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
{{ message }}
This repository has been archived by the owner on Jan 12, 2022. It is now read-only.
Implement various ideas for building a more clever Rete structure:
ASAP trimming of variables not used later
Clever node reuse
Find if a stub join also happens to join alternate stubs with the same underlying Rete node, and then consider the "free accidental" stub in further joins.
Suggested order of join preference:
trivial joins (stub a = stub b)
high selectivity filters
supressed by joincandidate if already checked on other side
apply first to narrower stubs
aggregator quasijoins, let-expressions (?)
true joins
except cartesians
TBD: ranking based on what constraints the join will enable
low-selectivity filters (injectivity)
cartesian joins
The text was updated successfully, but these errors were encountered:
Added the task of clever node reuse after the discussions at the latest meeting. Potential source of good test case: the thesis about GPU vs. EIQ performance supervised by @abelhegedus.
Implement various ideas for building a more clever Rete structure:
The text was updated successfully, but these errors were encountered: