November 2012+Merge+Notes

Click here to return to top level Jetstream Merge Notes page.

Modifications

 * Using a new formula for the memory usage prediction of ILU factorization. The new formula looks wierd, but it does save a lot memory usage for ILU factorization. This has been tested in a bunch of Euler, Laminor and RANS cases. Please contact Xiaodong if you get some problem with the memory allocation in factorization.
 * The reference quantities in "pyOPTIMA3D.py" now correspond to an altitude of 30,000ft based on the U.S. Standard Atmosphere model.
 * Added diablo% resid_switch to the list of input parameters, it's an absolute tolerance at which the solver switches to the inexact Newton phase. The default value is 1e-5, and is set to a larger value to speed up aero-structural calculations. Otherwise, the default value should be lower than the relative drop tolerance in most cases.
 * Added an optional as_rhs argument to solveFlowAdjoints. If present, it is added to the RHS of the adjoint equation for partitioned aero-structural calculations, otherwise the usual calculation is carried out.
 * Added an aerodynamic adjoint storage psi to Partition_Mod for aero-structural optimization
 * If jtstrm% as_opt is true, then calcdJdG_plus_psidRdG only calculates psidRdG for aero-structural optimization

Bug Fixes

 * Fixed problem with header of //surf_CPs_XXX.dat// files.
 * Fixed problem with wrong nnz (number of non-zero element) and the output in .solv file