Day 2 results: prediction vs reality
Measured on Apple Silicon laptop, macOS, Python 3, estimate.py.
Scoreboard
Paste the scoreboard and “two numbers to carry” output here.
One sentence per prediction
High or low, by how much, and why I now think that happened.
- P1 db size:
- P2 bulk insert:
- P3 durable rows/sec:
- P4 index size:
- P5 scan:
- P6 indexed lookup:
The fsync number
Real fsync (fullfsync=1): __ µs per commit Plain fsync (fullfsync=0): __ µs per commit What the gap between them means: