Skip to content

Latest commit

 

History

History
90 lines (65 loc) · 4.01 KB

readme.md

File metadata and controls

90 lines (65 loc) · 4.01 KB
Raw

Netlist

Circuit netlist generation & parsing

Netlist is a Python package for reading, manipulating, and writing popular circuit netlist formats. Particularly emphasis is placed on the device model and technology library constructs of netlist formats.

The Netlist AST

The data model used throughout the netlist package is defined in data.py. The primary top-level entity is a Program - a term which Spice-culture generally lacks, but means "a collection of input SourceFiles for simulation".

Netlist Dialects

"Spice" format is not a terribly well-defined term. Netlists come in a variety of dialects, each generally specific to a particular simulation program. These dialects represent the same conceptual content - circuits, instances of sub-circuits, models, parameters, and the like - but do so with varying syntax and semantics.

The netlist package is designed to support the most popular dialects. Enumerated, supported dialects are listed by the netlist.NetlistDialects enumeration. Support as of this writing includes:

Dialect Parsing Writing
"Generic" Spice* ✅
Hspice ✅
Ngspice
CDL
Xyce ✅
Spectre ✅
Spectre-Spice* ✅

Notes:

  • While "generic spice" format is something of an oxymoron, a base-case set of rules covers many common constructs, and should cover most syntax for an unlisted simulation program.
  • "Spectre-Spice" is the spice-dialect supported by Cadence's Spectre. It is largely identical to "generic spice", plus the capacity for simulator lang changes into spectre-syntax.

Supported Constructs

Netlist focuses on the netlist-constructs most commonly used for device models and technology libraries, including:

  • Circuit hierarchy via SubcktDef sub-circuit-definitions, Instances thereof, and Primitive instances
  • ModelDef statements for defining device models
    • Binned models, generally separated by physical quantities such as {lmin,lmax}, are grouped into ModelFamilys and ModelVariants.
  • Parameter declarations (ParamDecl) and values (ParamVal)
  • Mathematical expressions (Expr) including
    • UnaryOp, BinaryOp, and TernaryOp among parameters
    • Funcion Calls
  • Limited FunctionDef definitions, in supporting dialects
  • Options

Common unsupported constructs include most used for simulation stimulus-generation: analyses, measurements, probes, and similar.

Parsing

The primary entry point for parsing an existing netlist is netlist.parse. This function takes a Path or PathLike input, and returns a netlist Program.

from netlist import parse
program = parse("mymodels.sp")

The parse function accepts an optional dialect argument, which sets the initial spice-dialect for parsing. If not provided

Writing

The primary entrypoint for writing a netlist Program is netlist.netlist (as a verb). The netlist function accepts a source Program and an open IO object - commonly an open file-handle - as a destination.

from netlist import write
write(program, open("mymodels.scs", "w"), fmt=NetlistDialects.SPECTRE)

The netlist function accepts an optional dialect argument, which sets the dialect for writing.

Conversions

Real-life netlist-programs can be long - at times millions of text-lines - and can be broken up into a series of smaller files. For tasks not requiring any in-memory analysis or manipulation, but strict conversion from one format to another, netlist includes a convert function which maps potentially multi-file input to a directory-full of output.

from netlist import convert
convert(...)