huanglangwen
huanglangwen
Thank you! But setting `missingValue` to the grib file has a lot of overheads as we have 100GB+ grib files and there's no way to set it inplace. I'm wondering...
Is there any bound check in debug mode? Because having size signatures always give me an illusion that out of bound indexing are checked.
We could insert runtime bound check assertion to every indexing node when debug flag is set. Or bound check indices whenever the sdfg is called with new values of symbols.
I can give it a try. Can I have some hints where to get started?
No, because `vec(x)` keeps `SArray`. https://github.com/JuliaDiffEq/SparseDiffTools.jl/blob/aaaa9048de136cf222967d3f4c5d13315cbf9208/src/differentiation/compute_jacobian_ad.jl#L102
> Does this at least not break anything that currently exists? It break at least one test in OrdinaryDiffEq.jl: https://github.com/JuliaDiffEq/OrdinaryDiffEq.jl/blob/master/test/interface/data_array_test.jl
But someone may prefer their own `f_p` `g_p` and `g_u` compared to AD-generated one for reasons like analytical solutions, foreign language calls, etc. For the non-parameters, perhaps we need a...
For example, `f_p` is needed when doing back-integration in adjoint sensitivity. I can't find anyplace to specify them.
Oh, I didn't notice that. How about `g_u` and `g_p` because ` adjoint_sensitivities` only got `dg`?
Sorry, I can't get your point. How does the adjoint sensitivity equation go to your conclusion that only dg and g are needed?