Values <1.0E-100 included in SAVE file with improper formatting.
When running with tough3-eos3, we noticed that there were some values for air mass fraction (X2) that were below 1.0E-100, but the "E" was not included. Here is a screenshot of what I'm talking about:
Has anyone seen this before? We were able to continue running the simulation using this file as INCON, so TOUGH is reading something from it, but I don't think this was the intention of the code design.
1 reply
-
Hi Mikey,
Yes, I've seen it. It's standard FORTRAN (maybe compiler- or compiler-option dependent). Therefore, it is not really an issue, i.e., FOTRAN writes it out in this format and correctly reads it in, so no problem reading SAVE files as INCON files. But it's of course "ugly" and may cause problems if you use non-FORTRAN post-processors. Two "solutions" (none really satisfying):
(1) Change the FORTRAN FORMAT statement for reals from, e.g., E20.12 to E20.12E3, which writes three-digit exponents for all number and adds the "E". It's a practical solution in the code section where the SAVE file is written
(2) Post process. Under unix, use, e.g.
sed "s/\([0-9]\)-.../\1E-99/g" SAVE > INCON
or any other corresponding global string replacement command.
Good luck!
Stefan