Primary debug scenario

Symptom: Many-core system shows escalating latency after 16 cores active. Snoop rate is manageable. Retry rate spikes.

Triage steps

1. Profile retry rate per home node — is one HN overloaded?

2. Check RN-to-HN routing — are requests evenly distributed?

3. Measure directory lookup latency at the hot home node.

4. Check credit pools — is a VC exhausted causing backpressure?

5. Consider home node redistribution or additional HN instances.

Common pitfalls

• Treating CHI as 'AXI with snoops' — different topology, different debug.

• Ignoring retry counters in performance analysis.

• Assuming protocol family alone determines snoop filtering, topology or performance.