gate_length

Purely geometric (no connectivity needed). The minimum gate length — the poly width measured only where the poly actually forms a device gate of a given type — as opposed to min_width, which would measure GatPoly’s width everywhere, and instead of the channel width W (set by the perpendicular Activ edges), which must never be mistaken for the length L.

Semantics

Runs the same facing-wall width scan as min_width (see Architecture’s tiled-merge section), but masked: a measured width is only kept if its midpoint lies inside layers[1], a derived “gate region” layer such as GatPolyOverNsdActivNoTGO (GatPoly, intersected with N+ diffusion, outside a thick-gate- oxide device). Measuring the poly directly — not the Activ-clipped channel — gives the true gate length; different device flavours (N/P, thin/thick oxide) get their own rule instance with their own mask layer and, typically, their own minimum length.

Layers

Positional, exactly two:

1. The poly layer whose width is measured (GatPoly).

2. A derived mask layer selecting which parts of the poly count as this device type’s gate (e.g. GatPolyOverPsdActivTGO for a P-type thick-gate-oxide device). Only widths whose midpoint falls inside this mask are checked.

Parameters

None beyond the standard value (the minimum length, µm).

Violation markers

One edge marker per side of every under-length gate span found (both walls of the narrow poly, like min_width), but only where the span’s midpoint lies inside the mask layer.

KLayout equivalent

Corresponds to a width check on the poly layer, restricted (and/masked) to the per-device-type gate-region derived layer — not a single built-in KLayout primitive, but the same idiom KLayout DRC decks use to separate gate length from channel width.

Example

Four device flavours, each with its own gate-region mask and minimum length:

- id: Gat.a1
  check: gate_length
  layers: [GatPoly, GatPolyOverNsdActivNoTGO]
  value: 0.13
- id: Gat.a3
  check: gate_length
  layers: [GatPoly, GatPolyOverNsdActivTGO]
  value: 0.45