update example05a to include reference to Myy caching

Author: TmsRC

examples/example05a - nonlinear_and_linear_evolution_with_GS.ipynb

@@ -217,6 +217,7 @@
     "- `min_dIy_dI`: threshold value below which passive structure normal modes are dropped. Modes with norm(d(Iy)/dI)<`min_dIy_dI` are dropped, which filters out modes that do not actually couple with the plasma.\n",
     "- `max_mode_frequency`: threshold value for characteristic frequencies above which passive structure normal modes are dropped (i.e. the fast modes).\n",
     "- **`n_linearization_workers`**: number of independent worker processes used to build the current-response (`dIydI`) and profile-parameter-response (`dIydtheta`) Jacobian columns. The default, `1`, retains serial construction. Values larger than one parallelise both the initial linearisation and later relinearisations without limiting the inner threading of ordinary time-step GS solves.\n",
+    "- `cache_myy`: determines whether the Myy matrix should be calculated once and saved, or re-calculated as needed. The default is `True`, setting to `False` reduces memory usage significantly, allowing for more resolved meshes to be used. Note that not caching Myy can have serious performance impact when evolving the plasma non-linearly (i.e. when passing `linear_only=False` to the `nlstepper`); for linear evolution no noticeable impact should be expected.\n",
     "\n",
     "Other customisable inputs are available, do see the documentation or the later notebook on \"Growth Rates\" for more details. For example, one may explicitly set your own resistance and inductance matrices for the tokamak, rather than the geometrical value calculated internally in FreeGSNKE.\n",
     "\n",
@@ -241,6 +242,7 @@
     "    plasma_resistivity=1e-6,\n",
     "    max_mode_frequency=10**2.5,\n",
     "    n_linearization_workers=1,  # set >1 to parallelise Jacobian construction\n",
+    "    cache_myy=True,\n",
     ")"
    ]
   },