joesinger12
commented
5 years ago SC - There are certainly ways to track analysis progress in a more continual way. If we were doing it a year ago the most obvious would be a mechanism OpensStudio offers using Qt Signals and Slots.
Given where we are today and the strong likelihood of no longer using OS to manage the analysis simulations, I would recommend that we simply incorporate this into the scheduled work to replace the OS RunManager. The mechanism we use to track CSE simulation progress could probably be modified to do the same for E+ runs - giving the users a much more informative progress indication.
Re: CEC suggestions - coming up w/ estimates and showing progress based on time rather than physical/reported analysis progress is certainly something we could do. To date we split the 100% progress into several pieces, the largest of which are simulations and reporting - both of which CAN be pretty unpredictable in duration and can vary widely based on computer hardware. There are however ways to estimate simulation runtime based on complexity of the model. We could also store simulation runtimes from past runs and use those as estimates for future runs.
Lots of options ... if this is indeed something that the CEC is prioritizing.
Original comment by: joesinger12
12/28/18 CEC - I have been doing a few curiosity experiments to see how long the runs take. Of the handful that I have timed, it seems that there may be a way to predict how long a run will take, with the thought of maybe indicating on the progress pop-up the time left for the simulation to give the user an indication of how long it will take. It seems that if the time it takes to get to 14% is multiplied by 11.4 or so, that gives a good indicator of total run time. And if the time it takes to get to 39% is multiplied by 3.1, also gives a good indicator. What do you think?
Reported by: joesinger12
Original Ticket: cbecc-com/tickets/2918