← Back to team overview

dolfin team mailing list archive

[WIKI] DOLFIN_road_map

 

The following page has been updated:

  http://www.fenics.org/wiki/DOLFIN_road_map

The changes are as follows.

***  

---  

***************

*** 19,34 ****

       * [15]1 Version 1.0
       * [16]2 Version 0.9.4
       * [17]3 Version 0.9.3
!      * [18]4 Other planned features
!           + [19]4.1 Demos
!           + [20]4.2 Function
!           + [21]4.3 General
!           + [22]4.4 Linear Algebra
!           + [23]4.5 Mesh
!           + [24]4.6 ODE solvers
!           + [25]4.7 Parallelization
!           + [26]4.8 Parameter system
!           + [27]4.9 User solvers
  
     [[28]edit]
  
--- 19,34 ----

       * [15]1 Version 1.0
       * [16]2 Version 0.9.4
       * [17]3 Version 0.9.3
!      * [18]4 Parallelisation
!      * [19]5 Other planned features
!           + [20]5.1 Demos
!           + [21]5.2 Function
!           + [22]5.3 General
!           + [23]5.4 Linear Algebra
!           + [24]5.5 Mesh
!           + [25]5.6 ODE solvers
!           + [26]5.7 Parameter system
!           + [27]5.8 User solvers
  
     [[28]edit]
  
***************

*** 86,94 ****

  
     [[34]edit]
  
  Other planned features
  
!    [[35]edit]
  
  Demos
  
--- 86,116 ----

  
     [[34]edit]
  
+ Parallelisation
+ 
+      * [DEL: Mesh parsing :DEL]
+      * [DEL: Partitioning (ParMETIS) :DEL]
+      * [DEL: Mesh distribution :DEL]
+      * [DEL: Coefficients/interpolation :DEL]
+      * [DEL: Mesh entity numbering :DEL]
+      * [DEL: DOF renumbering :DEL]
+      * [DEL: SparsityPattern :DEL]
+      * Test cases / unit tests (see sandbox/demo.py)
+      * Partition and distribution of built-in meshes
+      * DG (facet assembly)
+      * Benchmarking
+      * [DEL: VTK output :DEL]
+      * Viper visualisation
+      * [DEL: Dirichlet BC :DEL]
+      * Neumann BC (should work, needs to be tested)
+      * Get all demos to run in parallel
+      * More advanced renumbering
+ 
+    [[35]edit]
+ 
  Other planned features
  
!    [[36]edit]
  
  Demos
  
***************

*** 100,119 ****

            + This section may also relate/link to the higher order mesh
              stuff.
  
!    [[36]edit]
  
  Function
  
       * Direct support for tensor valued functions
       * Input/output for time-series
  
!    [[37]edit]
  
  General
  
       * Remove as much const_cast as possible
  
!    [[38]edit]
  
  Linear Algebra
  
--- 122,141 ----

            + This section may also relate/link to the higher order mesh
              stuff.
  
!    [[37]edit]
  
  Function
  
       * Direct support for tensor valued functions
       * Input/output for time-series
  
!    [[38]edit]
  
  General
  
       * Remove as much const_cast as possible
  
!    [[39]edit]
  
  Linear Algebra
  
***************

*** 121,127 ****

         and call it if necessary (GNW: I don't think that we should do
         this)
  
!    [[39]edit]
  
  Mesh
  
--- 143,149 ----

         and call it if necessary (GNW: I don't think that we should do
         this)
  
!    [[40]edit]
  
  Mesh
  
***************

*** 130,136 ****

       * Support for non-affine cells
       * Rewrite Point class with public x, y, z member and use the trick
         described here to allow indexed access:
!        [40]http://www.mr-edd.co.uk/?p=113
       * Higher Order (Curved) Mesh Capability
            + Continue to get P2 (quadratic) 2-D triangles implemented.
            + This will require modifying ufc::cell and related routines so
--- 152,158 ----

       * Support for non-affine cells
       * Rewrite Point class with public x, y, z member and use the trick
         described here to allow indexed access:
!        [41]http://www.mr-edd.co.uk/?p=113
       * Higher Order (Curved) Mesh Capability
            + Continue to get P2 (quadratic) 2-D triangles implemented.
            + This will require modifying ufc::cell and related routines so
***************

*** 139,166 ****

              NOT affinely mapped.
            + Have a demo for this.
  
!    [[41]edit]
  
  ODE solvers
  
       * Solve dual and compute error estimate
       * Automatic computation of stability factors as function of time T
- 
-    [[42]edit]
- 
- Parallelization
- 
-      * Modify Function to work with distributed vectors (for parallel
-        assembly)
-      * Refine linear algebra classes for parallel assembly and solve
-      * I/O
-      * Store both global and local indices for cell/vertex, add mesh
-        functions for this
-      * Adapt BoundaryComputation for parallel mesh, use knowledge about
-        which vertices are shared to decide if a facet is a global or
-        local boundary
-      * Use parallel mesh information in DofMap to simplify implementation
-        - should only need map from local to global cell/vertex
  
     [[43]edit]
  
--- 161,172 ----

              NOT affinely mapped.
            + Have a demo for this.
  
!    [[42]edit]
  
  ODE solvers
  
       * Solve dual and compute error estimate
       * Automatic computation of stability factors as function of time T
  
     [[43]edit]