>> I have determined that the file to which these descriptors >> point is the one that is being passed to PETSc's "-save_to" >> option. > > As Satish says, this is a PISM thing, and has nothing to do with PETSc. > PISM is using NetCDF directly here, but nc_close() should be called > after writing those snapshots. It would be best if you could post any > additional info to this bug report > > https://gna.org/bugs/index.php?15074 > > Jed
Interesting; thanks for the info. Exactly the sort of insight I was hoping for when posting. When I initially went looking in the code to see where the filename, that I knew was in the argument list to the top level application, PISM, was being used, cd pism-stable0.2.1 find . -type f -print | xargs grep save_to I found ierr = PetscOptionsGetString(PETSC_NULL, "-save_to", snapshots_filename, ... and so assumed it was being passed through into the PETSc "layer". I did not get a chance to dig much further, hence posting on here where there is a lot of knowledge. I do appreciate that we were trying to get the whole software stack going on a platform here at VUW, that's not the most widely used, indeed, the researchers who were hoping to make use of the facility have decided to direct their efforts elsewehere for now. I obviously have the code to play around with but have to claim to being a mere implementer as opposed to someone who knows the software stack, and the way the application should be used, in any great detail, however, if I do get a chance to scratch the itch that this has started, (and I hope to) then I'll feed back what I can, as and when I can. Thanks once again for directing my investogation to the right areas, Kevin -- Kevin M. Buckley Room: CO327 School of Engineering and Phone: +64 4 463 5971 Computer Science Victoria University of Wellington New Zealand
