HANG when aborting timed loops: Is 42CHH33W really fixed??

[eberaud]

I am experiencing frequent occurences where the Stop Timed Structure.vi hangs. I found online that issue #42CHH33W used to be a possible cause, but it's supposed to be fixed in LV8.2 and I'm using LV2015! I tried hard to identify a pattern for the hanging, and implemented a system based on semaphore that would guarantee that I only abort the loop if it's sleeping (waiting to start its next iteration). If it's awake, I simply rely on a Boolean to exit the loop. Even now it still hangs sometimes. Getting a little frustrated here 

Anybody has experienced such scenario before? Do you think I should try the suggested workaround (putting Stop Timed Structure.vi in a another timed loop that is in sync with the timed loop I want to abort)? What does "in Sync" really mean?

Thanks

[eberaud]

Bumping this thread as I really would like to see some comments, even if you can't propose any fix. At least I would like to know that I am not the only one who encountered this hang issue. Please 

[smithd]

Well since you want comments too :

At one point I was working with some folks on making a timed loop you can plug into. We played around with a variety of features including using a custom timing source, using the timed loop abort functionality, etc. I can't remember them all anymore, but we had a lot of issues. For example, we had totally bizarre issues using a named timing source (if you didn't know, you can wire in a string name like "1 MHz absolute" rather than using the dialog)...sometimes it would work fine, sometimes the loop just would not ever run. On the 9068 in 2013 (first release of linuxrt), the 1 kHz clock in our testing had jitter of up to 1 clock cycle -- that is, if you request 1 ms it might give you 2. I don't know if this is fixed now, I just only ever use the MHz clock.

All this is basically to say that after really trying to push on the features of the timed loop, I came away less a decent chunk of time and uninterested in using it

The solution we went with was to just make a really simple class for a timer function, you pass in a time to wait in ns and you run this inside of a non-timed loop (to avoid timed loop overhead), which in turn was inside of a timed sequence structure (to make it single threaded and give rt priority). We didn't see any meaningful performance degradation going this route. For your purposes, you could create a 'timing source' which waits on a rt-fifo in polling mode for example, and tells you if an exit or some other message was received or if the wait time elapsed. Admittedly you do have to write the math for this and keep track of when iterations start and stop, and make it work cross platform -- all the stuff the timed loop does for you -- but you have total control over it and can stop hitting your head against the wall, so thats what we did.

 

Edited August 31, 2016 by smithd

[eberaud]

Thanks smithd!

I should have mentioned that I'm just running on regular Windows, no RT anywhere. You will probably reply: then why on earth are you using a timed loop? Well back in the days I naively thought that timed loop would be more deterministic and have higher performance, but later on I read that not only is it not more deterministic when run on Windows, it also has additional overhead and end up using more CPU than a regular while loop. Yet we kept it because we liked the possibility to abort, and the fact that we can use the "Actual Start [ i ]" left data node in order to log the time stamp of the actual iteration in our log file.

So it seems that you also found stinky stuff with the timed loop eh... Well I might end up getting rid of it altogether, replace it with a while loop, and handle the timing myself...

Edited August 31, 2016 by Manudelavega

[eberaud]

I found the culprit!! 

Back in the days a coworker of mine had the noble intention of increasing the performance of our application by adding some code that would assign a HIGH priority to its thread in Windows (you can do it manually through the task manager).

It took me a while to figure that out, I’ve been stripping code from our application bit by bit, until that little subvi was the only thing remaining!

The online documentation does recommend not to modify the priority and warns that unexpected behaviors might occur. So I’m not sure if NI will try to fix anything. But at least you guys should add this to your troubleshooting database so that next time you hear about this symptom you can tell the developer to check this setting.

I’m attaching a project that will reproduce the issue very quickly if you set “HIGH PRIORITY”. It will run smoothly all night long if you set “NORMAL PRIORITY”.

Closure feels so good, workarounds suck 

Stop Timed Structure Hang.zip

[smithd]

But...but...setting high priority should make the whole labview execution system high priority. Very weird.