Rust BV's progress reporting undercounts arcs
seemingly constant regardless of the graph size:
- 2023-09-06: 517,399,308,984 arcs in Java, 517,397,741,568 in Rust
- 2023-09-06-popular-1k: 11,322,432,687 arcs in Java, 11,320,852,480 in Rust
- 2021-03-23-popular-3k-python: 1,221,283,907 arcs in Java, 1,219,690,496 in Rust
Java values come from https://docs.softwareheritage.org/devel/swh-dataset/graph/dataset.html ; Rust values come from the BV log before sorting arcs, eg.
2024-03-22T16:37:47+00:00 - INFO - Flushing remaining buffers to BatchIterator...
2024-03-22T16:38:01+00:00 - INFO - Done sorting all buffers.
2024-03-22T16:38:10+00:00 - INFO - Completed.
2024-03-22T16:38:10+00:00 - INFO - Elapsed: 4h 43m 40s [517,397,741,568 arcs, 30398380.80 arcs/s, 32.90 ns/arc]; used/avail/free/total mem 176.21GB/1.57TB/1.57TB/4.33TB
2024-03-22T16:38:10+00:00 - INFO - Building BVGraph