventing the sending of the enqueuer back to the Launch VI)?

 

I'm not, it's so strange, why can I run a few Actors in parallel but not too many?!?

I might do a 2012-64bit version to see if that makes the Queue Reference work better.

[Aristos Queue]

Haven't seen anything like this, and I've ramped lots of actors at the same time of the same and of various types.

 

Post your question on the http://ni.com/actorframework community site... let's see if any of the other contributors have seen this.

 

As a test, try adding a time delay in your code between calls to Launch Actor.vi. I *really* hope that does NOT fix it, because if it does fix it, you're really going to stress me out that there's some sort of deep race condition, probably in LV itself, and those are NEVER easy to track down, and we are WAY late in the 2013 release cycle.

[drjdpowell]

Do you have a four-core machine?  I ask because then four clones will initially be reserved of the various shared-clones at start, and then the fifth instance will require a new clone.  In working with shared clone pools (through an 0x40 reference to "Open VI Ref”) I have sometimes noticed what seems to be a delay in getting that extra clone when it is asked for soon after the reference is opened.   BUT, I have never seen this delay actually stop code from working.

[MikaelH]

Do you have a four-core machine?  I ask because then four clones will initially be reserved of the various shared-clones at start, and then the fifth instance will require a new clone.  In working with shared clone pools (through an 0x40 reference to "Open VI Ref”) I have sometimes noticed what seems to be a delay in getting that extra clone when it is asked for soon after the reference is opened.   BUT, I have never seen this delay actually stop code from working.

I did see this problem on the production machine which has 2 cores, but not seen it yet on my 4 core machine (but I haven't tried to run too many DUTs).

I don't think it's an Actor related issue, I think it could be  a "Reentrant,Fire & Forget(0xC0)", "Preallocated Clone" , "Queue Reference" - Problem.

If I have time I'll create a test application to see if I can cause it to fail.

It might be the production computer that has some problems.

[MikaelH]

So I made a test application using roughly the same design as my real application.

ActorTest1.zip

And of course this works fine on both computers, I can spawn 50 Actors at the same time without any problem.

 

I still might put in a small delay between the Launch of every actor just in case.