Crate libreda_sta

source ·
Expand description

Incremental graph-based static timing analysis (STA) for the LibrEDA framework.

Timing analysis is performed on netlist data structures that implement the [trait@NetlistBase] trait. This makes the STA algorithm easily portable.

The concept of timing, delays and constraints is abstracted by a set of traits. This architecture should allow to implement simple timing models such as the non-linear delay model (NDLM) and more complicated models (such as statistical models) in a consistent way.

Modules

  • Propagate forward and backward cones from a set of frontier nodes. The cones mark the regions where the timing must potentially be updated.
  • graphviz 🔒
    Dump the timing graph in the graphviz ‘dot’ format.
  • Graph levelization based on an operator formulation. Compute the level of all graph nodes in a directed acyclic graph. The level of a node is the minimum distance to a node without incoming edges with each edge having a distance of 1.
  • Public modules Interface definitions for cell delay look-up as it is provided for example by Liberty libraries.
  • Helper functions to interpret the liberty library which comes from the parser.
  • Model implementations (cell delay and constraints, interconnect delays).
  • Graph levelization based on an operator formulation. Compute the level of all graph nodes in a directed acyclic graph. The level of a node is the minimum distance to a node without incoming edges with each edge having a distance of 1.
  • Construct timing graphs (delay graph, constraint graph) from a netlist.
  • Traits for abstracting timing models.

Structs

Enums

  • Public exports. Type of a timing constraint.
  • Public exports. Type of a signal edge. Either rising (0->1), falling (1->0).
  • Public exports. Type of a signal edge. Either rising (0->1), falling (1->0) or both of them.
  • Error during static timing analysis. This includes errors in the liberty library, mismatches between netlist and library or invalid netlists (drive conflicts, etc).
  • Analysis mode.

Constants

Traits

Functions

  • Compute total net capacitances for each net of the top circuit. Simply sums up the input capacitances of all connected input pins for each net.