build1 distinct publisher
StoneKV's "765,846 writes per second" benchmark turned out to be a lie
The author of StoneKV already calls that figure a lie in his own headline. The write path he documents explains it, since a completed File::sync_all() cannot fit inside the per-write budget the rate implies.
Publishers:dev.to
Reality
- Evidence46
- Adoption
- Insufficient
- Hype gap−20
- Incentives55