Dynamic Decile grouping performance issues

From what I saw in this article, it relies on constant segmentation which is hard-coded in the segmentation tables. I need my segmentation information to also update dynamically (i.e. the upper & lower bounds of each segment should change based on the filtered dataset, as the upper & lower bounds are based on percentiles of the filtered dataset).

What I really need is a calculated column that updates upon filtering in order to calculate the upper & lower bounds, but that's not available. What I wrote uses a variable in the measure to calculate the upper & lower bound during visualization, but the performance is very poor. Below is the perfomance analyzer screenshot using hard-coded segmentation, using measures to only calculate the upper & lower bounds, or using a variable to store the upper/lower bounds and calculate segmentations from that.

While the method works for small datasets (only 10K records in the above), I haven't gotten it to work on larger datasets.