Work it through
Coverage without sensitivity
A test can hit the stall coverage bin even when its checker is disconnected or disabled.
Record scenario coverage and demonstrated checker sensitivity as separate evidence.
Your next page is on its way.
System verification strategy
Classify the subsystem, anchor the plan to its real standards, and assign each risk to the cheapest trustworthy verification engine.

A worked mechanism
Illustrative requirement: a stalled output must preserve payload until acceptance.
VALID=1; READY=0
The test creates a sustained stall, including the last beat.
DATA, LAST and qualifiers sampled
The monitor observes the interface independently of stimulus intent.
change DATA during the stall
A negative fixture must cause the stability assertion to fail.
Work it through
A test can hit the stall coverage bin even when its checker is disconnected or disabled.
Record scenario coverage and demonstrated checker sensitivity as separate evidence.
A passing run or a covered bin alone does not establish correctness. Formal proofs also depend on assumptions and scope.
For each high-risk requirement, identify the activating test, observation point, checker and a known violation it detects.
Select a subsystem. Inspect its verification responsibilities and the complementary evidence needed to support them.
Match completions against accepted requests by destination, ID and flow instead of one global FIFO. Allow independent flows to interleave, while recording any missing, duplicate or wrongly ordered completion within a required ordering domain.
AXI interconnect · NoC · Cache-coherent fabric
Source × destination × ID × burst · Legal route × congestion point · Outstanding depth × buffer · QoS age × arbitration outcome
Primary-source accuracy pass
Factory, events, barriers, components, TLM, RAL, phasing, and other APIs must be interpreted against the active UVM standard and its documented implementation deviations.
Arm · AMBA AXI/ACESame-ID ordering depends on channel and destination; AXI does not generally order different IDs, different masters, or reads against writes. The scoreboard must model the actual ordering domain.
AMD · CDC methodologyDynamic phase stress can expose logic assumptions, but it does not simulate analog metastability. Structural review, recognized synchronizers, implementation constraints, and MTBF analysis remain essential.
Accellera · IEEE 1801-2024Isolation, supplies, and state retention belong in the power intent model. A power-aware UVM monitor complements UPF-aware verification; it cannot prove physical isolation-cell behavior.
RISC-V · RVWMOA global shadow memory is not enough by itself. The checker must encode preserved program order, dependencies, fences, atomicity, and the values a load is legally allowed to observe.
OpenTitan · production DVThe testplan, functional coverage plan, DV architecture, independent checkers, regressions, waivers, and dashboards form one evidence chain—not separate completion percentages.
Layered evidence
The design risk chooses the engine; the engine does not choose the requirement.
Choose the dominant risk
AXI interconnect · NoC · Cache-coherent fabric
DDR controller · Multi-bank SRAM controller · L2 cache
DMA · AI accelerator · PCIe or NVMe data mover
ACE or CHI · CPU LSU · Coherent accelerator
Power controller · Interrupt controller · Reset or clock manager