Scenario: DDR Bandwidth Collapse

Scenario

A new traffic mix increases refresh stalls, row misses, and read/write turnarounds. The memory controller owner asks for more bus width.

OBSERVED METRIC
effective bandwidth drops from 42 GB/s to 21 GB/s on a video workload

45-MINUTE INTERVIEW FLOW
0-5: clarify protocol, layer, and traffic setup
5-15: draw transaction flow and ownership
15-25: isolate first bad transaction or stalled progress
25-35: propose reduced experiment and fix options
35-45: state compliance, workload, and regression plan

Common pitfalls

• Quote peak bandwidth

• Ignore bank locality

• Miss firmware QoS configuration

Scenario debrief

Score the answer: layer, transaction ID, metric, artifact, regression.

Reference stack

intent -> transaction -> channel -> physical

Metric pattern

baseline -> experiment -> target

Debrief prompts

1. Which protocol layer did you name first?

2. Which transaction identity did you carry through the story?

3. What reduced experiment did you propose?

4. What compliance and product regressions did you list?

Key takeaways

• Connect every protocol claim to a transaction identity and measurable metric.

• Store the artifact (waveform, log, counter) next to every signoff decision.

Common pitfalls

• Debugging timeouts without finding the first bad transaction.

• Quoting peak bus width without payload efficiency and retry overhead.

• Treating VIP compliance as a substitute for system integration replay.