Computer Architecture · All levels

Routing and Flow Control — Interview Drills

Interview Drills for Routing and Flow Control (NoC and Interconnect Architecture).

Interview drills

Practice aloud for NoC and Interconnect Architecture → Routing and Flow Control. Use METRIC → HYPOTHESIS → FIX → REGRESSION.

Why does deadlock freedom not guarantee good latency?

diagram
[INT][ARCH][TOPIC]

Q: Why does deadlock freedom not guarantee good latency?

A:
Deadlock freedom only proves eventual progress. You still need fairness and bounded queue growth to control tail latency.

FOLLOW-UP TRAP: Equating deadlock-safe with performance-safe.

When should you prefer deterministic routing over adaptive?

diagram
[INT][ARCH][TOPIC]

Q: When should you prefer deterministic routing over adaptive?

A:
When workload paths are predictable, verification budget is tight, and deterministic routes avoid instability from noisy congestion signals.

FOLLOW-UP TRAP: Assuming adaptive routing is always superior.

What metric best indicates credit starvation?

diagram
[INT][ARCH][TOPIC]

Q: What metric best indicates credit starvation?

A:
Rising credit return latency combined with stable offered load and localized queue growth.

FOLLOW-UP TRAP: Using only average buffer occupancy.

10+ year interview answer bar

At senior/principal level, the interviewer is testing ownership judgment more than vocabulary. Answer Routing and Flow Control through failure mode, evidence, tradeoff, and release decision.

You inherit a late-stage Routing and Flow Control failure one week before release. What do you do in the first hour?

diagram
[INT][ARCH][STAFF]

Q: You inherit a late-stage Routing and Flow Control failure one week before release. What do you do in the first hour?

A:
Freeze the workload/model/RTL tag, name the failing metric (Routing/flow-control stress verification report), confirm counter setup, cluster the issue by structure or workload phase, assign the first experiment, and publish a validation/owner plan before changing architecture.

FOLLOW-UP TRAP: Jumping directly to a larger cache, wider pipe, or extra NoC link without preserving evidence.

When would you stop trying to improve Routing and Flow Control and escalate?

diagram
[INT][ARCH][STAFF]

Q: When would you stop trying to improve Routing and Flow Control and escalate?

A:
Escalate when the remaining risk crosses ownership boundaries, consumes shared margin, changes signed-off assumptions, or threatens QoS guarantees, coherency response latency, and software-visible jitter.. Bring exact report lines and options, not vague concern.

FOLLOW-UP TRAP: Escalating without data or continuing alone after a cross-team decision is needed.

Whiteboard diagram to draw

diagram
VISUAL MODEL — NoC and Interconnect Architecture / Routing and Flow Control

        workload / trace
              │
              ▼
   metric symptom (Routing/flow-control stress verification report)
              │
              ▼
     likely microarchitectural mechanism
              │
      ┌───────┼────────┐
      ▼       ▼        ▼
  pipeline  memory    fabric/coherency
  stalls    misses    queues / ordering
      │       │        │
      └───────┼────────┘
              ▼
        bounded design change
              │
              ▼
   validation workload + PPA regression

Architecture deep dive

NoC is a queueing system — bandwidth, latency, and deadlock are coupled.

Concept diagram

diagram
NoC TOPOLOGY SKETCH

CPU0 ──┐      ┌── LLC0 ── DRAM0
       R0 ─── R1
CPU1 ──┘      │
              R2 ─── R3 ── GPU/DMA
              │      │
             NPU    LLC1 ── DRAM1

Look for: hot links, cyclic dependencies, VC starvation, and tail latency.

Metric graph

diagram
LATENCY DISTRIBUTION

p50    ██████  32 ns
p90    ████████████  71 ns
p99    ████████████████████████  210 ns
p99.9  █████████████████████████████████  480 ns

Averages hide QoS failures.

Metrics and artifacts

  • link utilization

  • average latency by master

  • retry/backpressure counts

  • QoS violation log

Mini case study

Average latency looks fine but tail latency spikes for CPU coherent reads when GPU DMA runs. QoS and separate VCs fix the starvation without doubling link width.

Debug branches

  • If deadlock, check credit loops and routing restrictions first.

  • If latency tail long, inspect arbitration and buffer depth.

Senior review question

Ask: what single metric would prove this concept is working or failing on your workload?

Key takeaways

  • Connect every architecture claim to a workload and measurable metric.

  • State verification and PPA impact before proposing design changes.

Common pitfalls

  • Feature-driven design without MPKI/IPC/bandwidth evidence.

  • Ignoring coherency and NoC traffic in cache and accelerator sizing.

Study notes

Re-read this topic with one concrete workload.