GeneL
commented
5 months ago CORRECTION! The program did not freeze, it just took much longer to execute. This is the timing I got on Windows 10:
Elapsed time to load edges 2199 : 1.55ms Elapsed time to freeze: 2.00ms Elapsed time to prepare: 235.42ms Elapsed time to get path: 235.87ms ShortestPath { source: 388, target: 99, weight: 72, nodes: [388, 392, 428, 879, 888, 889, 99] }
Elapsed time to load edges 2499 : 1.59ms Elapsed time to freeze: 2.07ms Elapsed time to prepare: 3.52s Elapsed time to get path: 3.52s ShortestPath { source: 388, target: 99, weight: 72, nodes: [388, 392, 428, 879, 888, 889, 99] }
Elapsed time to load edges 2999 : 1.58ms Elapsed time to freeze: 2.10ms Elapsed time to prepare: 50.74s Elapsed time to get path: 50.74s ShortestPath { source: 388, target: 99, weight: 36, nodes: [388, 392, 393, 99] }
Elapsed time to load edges 3499 : 1.62ms Elapsed time to freeze: 2.22ms Elapsed time to prepare: 274.39s Elapsed time to get path: 274.39s ShortestPath { source: 388, target: 99, weight: 36, nodes: [388, 387, 398, 99] }
Elapsed time to load edges 3540 : 1.59ms Elapsed time to freeze: 2.13ms Elapsed time to prepare: 267.99s Elapsed time to get path: 267.99s ShortestPath { source: 388, target: 99, weight: 36, nodes: [388, 387, 393, 99] }
Is this normal to see such a time difference with addition of a few hundred edges? Are there ways to speed the preparation up?
dabreegster
Your library works great for me when I use a set of edges with under 2200 records. When I try it with a larger set of edges the application hangs.. Is there a way to configure the app for a larger data set? Is there a different commands or function calls to use a larger data set? Thank you in advance.
Below is my source code:
let mut input_graph = InputGraph::new(); let json = read_json_file("disc.json".to_string()); let mut cnt = 1; for l1 in json.unwrap().len() { if cnt < 2200 { println!(" s: {:?} t: {:?}", l1["source"].as_u64().unwrap(), l1["target"].as_u64().unwrap()); input_graph.add_edge(l1["source"].as_u64().unwrap() as NodeId, l1["target"].as_u64().unwrap() as NodeId, 12); } cnt += 1; } input_graph.freeze(); let fast_graph = fast_paths::prepare(&input_graph); let shortest_path = fast_paths::calc_path(&fast_graph, 388, 99); println!("{:?}", shortest_path.clone().unwrap());