Table 5.
Average runtimes in milliseconds for random graphs with 100, 1000, and 10 000 vertices, for the Recursive Dijkstra's, Edge-List Bisection, and Edge-Elimination algorithm. Runtimes were calculated on a single core of an 8 core Linux Intel Xeon X5460 processor with clock speed of 3.16 GHz and 16 GB memory shared among 8 cores. Runtimes were not calculated for the Edge-Elimination algorithm for 10 000 vertices since the estimated runtime was too long. Also shown is the number of edges for each size of random graphs.
| Dense graphs | ||||
|---|---|---|---|---|
| Graph size (nodes) | Number of edges | Recursive Dijkstra's | Edge-List Bisection | Edge Elimination |
| 100 | 9500 | 0.31 | 2.93 | 258.61 |
| 1000 | 600 000 | 98.70 | 404.06 | 1.56 × 106 |
| 10 000 |
60 000 000 |
49 045.3 |
117 248.62 |
… |
| Sparse graphs | ||||
| Graph size (nodes) |
Number of edges |
Recursive Dijkstra's |
Edge-List Bisection |
Edge Elimination |
| 100 | 1000 | 0.23 | 0.54 | 13.83 |
| 1000 | 10 000 | 47.81 | 36.18 | 7554.45 |
| 10 000 | 100 000 | 11 188.2 | 7228.17 | … |