Track min and max of important quantities#376
Conversation
|
Hey Bruce, this is great! There are a bunch of things the linter picked up if you wouldn't mind fixing those. I left some other minor comments |
|
Is there a reason not to just output the extrema of all the mass fractions? I guess it can get verbose... |
|
And can you post an example of the log file with these modifications? Curious to see what it looks like. |
|
Here is what the output looks like: And here is the log file: |
|
I think I would be fine replacing the integrated quantities with this. |
|
@baperry2 : A small suggestion - can you either print on the screen as exponentials with only a few digits or print just the one decimal point. We perhaps don't need that much resolution in min and max. You can retain the full in the log file. |
|
I want to keep the integrated quantities. I use them for things like TG and quick plotting of integral quantities. |
I second this. Can we also make it look more like the other ones instead of whitespace before the name, make the whitespace between the name and the equal? Also, a space between |
|
Okay, I think I addressed all of the above comments. I also added tracking for the sum of mass fractions (actually sumY - 1 to make any discrepancy easier to track without requiring crazy precision). And I added an option to print extrema for all mass fractions by setting |
|
Excellent idea. Can you repost the new output here? |
|
Log file: Output: (without the option to print all species to keep these reasonably sized) |
|
Sweet. One last thing. Can you remove |
|
Do we need to list |
|
Not required, but makes it easy to make plots using grep in the log |
|
Maybe we could just use |
|
@drummerdoc I've found having TIME in the output handy for the same reason. But the datalog file is there for plotting purposes (except on runs when you didn't set it up, which you sometimes still want to plot). My preference would be to record datalog files by default. I need to reshuffle the log file in some way because I noticed it conflicts with the prob log file for some of the test cases. |
|
Ok, I removed |
|
there's even documentation!! |
6 participants
I made some modifications so the terminal output and datalog files track extrema of important quantities (e.g., species mass fractions) for debugging purposes and thought it might be generally useful. This tracking occurs at the same intervals as tracking integrated quantities but can be turned off independently.
Some notes:
maxDerive, which in the previous implementation ignored coarse cells covered by fine cells (good), but assumed maxima were positive (bad, e.g. for eint). The new implementation does not ignore covered cells, but this should be okay - average down is always called before the tracking, so covered cells should never have more extreme values than the cells covering them.pelec.fuel_name,pelec.ftrac_spec_name, and a new input calledpelec.extrema_spec_nameare also reported if these input parameters are specified.