The memory footprint of the cached MLI_SHA_2E model is too large.

[pacesm]

The memory footprint of the cached MLI_SHA_2E (max. degree 1319) model is too large.

This is not an issue with one working processes, but with multiple processes the memory usage goes quickly out of hand.

The model caching needs to be redesigned to allow large models such as MLI_SHA_2E.