Skip to content
STIMSMITH

Directed Random Stimulus

Concept WIKI v1 · 5/28/2026

Directed random stimulus is a verification stimulus style used alongside constrained-random and fully directed scenarios. In processor verification, it steers random instruction streams toward hard-to-hit conditions—such as arithmetic rounding and overflow cases—by seeding or directing parts of the scenario with special values while still allowing randomized generation.

Definition

Directed random stimulus is a stimulus-generation style that combines directed setup with randomized scenario generation. In the processor-verification example described by Chen, it is presented as a second scenario type alongside constrained-random scenarios and directed scenarios. Its purpose is to make hard-to-hit bug classes more likely than they would be under pure random stimulus alone. [1]

Use in processor verification

Directed random stimulus is useful when pure random inputs are unlikely to expose specific numerical corner cases. For arithmetic operations, the cited example notes that rounding and overflow bugs can be difficult to find with purely random numbers. A directed-random scenario can address this by preloading part of data memory with special values, such as walking patterns of 0s and 1s, values close to zero, and values close to minimum or maximum numeric limits. The scenario can then read those values from memory before executing an arithmetic instruction stream. [1]

Relationship to other scenario types

The cited stimulus-generation flow distinguishes three scenario styles:

  • Constrained-random scenarios, such as constraining a long sequence to computational operations only.
  • Directed random scenarios, which bias or seed the generated stimulus toward specific hard-to-hit data conditions while still using randomized instruction streams.
  • Directed scenarios, such as reading a pre-assembled program as a directed scenario. [1]

A scenario generator can work with all three types. Given a set of scenario objects, it randomly selects and randomizes one scenario, repeating the process until a user-specified stopping condition is reached. [2]

Practical role

Directed random stimulus is therefore useful as a middle ground: it is more targeted than unconstrained pure random stimulus, but less fixed than a fully directed scenario or pre-assembled program trace. This makes it suitable for focusing verification effort on specific risk areas while retaining the broader exploration benefits of randomization. [1]

LINKED ENTITIES

2 links
Scenario Generator USES
Directed Test USES

CITATIONS

2 sources
2 citations
[1] Directed random stimulus is a scenario style used alongside constrained-random and directed scenarios, and it targets hard-to-hit numerical cases such as rounding and overflow by using special values before randomized arithmetic instruction streams. Applying constrained-random verification to microprocessors
[2] A scenario generator can work with constrained-random, directed-random, and directed scenarios by randomly selecting and randomizing scenario objects until a user-specified condition is reached. Applying constrained-random verification to microprocessors