Main Page   Namespace List   Class Hierarchy   Compound List   File List   Namespace Members   Compound Members   File Members   Related Pages   Examples  

WishList.html

Go to the documentation of this file.
00001 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
00002 <html>
00003 <HEAD>
00004 <LINK HREF="doxygen.css" REL="stylesheet" TYPE="text/css">
00005 </HEAD>
00006 <BODY BGCOLOR="#d4ebf9">
00007 <IMG SRC="../Images/tBird2D.gif" WIDTH=100 ALIGN="LEFT"/>
00008 <br>
00009 <br><br>
00010 <br><br>
00011 
00012 <h1> Recent additions </h1>
00013 
00014 <table border="4" width="100%">
00015 <tr>
00016 <td> <b> Feature </b> </td>
00017 <td> <b> Motivation </b> </td>
00018 <td> <b> Status </b> </td>
00019 </tr>
00020 
00021 <tr>
00022 <td> Expressions defined on regions </td>
00023 <td> Multi-material problems </td>
00024 <td> Has passed simple tests </td>
00025 </tr>
00026 
00027 <tr>
00028 <td> High-level nonlinear solver objects </td>
00029 <td> Nonlinear problems </td>
00030 <td> Has passed fairly serious tests. 
00031 Needs better user interface for parameter input. Newton 
00032 and Picard solvers have been implemented so far.</td>
00033 </tr>
00034 
00035 <tr>
00036 <td> Parameter expression objects </td>
00037 <td> Design parameters in optimization, continuation parameters
00038 in nonlinear solves </td>
00039 <td> Has passed simple tests. 
00040 Single and listed parameters have been implemented. For problems
00041 with a moderate number (10-100) of design parameters, I'd like to
00042 have a vector of parameters to avoid working with a large symbolic list
00043 at the expression level. This needs to be designed.  </td>
00044 </tr>
00045 
00046 </table>
00047 
00048 <h1>Cool features to add</h1>
00049 
00050 <table border="4" width="100%">
00051 <tr>
00052 <td> <b> Feature </b> </td>
00053 <td> <b> Motivation </b> </td>
00054 <td> <b> Dependencies </b> </td>
00055 <td> <b> Work level </b> </td>
00056 <td> <b> Status </b> </td>
00057 </tr>
00058 
00059 <tr>
00060 <td> Non-affine quad and brick cells </td>
00061 <td> Support for exodus files based on fastq and cubit meshes </td>
00062 <td> None </td>
00063 <td> Moderate </td>
00064 <td> Started. The major bottleneck is the need for to modify the WorkSet
00065 code to handle non-affine cells</td>
00066 </tr>
00067 
00068 <tr>
00069 <td> autoconf </td>
00070 <td> Conform to standard configuration procedure </td>
00071 <td> None </td>
00072 <td> Low </td>
00073 <td> Not begun </td>
00074 </tr>
00075 
00076 <tr>
00077 <td> symmetric quad rules for tets </td>
00078 <td> performance </td>
00079 <td> None </td>
00080 <td> Very low </td>
00081 <td> Designed, waiting to be done </td>
00082 </tr>
00083 
00084 <tr>
00085 <td> symmetric essential BCs </td>
00086 <td> Allows use of CG. Simplifies formulation of eigenvalue problems. </td>
00087 <td> None </td>
00088 <td> Moderate </td>
00089 <td> In design phase </td>
00090 </tr>
00091 
00092 <tr>
00093 <td> Eigenvalue problems </td>
00094 <td> Model reduction, vibrational problems. </td>
00095 <td> Symmetric BCs </td>
00096 <td> Moderate </td>
00097 <td> Not begun </td>
00098 </tr>
00099 
00100 <tr>
00101 <td> Interface to rSQP++ </td>
00102 <td> Optimization </td>
00103 <td> None </td>
00104 <td> Low </td>
00105 <td> Started. Roscoe and I need to be careful about coordination 
00106 of our smart pointers and memory models, but it's otherwise straightforward
00107 as long as we both use Petra. </td>
