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
Currently in run_by_unit.sh, it has HUC processing duration system, where each run_unit.sh calculates duration for a huc and updates a shared file in the logs folder. It currently looks like this:
However.. As multiple processes all try to write to the same file, colisons occur and the data is always somewhat incomplete and I need more performance data to find bottlenecks
I would like to replace / upgrade this system to:
have run_by_unit.sh continue to track the duration time for the entire huc (including branches), but now also add new parameters for the total number of branches and the total processing time for the branches as a whole. This would include the processing time for both branch zero and its other branches as one number.
the number of branches including branch zero.
put that value a new output text file in the HUC output directory itself.
Add a new column to this output which is the HUC number, careful on zero padding. (make this the first column).
At the start of post processing, create a new tool which simply collects all of the HUCs files and concatenates them.
It should continue to be a csv output and continue with a semi-weird pattern for durations. It currently shows the entire huc processing time in two columns. One is total duration as min, second in normal time format. ie) 2:55. The second duration is a good duration in 10 base (percent), ie: 2.91 (two decimals). That second column makes it easier for averaging and summing later. Ensure this two column pattern for duration is applied to the branch durations columns
Nice to have but optional, but very nice to have.
In post processing, when we concatenate all independent HUC duration files, add new rows for:
total number of hucs processed
sum of the total overall huc processing time
sum of branch processing time.
sum of the number of branches processed.
The text was updated successfully, but these errors were encountered:
Currently in
run_by_unit.sh
, it has HUC processing duration system, where each run_unit.sh calculates duration for a huc and updates a shared file in the logs folder. It currently looks like this:However.. As multiple processes all try to write to the same file, colisons occur and the data is always somewhat incomplete and I need more performance data to find bottlenecks
I would like to replace / upgrade this system to:
run_by_unit.sh
continue to track the duration time for the entire huc (including branches), but now also add new parameters for the total number of branches and the total processing time for the branches as a whole. This would include the processing time for both branch zero and its other branches as one number.It should continue to be a csv output and continue with a semi-weird pattern for durations. It currently shows the entire huc processing time in two columns. One is total duration as min, second in normal time format. ie) 2:55. The second duration is a good duration in 10 base (percent), ie: 2.91 (two decimals). That second column makes it easier for averaging and summing later. Ensure this two column pattern for duration is applied to the branch durations columns
Nice to have but optional, but very nice to have.
The text was updated successfully, but these errors were encountered: