netlist_carpentry.routines.opt.constant_folds

A collection of constant folding algorithms.

Functions:

• opt_constant –

  Executes several optimization routines on the given module.

• opt_constant_mux_inputs –

  Optimizes multiplexers, where both inputs are constant, by replacing them with the appropriate constant signal.

• opt_constant_propagation –

  Executes constant propagation to simplify the circuit.

opt_constant

opt_constant(module: Module) -> bool

Executes several optimization routines on the given module.

These routines currently include constant propagation and constant multiplexer replacement. More passes may follow in the future

Parameters:

• module

  (Module) –

  The module to be optimized.

Returns:

• bool ( bool ) –

  True if any optimizations were executed, False otherwise.

opt_constant_mux_inputs

opt_constant_mux_inputs(module: Module) -> bool

Optimizes multiplexers, where both inputs are constant, by replacing them with the appropriate constant signal.

If both inputs are constant and equal, the output is equal to the constant input. If both inputs are constant and unequal (i.e. one is 0 and one is 1), the output is either equal to S or !S, depending on which input is 0 and which is 1. If S is constant, the instance can be removed, since the output follows the corresponding input signal.

Parameters:

• module

  (Module) –

  The module in which constant multiplexers should be optimized.

Returns:

• bool ( bool ) –

  True if any optimizations were executed. False if this module is already optimized in regards to constant multiplexers.

opt_constant_propagation

opt_constant_propagation(module: Module) -> bool

Executes constant propagation to simplify the circuit.

Constant propagation replaces expressions with known constant values. By substituting constants it simplifies expressions, exposes dead instances and other optimization opportunities, and can reduce circuit size. For example, if the framework knows A = 0 and later sees B = A && 1, it can replace B with 0 and eliminate the original assignment, since this expression can never be 1, as A is known to be 0, and 0 && x is always 0.

Parameters:

• module

  (Module) –

  The module to perform constant propagation in

Returns:

• bool ( bool ) –

  True if any optimizations were executed. False if this module is already optimized in regards to constant propagation.