sqpeWhat it contributes
Reorders table rows and fields to increase shared prompt prefixes and improve KV-cache reuse during batch model inference.
How to read the evidence
Reported cost savings depend on model-serving behavior and provider pricing; this catalog does not repeat them as current estimates.
We checked the primary publication record and abstract. This profile summarizes author-reported work; it does not claim a full-paper audit, implementation audit, or independent reproduction.
Explore the connections
- Kalypso — Query-plan serving API
- Databricks — Databricks SQL