A chip has many failing paths and finite time. Path groups organise them, summaries rank them, and a priority rule decides the order of work. This topic covers grouping, reading per-group results, choosing what to fix first, and tracking progress so that improvement in one group is not hidden by regression in another.
Lesson | After it you can... |
|---|---|
Explain how paths are grouped and why it matters. | |
Read WNS, TNS and counts per group. | |
Rank work by impact, not by worst number alone. | |
Track and compare runs per group. |
One very bad path can hide that the rest of the design is nearly clean, or a small WNS can hide thousands of failing endpoints. Per-group numbers expose where the work is.