When computing a quantities like the criticial curves, they are evaluated using an evaluation_grid which is typically higher resolution than the data.
For example, for a 100 x 100 strong lens image cut-out, this grid will typically be higher resolution, going to 2000 x 2000 pixels.
For large images (e.g. galaxy clusters) this produces grids which increase in shape from ~1000 x 1000 pixels to ~20000 x 20000, which significantly slows the code down.
This PR puts a cap on the shape of the evaluation grid to ensure fast run times.
When computing a quantities like the criticial curves, they are evaluated using an
evaluation_grid
which is typically higher resolution than the data.For example, for a 100 x 100 strong lens image cut-out, this grid will typically be higher resolution, going to 2000 x 2000 pixels.
For large images (e.g. galaxy clusters) this produces grids which increase in shape from ~1000 x 1000 pixels to ~20000 x 20000, which significantly slows the code down.
This PR puts a cap on the shape of the evaluation grid to ensure fast run times.