InferWidths
Related Docs:
object InferWidths
| package passes
Perform the transform, encode renaming with RenameMap, and can delete annotations Called by runTransform.
Perform the transform, encode renaming with RenameMap, and can delete annotations Called by runTransform.
Input Firrtl AST
A transformed Firrtl AST
A function that, given *another* transform (parameter a) will return true if this transform invalidates/undos the
effects of the *other* transform (parameter a).
A function that, given *another* transform (parameter a) will return true if this transform invalidates/undos the
effects of the *other* transform (parameter a).
transform
A convenience function useful for debugging and error messages
A convenience function useful for debugging and error messages
All transform that must run before this transform
All transform that must run before this transform
The use of a Seq here is to preserve input order. Internally, this will be converted to a private, ordered Set.
Perform the transform and update annotations.
Perform the transform and update annotations.
Input Firrtl AST
A transformed Firrtl AST
A mathematical transform on some type
A mathematical transform on some type
an output object of the same type
All transforms that must run after this transform
All transforms that must run after this transform
This is a means of prerequisite injection into some other transform. Normally a transform will define its own prerequisites. Dependents exist for two main situations:
First, they improve the composition of optional transforms. If some first transform is optional (e.g., an expensive validation check), you would like to be able to conditionally cause it to run. If it is listed as a prerequisite on some other, second transform then it must always run before that second transform. There's no way to turn it off. However, by listing the second transform as a dependent of the first transform, the first transform will only run (and be treated as a prerequisite of the second transform) if included in a list of target transforms that should be run.
Second, an external library would like to inject some first transform before a second transform inside FIRRTL. In this situation, the second transform cannot have any knowledge of external libraries. The use of a dependent here allows for prerequisite injection into FIRRTL proper.
(Since version FIRRTL 1.3) Due to confusion, 'dependents' is being renamed to 'optionalPrerequisiteOf'. Override the latter instead.
The use of a Seq here is to preserve input order. Internally, this will be converted to a private, ordered Set.
firrtl.passes.CheckTypes for an example of an optional checking firrtl.Transform
(Since version FIRRTL 1.3) Use Dependency API methods for equivalent functionality. See: https://bit.ly/2Voppre
(Since version FIRRTL 1.3) Use Dependency API methods for equivalent functionality. See: https://bit.ly/2Voppre
Infers the widths of all signals with unknown widths
Is a global width inference algorithm - Instances of the same module with unknown input port widths will be assigned the largest width of all assignments to each of its instance ports - If you don't want the global inference behavior, then be sure to define all your input widths
Infers the smallest width is larger than all assigned widths to a signal - Note that this means that dummy assignments that are overwritten by last-connect-semantics can still influence width inference - E.g. wire x: UInt x <= UInt<5>(15) x <= UInt<1>(1)
Since width inference occurs before lowering, it infers x's width to be 5 but with an assignment of UInt(1):
wire x: UInt<5> x <= UInt<1>(1)
Uses firrtl.constraint package to infer widths