Dont-Touch & Preserve Policy: Theory Deep Dive

Foundational theory

Dont-Touch & Preserve Policy is a core part of Constraints & QoR. dont_touch and preserve are surgical controls; overuse blocks optimization and underuse breaks intended structures such as CDC wrappers or DFT hooks. Senior synthesis engineers connect every QoR claim to constraint context, compile setup, and reproducible artifacts.

Core concepts explained

• dont_touch and preserve are surgical controls; overuse blocks optimization and underuse breaks intended structures such as CDC wrappers or DFT hooks.

• Primary metric: preserved object count, blocked optimization opportunities, ECO rework

• Primary artifact: constraint audit, preserve list, optimization-block report

• Owners: synthesis lead, RTL owner, DFT owner

• Compare timing, area, and power together

• Preserve run manifest for every regression jump

Why this matters at closure

At tapeout pace, Dont-Touch & Preserve Policy decisions can either shorten closure loops or create hidden debt. QoR quality starts with clean synthesis constraints and policy hygiene.

Mental model

OBJECT MARKED preserve/dont_touch?
        |
  functional or signoff reason?
     /                \
   yes                 no
   |                   |
keep + audit       remove, re-compile

Worked intuition

1. Freeze RTL tag, constraint tag, and compile switches.

2. Open preserved object count, blocked optimization opportunities, ECO rework and isolate the first meaningful regression.

3. Classify whether issue is constraints, mapping transform, or physical estimate.

4. Collect constraint audit, preserve list, optimization-block report and owner signoff evidence.

5. Pick minimal reversible fix and define rollback criteria.

6. Run timing + power + area regression matrix before merge.

Common misconceptions

• Better WNS always means better overall QoR.

• dont_touch is harmless if timing still passes.

• Retiming gain is free and always safe for formal.

• Topographical estimates are equivalent to signoff route outcomes.

Visual reinforcement

Preserve policy decision tree

OBJECT MARKED preserve/dont_touch?
        |
  functional or signoff reason?
     /                \
   yes                 no
   |                   |
keep + audit       remove, re-compile

Layer responsibilities

SYNTHESIS OWNERSHIP LAYERS — Dont-Touch & Preserve Policy

layer               owns                          failure mode
----------------    ---------------------------   -------------------------
constraints         clocks/exceptions/policy      fake QoR optimism
mapping             cell choices/structure        depth/fanout regressions
optimization        timing/power tradeoffs        one-metric overfitting
physical-aware      topo/congestion estimates     handoff delta surprises
closure             ECO order/regression          fixes break other corners

Synthesis deep dive

Constraint hygiene determines whether synthesis QoR is trustworthy.

Concept diagram

CONSTRAINT FLOW

RTL intent -> synthesis SDC -> compile checks -> QoR dashboard -> review gate

Metric graph

QOR HEALTH

constraint lint pass   ███████████████
unconstrained count    ███
policy violations      ██

Reports and artifacts

• compile QoR

• unconstrained endpoints

• constraint lint

• policy audit log

Mini case study

Nightly WNS improved but unconstrained paths jumped; strict SDC lint restored true QoR picture.

Debug branches

• Diff SDC first

• Audit dont_touch scope

• Track unconstrained trend weekly

Senior review question

Ask: what evidence proves this QoR move is real and stable?

Key takeaways

• State exact run context (RTL, SDC, libs, switches) with every QoR claim.

• Re-run timing, area, and power regressions after each synthesis ECO.

Common pitfalls

• Comparing runs with mismatched constraints or library views.

• Timing-only fixes that violate power or area budgets.

• Skipping equivalence checks after structural changes.

Theory reinforcement

QoR quality starts with clean synthesis constraints and policy hygiene.