I have a structural model that includes contact analysis. I noticed a strange behavior in the results directory size reported in the .rpt files. For instance, in a series of runs where the only parameter changed was the plotting grid, the size reported by Windows 10 Explorer matched reasonably well with size reported in .rpt until the plotting grid hit 7. Then things diverged wildly. My initial guess is that at some point individual file sizes (in bytes) are being put into a 32 bit signed integer variable and anything beyond 2^31 rolls over. This creates error that is then carried forward into the directory size calculations.
