I have a workflow that runs with dask and/or with concurrent.futures. Basically, depending on the machine that the code runs, it runs on dask or with the futures executor. Now, inside my code I am using the performance report from dask, like:
from dask.distributed import performance_report
executor = concurrent.futures.ProcessPoolExecutor() if args.futures else dask.compute() #### this is a simplification of the actual executor.
with performance_report(filename="dask-report.html"):
executor(job)
Now, in the way that my code (successfully) runs, even if the executor is with futures, I get a performance report from Dask.
My question is, is this performance report doing/measuring something? Again, when opening the HTML file, I see some performance plots for CPU and memory, which I would like to know if they are accurate.
thanks