# MOSFET, FinFET, GAAFET, and CFET architecture comparison replay

- Version: 2026-08-22-v91
- Experiment: SA-DEV-GAAFET-COMPARISON-002
- Public routes: /simulations/mosfet-finfet-gaafet-cfet-architecture-comparison-replay, /learn/mosfet-finfet-gaafet-cfet-architecture-comparison, /labs/mosfet-finfet-gaafet-cfet-architecture-comparison
- Dataset: mosfet-finfet-gaafet-cfet-architecture-comparison-replay-v91.csv
- Rows: 4096
- Execution: precomputed browser CSV replay only

## Purpose

This replay teaches architecture comparison as a contextual evidence problem. It keeps electrostatic control, drive per footprint, contact and capacitance burden, variability, thermal coupling, integration risk, evidence maturity, and circuit handoff together instead of declaring a universal winner.

## What Is Included

- Original synthetic rows across architecture, scaling context, contact strategy, variability mode, circuit context, and review posture.
- Normalized scores for electrostatic control, drive per footprint, capacitance burden, contact-resistance burden, variability sensitivity, thermal coupling, integration risk, evidence maturity, circuit handoff, and comparison readiness.
- Decision gates for architecture-context, evidence, electrostatic, parasitic, variability, thermal, integration, circuit-handoff, owner-review, bounded-teaching, and bounded-comparison states.
- Link-only source spine from official public research, roadmap, and metrology pages.

## What Is Not Included

- No live TCAD, SPICE, compact-model fitting, extraction, electrothermal, process, or optimizer execution.
- No PDK, compact model, deck, process flow, layout, extraction technology file, standard-cell library, measured silicon, benchmark, roadmap table, or copied source figure.
- No architecture, foundry node, vendor, product, cost, yield, readiness, signoff, qualification, or production ranking.

## Replacement Evidence Needed

Engineering reuse needs owned calibrated device models, I-V/C-V and short-channel metrics, exact geometry and process context, variability distributions and metrology, contact resistance and local-RC extraction, self-heating and thermal-boundary evidence, circuit-level correlation, tool/version provenance, and device/integration/extraction/circuit owner approval.