Repository navigation

rqe-optimizer: let mergeable finer-group deployments serve coarser RAQEs - #190

Merged

zzylol merged 13 commits into

mainfrom

187-rqe-optimizer-let-mergeable-finer-group-deployments-serve-coarser-raqes

Oct 10, 2026

Commits on Oct 10, 2026
- rqe-optimizer: let mergeable finer-group deployments serve coarser RAQEs
  
  authored and
  zzylol
  committed
  79d3983
- docs: split RQE optimizer design by concern
  
  milindsrivastava1997
  authored and
  zzylol
  committed
  8e255d5
- docs: clarify RQE optimizer cost model status
  
  milindsrivastava1997
  authored and
  zzylol
  committed
  bbdb066
- docs: one cost table for every phase, with roll-ups
  
  zzylol
  and
  claude
  committed
  067c65c
- docs: KLL's memory is sketch-bench's nominal footprint, not measured at 1e6
  
  zzylol
  and
  claude
  committed
  b0cb5a7
- rqe-optimizer: prune on every cost cost-by-use bills; fix stale comments
  
  zzylol
  and
  claude
  committed
  5eac9b9
- docs: consolidate the optimizer docs; add the Hydra admission design
  
  zzylol
  and
  claude
  committed
  3995c78
- rqe-optimizer: univmon-cardinality merges lossily; increase's real reason
  
  zzylol
  and
  claude
  committed
  f6ce9b7
- rqe-optimizer: check group counts only between label sets a plan uses
  
  zzylol
  and
  claude
  committed
  d289109
- docs(hydra): hydra-kll is measured in the same Hydra rerun
  
  zzylol
  and
  claude
  committed
  25b34b3
- docs: move the Hydra admission design out of this PR
  
  zzylol
  and
  claude
  committed
  c02a7de
- rqe-optimizer: tests for roll-up DD memory, facts and pruning; doc fixes
  
  zzylol
  and
  claude
  committed
  6b1637a
- rqe-optimizer: name every lossy non-heap sketch that needs merge curves
  
  zzylol
  and
  claude
  committed
  d288a5e

Navigation Menu

Repository navigation

rqe-optimizer: let mergeable finer-group deployments serve coarser RAQEs - #190

rqe-optimizer: let mergeable finer-group deployments serve coarser RAQEs#190
zzylol merged 13 commits into
mainfrom
187-rqe-optimizer-let-mergeable-finer-group-deployments-serve-coarser-raqes

Commits on Oct 10, 2026

rqe-optimizer: let mergeable finer-group deployments serve coarser RAQEs

docs: split RQE optimizer design by concern

docs: clarify RQE optimizer cost model status

docs: one cost table for every phase, with roll-ups

docs: KLL's memory is sketch-bench's nominal footprint, not measured at 1e6

rqe-optimizer: prune on every cost cost-by-use bills; fix stale comments

docs: consolidate the optimizer docs; add the Hydra admission design

rqe-optimizer: univmon-cardinality merges lossily; increase's real reason

rqe-optimizer: check group counts only between label sets a plan uses

docs(hydra): hydra-kll is measured in the same Hydra rerun

docs: move the Hydra admission design out of this PR

rqe-optimizer: tests for roll-up DD memory, facts and pruning; doc fixes

rqe-optimizer: name every lossy non-heap sketch that needs merge curves