Closed Xiangyu-Hu closed 10 months ago
Probably, running for less physical time?
If we want to shorten the simulation time in CI, the modifications must include lower particle resolution, shorter channel and shorter physical time. I have done it, tomorrow I will check results and re-build regression since time sep changed.
If 1000 seconds in CI cannot be achieved with this new configuration I suggest to de-activate the case from running in CI.
With this configuration takes (circa) 1600s on linux, 2400s on windows and 1400s on macOS.
It's better to deactivate the case for CI or remove it. In the following weeks I shall prepare a different configuration for testing purposes.
OK. You can replace the 3d case with a new one. Thanks.
Modified case. Now it should take from 250 to 300 seconds in CI.
Good. I think that it is good enough.
We can replace the old one with this.
Is already substituted and running in #358. Is not as fast as I hoped, in linux takes circa 600s. If you need faster I have to think another different configuration.
The nonlinear wave interaction case in 3d takes too long time. It takes about 2500 seconds in CI. @ncsalis Could it possible to shorten the time to about 1000 seconds?