Follow-up from "GSoC19 Executing Events in Parallel"
The following discussions from !227 (merged) should be addressed:
-
@kwolters started a discussion: (+1 comment) Oh and one last thing which would be good to verify in my opinion is too quickly check if the physics output continues to match our expectations (it should, but never harms to verify I would say). Easiest way to do this is to just rerun some of the scripts we used for producing publication output. Could you do this @simonspa (or did you maybe already do this)?
-
@simonspa started a discussion: (+1 comment) Here is still a
FIXME
. Did we figure this out in the meantime? -
@simonspa started a discussion: (+2 comments) Ah, I see you did already modify this thread pool implementation since
SafeQueue
changed. Since you now also simply submitstd::function<void>
in the "main" threadpool, you can simply delete this one and directly takeallpix::ThreadPool
here. -
Revisit the final statistics message to see if we can make this clearer I see the difference. I was just saying, I would usually be interested in just the average simulation time. I will add it this way:
|19:00:36.980| (STATUS) Average processing time is 124 ms/event, event generation at 8 Hz
|19:00:36.980| (STATUS) Average processing time for an event is 1 s
okay? -
Make sure that the manual (especially FAQ section code snippets) is updated with the new ways of accessing things and registering for messages.