00108 </tr>
00109 
00110 <tr>
00111 <td> Interface to O3D </td>
00112 <td> Optimization </td>
00113 <td> None </td>
00114 <td> Moderate </td>
00115 <td> In design phase. HCL needs some row access operators </td>
00116 </tr>
00117 
00118 <tr>
00119 <td> Interface to Opt++ </td>
00120 <td> Optimization </td>
00121 <td> None </td>
00122 <td> Moderate </td>
00123 <td> Need either (a) HCL interface to Newmat, or (b) Opt++ access to HCL </td>
00124 </tr>
00125 
00126 <tr>
00127 <td> Interface to NOX </td>
00128 <td> Nonlinear solvers </td>
00129 <td> None </td>
00130 <td> Low </td>
00131 <td> In design phase </td>
00132 </tr>
00133 
00134 <tr>
00135 <td> Interface to PETSC solvers </td>
00136 <td> Access to solvers, compatibility with Omar and George </td>
00137 <td> None </td>
00138 <td> Unknown </td>
00139 <td> Not begun </td>
00140 </tr>
00141 
00142 <tr>
00143 <td> Interface to SUNDIALS </td>
00144 <td> Access to time integrators, access to checkpointing
00145 code for adjoint problems </td>
00146 <td> None </td>
00147 <td> Unknown </td>
00148 <td> Not begun </td>
00149 </tr>
00150 
00151 <tr>
00152 <td> Exodus mesh reader </td>
00153 <td> Sandia "real world" problems </td>
00154 <td> Needs non-affine quads and bricks </td>
00155 <td> Low, but moderate dependencies </td>
00156 <td> Waiting for dependencies </td>
00157 </tr>
00158 
00159 <tr>
00160 <td> Expression objects for functionals </td>
00161 <td> Symbolic objective functions in optimization </td>
00162 <td> None </td>
00163 <td> Moderate </td>
00164 <td> In design phase </td>
00165 </tr>
00166 
00167 <tr>
00168 <td> Solving for unmeshed unknowns </td>
00169 <td> Solving for unmeshed design parameters in optimization. Currently, 
00170 every unknown and test function is assumed to be represented on a 
00171 finite-element mesh. There are ParameterExpr objects for design parameters,
00172 but they can't appear as unknowns. We need to add UnknownParameter expressions.
00173 </td>
00174 <td> Vectors of design parameters </td>
00175 <td> Moderate </td>
00176 <td> In design phase </td>
00177 </tr>
00178 
00179 <tr>
00180 <td> Vectors of design parameters </td>
00181 <td> Avoid symbolic vectors for moderate numbers of design parameters </td>
00182 <td> None </td>
00183 <td> Low </td>
00184 <td> In design phase </td>
00185 </tr>
00186 
00187 <tr>
00188 <td> Integral equation boundary conditions </td>
00189 <td> Enclosure radiation problems, far-field wave BCs </td>
00190 <td> None </td>
00191 <td> High </td>
00192 <td> In research phase. </td>
00193 </tr>
00194 
00195 <tr>
00196 <td> Sierra interface </td>
00197 <td> Making life simpler for Sierra developers </td>
00198 <td> None </td>
00199 <td> High </td>
00200 <td> In research phase. </td>
00201 </tr>
00202 
00203 <tr>
00204 <td> Adaptive mesh refinement </td>
00205 <td> Performance </td>
00206 <td> None </td>
00207 <td> High </td>
00208 <td> In research phase </td>
00209 </tr>
00210 
00211 
00212 <tr>
00213 <td> Aztec plugin </td>
00214 <td> Flexibility and performance </td>
00215 <td> None </td>
00216 <td> Moderate </td>
00217 <td> Not begun </td>
00218 </tr>
00219 
00220 <tr>
00221 <td> ML plugin </td>
00222 <td> Flexibility and performance </td>
00223 <td> Aztec plugin? </td>
00224 <td> Moderate </td>
00225 <td> Not begun </td>
00226 </tr>
00227 
00228 </table>
00229 
00230  </body>
00231 </html>

Contact:
Kevin Long (krlong@ca.sandia.gov)


Documentation generated by