How efabless/globalfoundries-pdk-libs-gf180mcu_fd_pr works
Core components
• cells/klayout/pymacros – Python “PCells” (parameterized cells) that generate GDS layouts
• cells/xschem – Schematic symbol definitions for Xschem
• models – Raw SPICE measurement Excel data, cleaned datasets, and extraction scripts
• models/ngspice & models/xyce – Spice/Xyce netlists and regression‐test suites
• rules/klayout – DRC/LVS rule macros for KLayout
• tech/klayout – KLayout technology files (layer mappings, styles)
• third_party/make-env – Environment‐setup scripts (Conda, Make, CI)
• Makefile / CI workflows – Orchestrator for builds, tests, regression runs
Relationships & data flows
a) “Raw Data” (Excel in models/180MCU_SPICE_DATA) feeds
→ Cleaning Scripts (models_clean/scripts) →
→ “Clean Data” (CSV in models_clean/gf180mcu_data)
b) Clean Data + PCell definitions (cells/klayout/pymacros)
→ Generate GDS PCells → loaded via tech/klayout into KLayout
c) Clean Data + Symbol definitions (cells/xschem)
→ Generate Xschem symbol library → used in schematic capture
d) Clean Data + SPICE Netlists (models/ngspice & models/xyce)
→ Regression Scripts (models/ngspice/testing & models/xyce/testing)
→ External Simulators (Ngspice, Xyce)
→ Test Results
e) KLayout PCells + tech files + DRC/LVS macros
→ Layout verification runs via rules/klayout macros
f) Makefile and GitHub Actions (/.github/workflows) coordinate all of the above
Architectural patterns
– Modular plugin‐style: each subdirectory is a self‐contained plugin for an EDA tool
– Data pipeline: raw measurement → clean data → device macros & netlists
– Test harness: Python pytest for PCells, Spice regression suites for models
– CI orchestration: Makefile targets + GitHub Actions for repeatable builds/tests
Essential elements
– Clear component labels (e.g., “PCells Python Macros”, “Clean SPICE Data”, “DRC/LVS Rules”)
• Code modules
• Data stores
• External EDA tools