I am trying to run Big Stitcher on a computing cluster in headless mode. I recently noticed that the number of threads that are spawned is much higher than I would expect. For example, when running a script to save a dataset as an N5 file, I set the number of threads to be 24, yet the # of threads that are spawned is 119. (Initially 95, and then 24 when the program begins to write the data). Another example is when I run Big Stitcher Pairwise Shift, well over 1000 threads are spawned. I also ran a test by just running a script that waits for a few minutes (“wait(10000);”) and even that spawns 95 threads. The amount of threads being spawned is overwhelming our institution’s cluster and I’m not sure how I can limit the # of threads.
I have tried using the following command, run(“Memory & Threads…”, "maximum=1000 parallel=16 run”);. But this seems to not have any effect.
I would greatly appreciate any help possible to control the number of threads.