Tidy benchmarking
Closed this issue · 1 comments
ElliottKasoar commented
A general pass of the repo is needed to tidy up, potentially including:
- Ensuring all unused variables are removed/used
- Add functioning assertions to all benchmarks
- Remove code that is no longer needed (#18?)
- Refactor shared code (e.g. between assert functions in #22)
- Ensure changes in FTorch are reflected in code here
- Restructure code into separate folders, as mentioned in #20
TomMelt commented
when we restructure code into folders, the following source files need to go together.
cg_drag_forpy.f90
cg_drag_torch.f90
benchmarker_forpy.f90
benchmarker_torch.f90
These test the structure of gravity wave drag routines in MIMA. Possibly rename module files to make it clear they are modules.