You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I think even if #112 is somehow resolved (eg by using KernelAbstractions --- a good solution for that issue still isn't clear to me), we probably still want to implement a special optimized method for two layer case.
Currently we implement "optimized" functions for extracting PV from the QG streamfunction, and vice versa, in the single layer case:
GeophysicalFlows.jl/src/multilayerqg.jl
Lines 485 to 495 in f4c68a5
The inversion is also trivial in the two layer case, so we should support that.
To implement this I propose generalize
SingleLayerParams
toNLayerParams{N, ...}
, whereN
is the number of layers. Then we have aliasesand we can implement optimized inversion routines for these two cases.
cc @szy21
The text was updated successfully, but these errors were encountered: