Why is the number of generated parallel loop instances for a For loop (in the For Loop Iteration Parallelism dialog) limited to 64? LabVIEW can handle more than 64 parallel threads, can't it? I will have a machine with 104 cores--will two For loops each configured to generate 52 parallel loop instances give the results that I want (i.e., 104 parallel executions of my VI)? Will a (non-parallelized) For loop opening my VI in call and collect mode 104 times give me the same results? Is a parallelized For loop containing a call to a single VI any different from calling the VI with call and collect inside a For loop?
Solved! Go to Solution.
You can change it to allow up to 256 parallel instances by adding an entry to the labview.ini file.
Thanks, altenbach, for that information! So the limit of 64 is not a fundamental limit of the For loop and is there just to protect us.
Still, I'm still curious--is doing multiple call and collects the same as running a parallelized For loop? Is there some optimization in the parallelized For loop that can't be had with the call and collect method?