-
Notifications
You must be signed in to change notification settings - Fork 41
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
use new compareScenarios2()
in test instead of old compareScenarios()
#249
Conversation
On my machine, the test now takes 3min. |
Thanks, Christof! May it be worth adding a second fulldata.gdx, maybe a H21 run (if that is possible in a cs2 to have both simultaneously?), or with sector-differentiated carbon prices (would have catched this one), from a coupled run or something like that? So to say, design a second mif which is maximally different from the current one? And have you thought about #172? |
Could the library tests for alternative gdxs be done in parallel? Maybe @giannou already know about that? |
@Renato-Rodrigues: As far as I understand, @chroetz managed to bring the test time down to 3 minutes (convGDX2mif + cs2), so even if we add a second gdx, it would still be a big improvement. |
compareScenarios2()
in test instead of old compareSceanrios()
compareScenarios2()
in test instead of old compareScenarios()
The problem is keep summing up tests above tests without worrying about the total buildLibrary time. If tests can run in parallel, I completely agree with adding as many tests as we can. I am not ok with having a lot of tests added sequentially because we need then to wait half hour or more for the buildLibrary to finsih to be able to commit an urgent model breaking change. |
Yes, I agree. So: If we don't provide a second gdx file, but just a second mif file (that then has to be updated regulary to keep track of the changes in the reporting), this shouldn't extend the runtime of cs2 at all. |
We used to have parallelization in the tests, we can maybe reintroduce it. I think it was removed for reasons that are not relevant any more. |
gdx files are not updated as frequently as mif files. We add or change reporting variables more often than model variables. So I don't think the mif file is a good approach. I think we are going to far from the topic of the PR, maybe we should discuss that in an issue instead. |
|
I suggest to merge that now because it is already a big relief, and discuss further which gdx files could be added, if this was desired. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I took the test and ran it without deleting the file and had a look at it. Seems to work and is much faster than before.
See #238.