Open SeanBryan51 opened 1 year ago
This is also related to how the flux site experiments are structured. Currently, the user can only run sites from the five site experiment, forty two site experiment or a subset of sites from the five site experiment. There is a use case where the user will want to run any specific site that is part of the test suite.
This is also related to how the analysis scripts in the benchcab-evaluation workspace on modelevaluation.org are designed. Ideally we should have a single analysis script that can run an an analysis for any number of sites that are a part of the test suite. The user should be able to specify which sites they want to test.
Hey team! Please add your planning poker estimate with Zenhub @ccarouge @SeanBryan51
We should add support for debugging CABLE with gdb, ddt, totalview, etc so that CABLE developers can easily investigate bugs or errors at runtime.
For example,
pytest --trace
launches a debugger for each test case. We could have a similar option which launches an interactive debugging session for each configuration we are running.