Multi-bit Crossing Strategies

What this topic teaches

Multi-bit Crossing Strategies focuses on closing CDC/RDC risk with mechanism-level reasoning. Bundled-data handshakes, FIFOs, source-synchronous strobes, and encoded protocols each trade latency, area, and proof complexity. Senior signoff depends on proving behavior with targeted evidence, not just clearing tool warnings.

The senior-engineer question

When coherency guarantee coverage, data corruption incidents, protocol overhead regresses, can you classify the hazard, identify accountable owners, and choose the smallest fix or waiver backed by evidence?

CDC/RDC SIGNOFF FLOW — Multi-bit Crossing Strategies

crossing inventory + reset map
          |
          v
crossing classification (level/pulse/bus/reset)
          |
          v
structure + protocol + reset checks
          |
          v
critical issues + waiver review
          |
          v
fix / validate / regress / signoff

Picture the crossing behavior

Draw the behavior before touching tools. These visuals are the expected whiteboard baseline for reviews and interviews.

Strategy matrix

bundle+valid/ready  : moderate latency, strong coherency
async FIFO         : high robustness, area cost
source strobe      : throughput friendly, timing contract heavy
token protocol     : explicit ordering, more control logic

Crossing sequence

CROSSING FLOW — Multi-bit Crossing Strategies

source clock domain -> launch signal -> crossing structure -> destination sample
      |                    |                 |                    |
   source FF           protocol           sync / fifo         destination FF

Key metric: coherency guarantee coverage, data corruption incidents, protocol overhead

Ownership layers

CDC/RDC OWNERSHIP LAYERS — Multi-bit Crossing Strategies

layer                 owns                            common failure
------------------    -----------------------------   -----------------------------
design intent         crossing architecture           wrong topology selected
protocol semantics    req/ack, fifo, ordering        liveness/deadlock bugs
reset behavior        assert/deassert sequencing      boot instability
analysis setup        tool rules + waivers            false confidence
signoff governance    risk acceptance + dashboard     stale critical waivers

Evidence to collect

• Primary metric: coherency guarantee coverage, data corruption incidents, protocol overhead.

• Primary artifact: strategy decision matrix, interface timing contract, verification plan.

• Owners to involve: architecture owner, RTL owner, CDC lead.

• At least one reproducer tied to mode/reset/traffic context.

• Decision record: fix, waive, or escalate with rationale.

Ownership map

OWNERSHIP MAP — Multi-bit Crossing Strategies

artifact                  owner
----------------------    -------------------------
design intent           architecture owner
verification evidence   RTL owner
signoff decision        CDC lead

Every open CDC/RDC issue needs one accountable owner before waiver or fix.

Subpages in this topic

Each topic includes mechanism, I/O contract, metrics, debug, worked example, pitfalls, interview drills, checklist, theory, design tradeoffs, expanded case study, walkthrough, comparison matrix, software view, and silicon impact.

Key takeaways

• Classify crossing/reset hazards before proposing fixes.

• Pair structural results with protocol/reset behavioral proof.

• Treat waivers as bounded risk contracts, not cleanup shortcuts.

Common pitfalls

• Mass-waiving warnings near tapeout.

• Assuming local IP cleanliness guarantees SoC behavior.

• Skipping reconvergence and reset stress after CDC fixes.

CDC/RDC deep dive

Reconvergence failures are timing + logic coupling hazards.

Concept diagram

RECONVERGENCE

branch A sync --                 -> merge logic -> illegal combination window
branch B sync --/

Metric graph

HAZARD REPRODUCTION RATE

before fix: 1/2000 runs
after fix: 0/100000 runs

Reports and artifacts

• reconvergence warnings

• gray transition checks

• hazard replay traces

• multi-bit strategy map

Mini case study

Gray pointer was legal, but downstream merge decoded mixed-era values during burst stress.

Debug branches

• Trace fanin cones

• align event windows

• verify decode assumptions

Senior review question

Ask: what evidence proves this risk is closed for silicon, not just tool-clean?

Key takeaways

• State crossing class, assumptions, and owner with every issue.

• Run structural and dynamic regressions after each fix.

Common pitfalls

• Treating all warnings as equivalent risk.

• Waiving issues without containment evidence.

• Skipping reset and reconvergence stress after CDC fixes.