You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
return@sprintf"%5.2f %2s/it" (ns_per_iter / divideby) unit
end
end
At the moment speed is always shown as time/iteration which is nice because it's general. However, when running many simulations, the loop is over time so often people think in terms of simulated time per wall clock time. E.g. in weather forecasting we often think in terms of simulated days per day. Is there general interest for a PR to extend Progress(...,speednorm::Dates.Period) with speednorm of units simulated time / iteration (= the time step) to allow for say
1 m/it
to be displayed with speednorm=Dates.Minute(30) meaning that every iteration simulates a 30min time step as
Weather is speedy:100%|███████████████████████████████| Time:0:00:29 (80.91 years/day)
Meaning the model ran at a speed such that in 1 day it would simualte 80 years of the Earth's atmosphere. If there's more general interest to make this feature available here too let me know. I believe it would be of interest to anyone who would use it to monitor the progress of a simulation which integrates in time
ProgressMeter.jl/src/ProgressMeter.jl
Lines 683 to 700 in 83e619f
At the moment speed is always shown as time/iteration which is nice because it's general. However, when running many simulations, the loop is over time so often people think in terms of simulated time per wall clock time. E.g. in weather forecasting we often think in terms of simulated days per day. Is there general interest for a PR to extend
Progress(...,speednorm::Dates.Period)
withspeednorm
of units simulated time / iteration (= the time step) to allow for sayto be displayed with
speednorm=Dates.Minute(30)
meaning that every iteration simulates a 30min time step as?
@samhatfield what we just discussed
The text was updated successfully, but these errors were encountered: