chisel3.util.random

Fibonacci Linear Feedback Shift Register (LFSR) generator.

A Fibonacci LFSR can be generated by defining a width and a set of tap points (corresponding to a polynomial). An optional initial seed and a reduction operation (XOR, the default, or XNOR) can be used to augment the generated hardware. The resulting hardware has support for a run-time programmable seed (via PRNGIO.seed) and conditional increment (via PRNGIO.increment).

$seedExplanation

In the example below, a 4-bit Fibonacci LFSR is constructed. Tap points are defined as four and three (using LFSR convention of indexing from one). This results in the hardware configuration shown in the diagram.

val lfsr4 = Module(new FibonacciLFSR(4, Set(4, 3))
//                 +---+
// +-------------->|XOR|-------------------------------------------------------+
// |               +---+                                                       |
// |   +-------+     ^     +-------+           +-------+           +-------+   |
// |   |       |     |     |       |           |       |           |       |   |
// +---+  x^4  |<----+-----|  x^3  |<----------|  x^2  |<----------|  x^1  |<--+
//     |       |           |       |           |       |           |       |
//     +-------+           +-------+           +-------+           +-------+

If you require a maximal period Fibonacci LFSR of a specific width, you can use MaxPeriodFibonacciLFSR. If you only require a pseudorandom UInt you can use the FibonacciLFSR companion object.

Attributes

See also: https://en.wikipedia.org/wiki/Linear-feedback_shift_register#Fibonacci_LFSRs $paramWidth $paramTaps $paramSeed $paramReduction $paramStep $paramUpdateSeed
Companion: object
Source: FibonacciLFSR.scala
Supertypes: trait LFSR

class PRNG

class Module

trait ImplicitReset

trait ImplicitClock

class RawModule

class BaseModule

trait Selectable

trait Selectable

trait IsInstantiable

trait InstanceId

class Object

trait Matchable

class Any
Show all
Known subtypes: class MaxPeriodFibonacciLFSR

Utility for generating a pseudorandom UInt from a FibonacciLFSR.

For example, to generate a pseudorandom 8-bit UInt that changes every cycle, you can use:

val pseudoRandomNumber = FibonacciLFSR.maxPeriod(8)

Attributes

Companion: class
Source: FibonacciLFSR.scala
Supertypes: class Object

trait Matchable

class Any
Self type: FibonacciLFSR.type

Galois Linear Feedback Shift Register (LFSR) generator.

A Galois LFSR can be generated by defining a width and a set of tap points. Optionally, an initial seed and a reduction operation (XOR, the default, or XNOR) can be used to augment the generated hardware. The resulting hardware has support for a run-time programmable seed (via PRNGIO.seed) and conditional increment (via PRNGIO.increment).

If the user specifies a seed, then a compile-time check is added that they are not initializing the LFSR to a state which will cause it to lock up. If the user does not set a seed, then the least significant bit of the state will be set or reset based on the choice of reduction operator.

In the example below, a 4-bit LFSR Galois LFSR is constructed. The tap points are defined as four and three (using LFSR convention of indexing from one). This results in the hardware configuration shown in the diagram.

val lfsr4 = Module(new GaloisLFSR(4, Set(4, 3))
// +-----------------+---------------------------------------------------------+
// |                 |                                                         |
// |   +-------+     v     +-------+           +-------+           +-------+   |
// |   |       |   +---+   |       |           |       |           |       |   |
// +-->|  x^4  |-->|XOR|-->|  x^3  |---------->|  x^2  |---------->|  x^1  |---+
//     |       |   +---+   |       |           |       |           |       |
//     +-------+           +-------+           +-------+           +-------+

If you require a maximal period Galois LFSR of a specific width, you can use MaxPeriodGaloisLFSR. If you only require a pseudorandom UInt you can use the GaloisLFSR companion object.

Value parameters

reduction: the reduction operation (either XOR or XNOR)
seed: an initial value for internal LFSR state. If None, then the LFSR state LSB will be set to a known safe value on reset (to prevent lock up).
step: the number of state updates per cycle
taps: a set of tap points to use when constructing the LFSR
updateSeed: if true, when loading the seed the state will be updated as if the seed were the current state, if false, the state will be set to the seed
width: the width of the LFSR

Attributes

See also: https://en.wikipedia.org/wiki/Linear-feedback_shift_register#Galois_LFSRs
Companion: object
Source: GaloisLFSR.scala
Supertypes: trait LFSR

class PRNG

class Module

trait ImplicitReset

trait ImplicitClock

class RawModule

class BaseModule

trait Selectable

trait Selectable

trait IsInstantiable

trait InstanceId

class Object

trait Matchable

class Any
Show all
Known subtypes: class MaxPeriodGaloisLFSR

Utility for generating a pseudorandom UInt from a GaloisLFSR.

For example, to generate a pseudorandom 8-bit UInt that changes every cycle, you can use:

val pseudoRandomNumber = GaloisLFSR.maxPeriod(8)

Attributes

Companion: class
Source: GaloisLFSR.scala
Supertypes: class Object

trait Matchable

class Any
Self type: GaloisLFSR.type

Trait that defines a Linear Feedback Shift Register (LFSR).

$seedExplanation

Attributes

See also: FibonacciLFSR

GaloisLFSR

https://en.wikipedia.org/wiki/Linear-feedback_shift_register
Companion: object
Source: LFSR.scala
Supertypes: class PRNG

class Module

trait ImplicitReset

trait ImplicitClock

class RawModule

class BaseModule

trait Selectable

trait Selectable

trait IsInstantiable

trait InstanceId

class Object

trait Matchable

class Any
Show all
Known subtypes: class FibonacciLFSR

class MaxPeriodFibonacciLFSR

class GaloisLFSR

class MaxPeriodGaloisLFSR

Utilities related to psuedorandom number generation using Linear Feedback Shift Registers (LFSRs).

For example, to generate a pseudorandom 16-bit UInt that changes every cycle, you can use:

val pseudoRandomNumber = LFSR(16)

Attributes

Companion: trait
Source: LFSR.scala
Supertypes: class Object

trait Matchable

class Any
Self type: LFSR.type

A reduction operation for an LFSR.

Attributes

See also: XOR

XNOR
Source: LFSR.scala
Supertypes: trait (Bool, Bool) => Bool

class Object

trait Matchable

class Any
Known subtypes: object XNOR

object XOR

A maximal period Fibonacci Linear Feedback Shift Register (LFSR) generator. The maximal period taps are sourced from LFSR.tapsMaxPeriod.

val lfsr8 = Module(new MaxPeriodFibonacciLFSR(8))

$paramWidth $paramSeed $paramReduction

Attributes

Source: FibonacciLFSR.scala
Supertypes: class FibonacciLFSR

trait LFSR

class PRNG

class Module

trait ImplicitReset

trait ImplicitClock

class RawModule

class BaseModule

trait Selectable

trait Selectable

trait IsInstantiable

trait InstanceId

class Object

trait Matchable

class Any
Show all

A maximal period Galois Linear Feedback Shift Register (LFSR) generator. The maximal period taps are sourced from LFSR.tapsMaxPeriod.

val lfsr8 = Module(new MaxPeriodGaloisLFSR(8))

Value parameters

reduction: the reduction operation (either XOR or XNOR)
seed: an initial value for internal LFSR state. If None, then the LFSR state LSB will be set to a known safe value on reset (to prevent lock up).
width: the width of the LFSR

Attributes

Source: GaloisLFSR.scala
Supertypes: class GaloisLFSR

trait LFSR

class PRNG

class Module

trait ImplicitReset

trait ImplicitClock

class RawModule

class BaseModule

trait Selectable

trait Selectable

trait IsInstantiable

trait InstanceId

class Object

trait Matchable

class Any
Show all

An abstract class representing a Pseudo Random Number Generator (PRNG)

Value parameters

seed: the initial state of the PRNG
step: the number of state updates per cycle
updateSeed: if true, when loading the seed the state will be updated as if the seed were the current state, if false, the state will be set to the seed
width: the width of the PRNG

Attributes

Companion: object
Source: PRNG.scala
Supertypes: class Module

trait ImplicitReset

trait ImplicitClock

class RawModule

class BaseModule

trait Selectable

trait Selectable

trait IsInstantiable

trait InstanceId

class Object

trait Matchable

class Any
Show all
Known subtypes: class FibonacciLFSR

class MaxPeriodFibonacciLFSR

class GaloisLFSR

class MaxPeriodGaloisLFSR

trait LFSR

Helper utilities related to the construction of Pseudo Random Number Generators (PRNGs)

Attributes

Companion: class
Source: PRNG.scala
Supertypes: class Object

trait Matchable

class Any
Self type: PRNG.type

Pseudo Random Number Generators (PRNG) interface

Value parameters

n: the width of the LFSR

Attributes

Source: PRNG.scala
Supertypes: class Bundle

class Record

trait Selectable

trait Aggregate

class Data

trait InstanceId

class Object

trait Matchable

class Any
Show all

Not XOR (exclusive or) reduction operation

Attributes

Source: LFSR.scala
Supertypes: trait LFSRReduce

trait (Bool, Bool) => Bool

class Object

trait Matchable

class Any
Self type: XNOR.type

XOR (exclusive or) reduction operation

Attributes

Source: LFSR.scala
Supertypes: trait LFSRReduce

trait (Bool, Bool) => Bool

class Object

trait Matchable

class Any
Self type: XOR.type

chisel3.util.random

Members list

Type members

Classlikes

Attributes

Attributes

Value parameters

Attributes

Attributes

Attributes

Attributes

Attributes

Attributes

Value parameters

Attributes

Value parameters

Attributes

Attributes

Value parameters

Attributes

Attributes

Attributes