Add support for Hilbert packing (bulk load)

I am dealing with quite large data set (~22M) and this data is static in nature. While I do not care much on insert performance, I would like to improve RTree structure resulting into search efficiencies. For this, I would like to introduce Hilbert packing in parallel to STR packing which was recently put in place. Based upon data compositions, shape and distribution one packing algo. may perform better than the other. More importantly, its quite straight forward.