Column: Silicon systems design


Figure 3: System-level observability across multi-die systems (Source: IEEE)


driven validation, cross-die observability, power-aware analysis and end-to- end protocol checking all contribute information that traditional metrics can’t capture on their own. Te goal is not to replace existing measures, but to contextualise them within a broader understanding of system behaviour. Figure 3 shows the system-level access


and observability architecture across multiple chiplets, highlighting that visibility into emergent behaviour depends on standards-based access paths rather than local coverage metrics alone. Artificial intelligence is increasingly


used to assist with this task, particularly when data volumes exceed the capacity of manual analysis. Pattern recognition across regressions, anomaly clustering and cross-domain correlation can highlight areas where high coverage masks fragile behaviour. Crucially, these tools support engineering judgement rather than attempting to replace it.


or removed, to manage simulation performance. What remains visible is oſten insufficient to diagnose subtle integration effects. Metrics continue to report success, even as visibility into true system behaviour diminishes. Tis is why experienced engineers oſten express unease late in a program despite strong metric reports. Teir intuition reflects an understanding that confidence can’t be inferred solely from coverage graphs.


Extending metrics without abandoning them Te limitations of verification metrics do not imply that they should be discarded. Tey remain essential for managing complexity and ensuring discipline. What must change is how they are interpreted and complemented. System-level verification demands


additional forms of evidence. Workload- 14 April 2026 www.electronicsworld.co.uk


Teams are forced to confront an uncomfortable reality: the metrics were not wrong, but they were incomplete


A shift in how confidence is earned Ultimately, verification metrics are tools, not guarantees. In complex systems, confidence is earned through visibility into interaction, not through the accumulation of local success indicators. Programs that recognise this early tend to invest in observability, standards-based access and system-level thinking from the outset. For verification engineers, the challenge


is not to generate more metrics, but to ask better questions of the ones we already have. When metrics stop correlating with confidence, the answer is rarely to push them harder. It is to look beyond them. As systems continue to scale,


understanding the limits of verification metrics will become as important as mastering their use. In that transition, engineers who treat metrics as evidence rather than assurance will be better positioned to deliver robust, trustworthy silicon.


Tis column continues in next month’s edition of Electronics World.


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40  |  Page 41  |  Page 42  |  Page 43  |  Page 44  |  Page 45  |  Page 46  |  Page 47  |  Page 48