Test+Problem+LOG

=LOG of changes to the Test Problem Utilities= Pages: Workshop, Overview, CADAC, Data Access, Work Area Setup, Path Setup, Utilities, Grid Data, SPH Data, Running, Reporting, **LOG**

Please note that times below are given in UT (Universal Time).


 * 27.nov 2007, 21:40 UT (13:40 KITP):** Corrected a bug in **grid2cores.f90**; when reading an MHD file it was reading the velocity field into the magnetic field slots.


 * 22. nov. 2007, 21:00 UT:** Activated OpenMP options with **-nosave** option -- this was necessary to avoid an OpenMP trap that did not occur earlier (e.g. exectables produced nov. 11).


 * 22. nov. 2007, 08:20 UT:** Memory allocation corrected in **g2g.x** (source code grid2grid.f90) -- affecting only MHD-conversion cases (the bug was causing a memory segmentation error). Versions of **g2g.x** from before 12. nov 2007, ~7:30 UT (revisions of grid2grid.f90 up to and including 1.18) were functioning correctly.


 * 21. nov. 2007, 20:00 UT:** Added output of statistics (INPUT and OUTPUT) to the **g2g.x** executable. Corrected some inconsistencies in the example **.g2g** files in the ICS directory.


 * 11. nov. 2007 22:20 UT;** Added an option to **g2g.x** to read in data for sink particles and add their mass back to the gas -- for visualization or diagnostic purposes. This is triggered by adding **do_sink=t** to the input namelist, and supplying a .txt file with data for the sink particles. For an example, see the execution of **g2g.x** (controlled by **test/g2g.in)** that takes place in the **test/** subdirectory when doing **make tests**. The size and scaling of the voxel regions where mass is added is controlled by the **size** and **power** parameters in the **&sink ... /** namelist.


 * 10. nov. 2007 14:20 UT;** Added SPH data file **1e9.sph** to the **data** directory. This file has ~1e9 SHP particles, with 'precision placement' of particles (see below). Previous versions of **g2s.x** failed for 1e9 particles (or required a larger-than-default UNIX stack size), because of the presence of a superfluous scratch variable.


 * 10. nov. 2007 13:00 UT;** Installed new **s2g.x**, based on rev. 1.11 of **sph2grid.f90**. This version distributes the mass and momentum of each particle over all cells within a local smoothing length (based on the local density, which is assumed to be available in the input SPH data file). This generally results in the generation of smoother and better grid data from SPH data files than with the previous, very rudimentary revisions of sph2grid.f90.


 * 9. nov. 2007 19:20 UT;** Installed new **g2s.x**, based on revision 1.6 of sph_positions.f90, with 'precision placement' of all particles in cells (in the original 1000^3 data set) with more than 10 particles. The probability density of particles is uniform over cells with less than 10 particles, and follows a tri-linear interpolation of density for cells with more than 10 particles. SPH files generated with g2s.x before this time (or copies of the utility code taken before this time) have uniform probability density over cells.


 * 5. nov. 2007 21:20 UT;** Changed sign convention in interpolate.f90, which is used by **g2g.x**. Regridded files produced before this time (or produced with copies of the utility code taken before this time) may have velocities shifted the wrong way (depending on input parameters). The previous version of **g2g.x** produced correct results if used with &in .. offset=0.5 ... / for Stagger Code input (e.g. the original initial value data files), while the updated version requires &in .. offset=-0.5 ... /. An offset of -0.5 now means that vector component data are centered on left hand side cube faces.