I noticed a drastic solve time difference between running a neohookean model with the multiscale utility, or the singlescale utility. This was because I was using a LU solver for the multiscale utility, whereas the singlescale application was not reading the petsc_options file which was causing petsc to default to the GMRES solver.
I noticed a drastic solve time difference between running a neohookean model with the multiscale utility, or the singlescale utility. This was because I was using a LU solver for the multiscale utility, whereas the singlescale application was not reading the petsc_options file which was causing petsc to default to the GMRES solver.