Skip to content

Recent changes double the length of time of runs #301

@andrewbaxter439

Description

@andrewbaxter439

As a quick check of GitHub Actions runs from #271 to #294, the length of time per year seems to more than double on changes in the second PR:

Time to create initial population structures: 29.305 seconds.
Scaling factor is 1363.0347626227203
Time to complete initialisation 1.3268333333333333 minutes.
Starting year 2019
Finished year 2019 (in 25.988 seconds)
It's New Year's Eve of 2019
Starting year 2020
Finished year 2020 (in 35.656 seconds)
...
It's New Year's Eve of 2026
Finished simulating population in 6.100416666666666 minutes

to:

Time to create initial population structures: 30.471 seconds.
Scaling factor is 1363.1462218395511
Time to complete initialisation 1.3090666666666666 minutes.
Starting year 2019
Finished year 2019 (in 82.796 seconds)
It's New Year's Eve of 2019
Starting year 2020
Finished year 2020 (in 93.208 seconds)
...
It's New Year's Eve of 2026
Finished simulating population in 12.776833333333334 minutes

All build and setup times are very similar, only the yearly times are lengthening. This would be a concern in large populations taking 10-20 mins per run, running a longer time. Might be worth reviewing new code in #294 to see if any parts are run in every yearly cycle (including year 1) which might be slowing things down?

https://github.com/centreformicrosimulation/SimPaths/actions/runs/19472726381/job/55724213431
https://github.com/centreformicrosimulation/SimPaths/actions/runs/19646570668/job/56263190107

Metadata

Metadata

Labels

No labels
No labels

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions