Comments (6)
Hi,
Yes it does support non-uniform spacing and we might support in the future. Problems come when we try to create a transpose operator. The whole function (stencils) changes and we don't know how to implement it without a matrix.
from diffeqoperators.jl.
It's fine to have no transpose in that case. non-uniform with the BC handling is super useful by itself, and many iterative solvers don't need the transpose.
from diffeqoperators.jl.
Upon further thought, we can get the transposed stencils very easily actually (except for :periodic
maybe)
We can generate the 'essential' full matrix linear operator ie. one which is much smaller is size (O((derivative_order+approximation_order)^2)
) and since it contains the essential information about all the coefficients, we can take it's transpose and get the transposed stencils. eg.
julia> A = LinearOperator{Float64}(2,4,1.0,100,:D0,:D0);
julia> full(A,5)
5×5 Array{Float64,2}:
-2.5 1.33333 -0.0833333 0.0 0.0
1.33333 -2.5 1.33333 -0.0833333 0.0
-0.0833333 1.33333 -2.5 1.33333 -0.0833333
0.0 -0.0833333 1.33333 -2.5 1.33333
0.0 0.0 -0.0833333 1.33333 -2.5
full(A,5)'
, in this case the same matrix, will have the transposed stencils which we can store for fast calculation of transposes.
Using A'
will just meaning changing a flag variable in the LinearOperator to start using the transposed stencils created at initialization.
from diffeqoperators.jl.
Don't worry about this until way way later. Non-uniform spacing with an iterative method that requires a transpose is a very niche topic.
from diffeqoperators.jl.
For the moment, anyone who needs the transpose can just use the full matrix operator, so the functionality is still there, just not optimal. With that caveat, supporting non-uniform spacing is actually very straightforward.
from diffeqoperators.jl.
work going on here #49
from diffeqoperators.jl.
Related Issues (20)
- Performance issues with nonlinear_diffusion! HOT 4
- Fix symbolic arrays test to allow MTK v5.21.0 HOT 7
- UndefVarError: dereference not defined when precompiling DiffEqOperators (Win10, julia 1.6.1-2) HOT 7
- Version Compatibility? HOT 1
- Error defining second-order differential HOT 8
- The latest versions of LoopVectorization (0.12.80-81) are not compatible and cause errors HOT 1
- Noob Help: Simple, Linear, Inhomogenous PDE HOT 3
- concretization of High Dimension PDE HOT 2
- LoopVectorization break HOT 5
- Unexpected concretization of Laplace operator HOT 3
- v4.35.0 failed to release HOT 2
- Feature request: Support Summation-By-Parts operators HOT 4
- Support Unitful HOT 6
- Incorrect boundary padded vector with composed PeriodicBC on 2d data HOT 6
- Implementing Drift-Diffusion model in higher dimensions HOT 3
- Uninitialized field `opnorm` in JacVecOperator HOT 4
- SplitODEProblem broken for MatrixFreeOperator HOT 3
- ERROR: LoadError: Some tests did not pass: 30 passed, 0 failed, 4 errored, 4 broken. HOT 18
- Warning related to "LoopVectorization.check_args"
- Lots of precompilation errors and noise HOT 7
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from diffeqoperators.jl.