Do queues require semaphores (i.e. are queue operations atomic?)

[chassan]

Hello,

Do queue functions require the use of semaphores to prevent mutliple vis from trying to access them simultaneously, or are the queue operations atomic, thereby eliminating the need for semaphores?

Thanks,

Chris

[Andrey_Dmitriev · Accepted solution]

No, semaphores normally not required.

[GregFreeman]

If you are worried about accessing the same queue in different spots, does this imply you are trying to send the same data to multiple different sections of code? If so, you should consider notifiers instead. Otherwise, if subVIs are operating in parallel, whichever executes first will get the data from the queue, the other subVIs will not.

[nathand]

This has been previously discussed here.

[DFGray]

Note that under the hood, semaphores are currently implemented using queues.

[chassan]

Hello for(imstuck),

Actually, I am actually using the queue to allow multiple different sections of code to report their statuses to one piece of code that cases out the statuses; like a multiplexer.

Would a notifier work this way?

At any rate, I've tested it out a bit with queues, and it seems to work well.

Thanks,

Chris Hassan 

[DFGray]

The queue is actually the best way to do this.  Notifiers are used for broadcast situations (one source to many listeners).  Queues are appropriate of many sources and one listener.

[chassan]

Hi DFGray,

Thanks for the information.  Your statement reflects what I thought to be true, but some company in my way-of-thinking is always comforting (whether we're actually right or wrong!).

Best Regards,

Chris