mr-fang-vlsi/bscost-net

BS Cost Net

Use this skill to move model work from single-design intuition to reproducible, multi-design evidence.

Background Knowledge Links

This skill should stay linked to:

1. KB workflow and comparison-policy context for backside cost-model evaluation.
2. Local paper-derived evidence or benchmark notes when choosing benchmark classes and publishable evidence framing.
3. eda-context-accessor when current KB or tool context needs to be refreshed before benchmark or evaluation design.

If the planned benchmark set or evaluation claim conflicts with current KB or paper-derived background knowledge, keep that conflict explicit and route KB feedback instead of silently proceeding.

Mandatory artifacts

1. contract.md/tsv and bucketed.md/tsv
2. design list + source links + commits
3. one summary stating whether model is stably better than HPWL

Operational References

1. Load references/background-knowledge-links.md when deciding which KB notes, paper-derived artifacts, and benchmark assumptions are authoritative for the current study.
2. Load references/benchmark_candidates.md when selecting the benchmark panel or deciding whether the current panel is too narrow.
3. Load references/web_research_protocol.md when network-backed benchmark expansion or provenance checking is required.
4. Load references/model_plan.md when designing signal-net and clock-net modeling tracks and aligned output fields.
5. Load references/evaluation-protocol.md when setting Gate-0/Gate-1 evaluation structure, HPWL comparison, or stability-panel reporting.
6. Load references/promotion-gate.md when deciding whether the model remains in shadow mode or is eligible for active optimization.