You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Actually, I think this is just the effects of the database file becoming larger than memory, and the deletions being the last part of the benchmark and hence starting from a full database.
I did additional benchmarks each starting with 16M insertions, and then deletions of that data or insertions of more data. The results show that deletions are slightly slower than insertions, but scale as expected:
Delete:
10k:
redb: Removed 10000 items in 5694ms
100k:
redb: Removed 100000 items in 31043ms
1M:
redb: Removed 1000000 items in 166154ms
2M:
redb: Removed 2000000 items in 343368ms
4M:
redb: Removed 4000000 items in 756057ms
8M:
redb: Removed 8000000 items in 1368607ms
16M:
redb: Removed 16000000 items in 1686246ms
Additional inserts (with seed = 37347842):
10k:
redb: Bulk loaded 10000 items in 4158ms
100k:
redb: Bulk loaded 100000 items in 26200ms
1M:
redb: Bulk loaded 1000000 items in 124340ms
2M:
redb: Bulk loaded 2000000 items in 250383ms
4M:
redb: Bulk loaded 4000000 items in 594556ms
8M:
redb: Bulk loaded 8000000 items in 1390448ms
16M:
redb: Bulk loaded 16000000 items in 3648112ms
Deletions should be about the same speed as insertions. However they seem to scale poorly. See the benchmarking done in #344 for more details
The text was updated successfully, but these errors were encountered: