Mercurial > repos > public > sbplib_julia
comparison Notes.md @ 890:eb03bda76bae feature/variable_derivatives
Add some ideas for how to handle communication of regions and region sizes
| author | Jonatan Werpers <jonatan@werpers.com> |
|---|---|
| date | Tue, 08 Feb 2022 10:58:14 +0100 |
| parents | fe8fe3f01162 |
| children | d8da3a1060b7 |
comparison
equal
deleted
inserted
replaced
| 889:069e58fb3829 | 890:eb03bda76bae |
|---|---|
| 152 - [ ] Use a trait to indicate if a TensorMapping uses indices with regions. | 152 - [ ] Use a trait to indicate if a TensorMapping uses indices with regions. |
| 153 The default should be that they do NOT. | 153 The default should be that they do NOT. |
| 154 - [ ] What to name this trait? Can we call it IndexStyle but not export it to avoid conflicts with Base.IndexStyle? | 154 - [ ] What to name this trait? Can we call it IndexStyle but not export it to avoid conflicts with Base.IndexStyle? |
| 155 - [ ] Figure out repeated application of regioned TensorMappings. Maybe an instance of a tensor mapping needs to know the exact size of the range and domain for this to work? | 155 - [ ] Figure out repeated application of regioned TensorMappings. Maybe an instance of a tensor mapping needs to know the exact size of the range and domain for this to work? |
| 156 | 156 |
| 157 ### Ideas for information sharing functions | |
| 158 ```julia | |
| 159 using StaticArrays | |
| 160 | |
| 161 function regions(op::SecondDerivativeVariable) | |
| 162 t = ntuple(i->(Interior(),),range_dim(op)) | |
| 163 return Base.setindex(t, (Lower(), Interior(), Upper()), derivative_direction(op)) | |
| 164 end | |
| 165 | |
| 166 function regionsizes(op::SecondDerivativeVariable) | |
| 167 sz = tuple.(range_size(op)) | |
| 168 | |
| 169 cl = closuresize(op) | |
| 170 return Base.setindex(sz, (cl, n-2cl, cl), derivative_direction(op)) | |
| 171 end | |
| 172 | |
| 173 | |
| 174 g = EquidistantGrid((11,9), (0.,0.), (10.,8.)) # h = 1 | |
| 175 c = evalOn(g, (x,y)->x+y) | |
| 176 | |
| 177 D₂ᶜ = SecondDerivativeVariable(g, c, interior_stencil, closure_stencils,1) | |
| 178 @test regions(D₂ᶜ) == ( | |
| 179 (Lower(), Interior(), Upper()), | |
| 180 (Interior(),), | |
| 181 ) | |
| 182 @test regionsizes(D₂ᶜ) == ((1,9,1),(9,)) | |
| 183 | |
| 184 | |
| 185 D₂ᶜ = SecondDerivativeVariable(g, c, interior_stencil, closure_stencils,2) | |
| 186 @test regions(D₂ᶜ) == ( | |
| 187 (Interior(),), | |
| 188 (Lower(), Interior(), Upper()), | |
| 189 ) | |
| 190 @test regionsizes(D₂ᶜ) == ((11,),(1,7,1)) | |
| 191 ``` | |
| 192 | |
| 193 | |
| 157 ## Boundschecking and dimension checking | 194 ## Boundschecking and dimension checking |
| 158 Does it make sense to have boundschecking only in getindex methods? | 195 Does it make sense to have boundschecking only in getindex methods? |
| 159 This would mean no bounds checking in applys, however any indexing that they do would be boundschecked. The only loss would be readability of errors. But users aren't really supposed to call apply directly anyway. | 196 This would mean no bounds checking in applys, however any indexing that they do would be boundschecked. The only loss would be readability of errors. But users aren't really supposed to call apply directly anyway. |
| 160 | 197 |
| 161 Preferably dimensions and sizes should be checked when lazy objects are created, for example TensorApplication, TensorComposition and so on. If dimension checks decreases performance we can make them skippable later. | 198 Preferably dimensions and sizes should be checked when lazy objects are created, for example TensorApplication, TensorComposition and so on. If dimension checks decreases performance we can make them skippable later. |