Fork me on GitHub
#aleph
<
2017-06-06
>
dm316:06:45

hmm, interesting

dm316:06:05

somehow fixed-thread-executor is not working as expected

dm316:06:19

i.e. as java.util.concurrent.Executors/newFixedThreadPool

dm316:06:33

@hansen-pansen try (e/fixed-thread-executor num {:control-period 100})

dm316:06:57

your test duration coincided with the default control period of the executor

dm316:06:15

that’s how often the instrumented executor will adjust the number of workers in the pool

dm316:06:54

there’s also :initial-thread-count parameter

hansen-pansen16:06:30

Thank you very much! I'll give it a try and will also check with a snapshot build of manifold, in case this has been fixed.

hansen-pansen17:06:36

@dm3 Again, thank you very much! Both :control-period and :initial-thread-count showed the expected behavior. I settled for :initial-thread-count, as this is more consistent of what I would expect from a fixed thread pool. As a side note, [manifold "0.1.7-alpha5"] behaves like 0.1.6.

dm317:06:46

I don’t think it should be fixed

dm317:06:01

you might create a PR with enhanced docs though

dm317:06:21

although

dm317:06:45

maybe fixed executor should have initial thread = num-threads from the start

mping20:06:59

Im having trouble using d/zip with `d/chain