Hello all,

I have implemented a VUMAT that is similar to the built-in *Brittle Cracking material model. Before element deletion, it has a short damage phase with a directional dependency for the crack. For this reason, no  less than 14 dependent variables (*depvar) are being used.

The calculation results are satisfactory, but I noticed the calculation time using the VUMAT has increased by a factor 10 compared to the same simulation using the *Brittle Cracking.
The VUMAT is now used as a fortran 77 '.for' file, compiled and linked for each analysis. 

Ideas what may have caused this reduction of calculation speed?
Tips or tricks?

Thanks!

joren