Single-Threaded Workers¶
By default, a worker will use as many threads as a node has cores. This allows
the worker to run many computations in parallel. You can limit the number of
threads a worker can use with the worker_options keyword argument. If you
set the number of threads to one, it will allow the worker to run computations
mostly synchronously.
import coiled
cluster = coiled.Cluster(worker_options={"nthreads": 1})