I have had my student to run a benchmark of this code.
While the papers claim that CH provides several orders of magnitude larger performance (like 1000x), here the improvement seems to be only 10%. Do you have any idea why this could be happening?
using BenchmarkTools, OpenStreetMapX
# I downloaded just the repo
include("ContractionHierarchies.jl/src/ContractionHierarchies.jl")
hmap = ContractionHierarchies.HierarchicalMap()
ContractionHierarchies.add_map(hmap,"europe","lithuania")
vanila_alg = get_map_data(raw"ContractionHierarchies.jl\data\europe\lithuania\local.osm.pbf");
local_src = ContractionHierarchies.local_node(hmap, LLA(56.029,21.182))
local_dst = ContractionHierarchies.local_node(hmap, LLA(54.085, 24.631))
I have had my student to run a benchmark of this code. While the papers claim that CH provides several orders of magnitude larger performance (like 1000x), here the improvement seems to be only 10%. Do you have any idea why this could be happening?
The benchmarks look like this:
This definitely not expected (this is a long route from one edge to another of Lithuania). Do you have any idea why this is happening?