I was thinking of a step or brep file but maybe I should consider a compound. The idea was to test if one could perform the import by starting at the lowest directory levels something that would display quicker. For Alice this appears to work but takes over a day to run. Now in a test repository have some python that creates from a gdml file a directory structure where each volume is a sub directory. Mine for example are angular deflection 15° and deviation 1%Ĭode: Select all CERN-Alice-MultiFile-gdml % grep ' VariCAD 2023 v2.08 free

Then finetune the values, till you are satisfied with the time difference of the operations and the appearance of the representation of the tessellation. This means, that now the tessellation time gets higher then the operations itself.

Repeat this step, till the recomputing time increases significant. Then change the values finer of the tessellation, recompute and measure the time.

Then made an array with a few thousands shapes, with planes and curved geometry (note, no chamfers or fillets, the stale on very large arrays recomputes pretty exact to 0.1s per chamfers or fillet, which is slower as they actual are, do not know why). But it can be, that it is the biggest part of the time.įirst, set up the tessellation pretty rough. Also the tessellation is only a part of loading a step file or of an operation in general. (Note, 0.1%? or 0.1 equals 10%? please be explicit!) Changing only one value is almost meaningless. The tessellation are two values, deviation from shape and the angular deflection of the tessellation.