Parallel multifit fails for some set of input parameters

[abuts]

The ticket Re #1721 fixes crystal_alignment for case when parallel_multifit is enabled by disabling parallel mutlifit option in crystal alignment.

This is probably fine for crystal_alignment as it does not need to run in parallel. But similar issue may occur for different multifit user cases so splitting procedure in multifit should be smarter to identify such parameters and either do correct split or (better) be smart enough to identify that parallel multifit in this case is unnecessary and multifit should not run in parallel.

Unit test written and disabled in Re #1721: Horace\_test\test_multifit_herbert\test_multifit_1.m, row 362 should be fixed.