Skip to content
STIMSMITH

Program-based stimulus

Concept WIKI v1 · 5/26/2026

A program-based stimulus is a form of verification stimulus for simulation-based processor verification in which the stimulus is represented as a program loaded directly into the processor's program memory, rather than as raw bit vectors applied to input ports.

Definition

In simulation-based processor verification, a program-based stimulus is a stimulus represented by a program that is loaded directly into the processor's program memory. This contrasts with another stimulus form in which bit vectors are applied directly to the processor's input ports.

Context in processor verification

Simulation-based processor verification commonly uses pseudorandom generators to generate stimuli, applies those stimuli to processor inputs, and monitors functional coverage to assess verification completeness. Within that workflow, stimuli may take different forms, including input-port bit vectors or programs loaded into program memory.

Role in coverage-driven verification

The cited processor-verification work describes a coverage-driven approach in which constraints for a pseudorandom generator are dynamically altered using a recurrent neural network that receives coverage feedback from simulation of the design under verification. In that setting, program-based stimuli are one of the supported forms of stimuli that can be generated and evaluated for coverage.

Regression-test relevance

The same work reports that its experiments not only reached coverage closure sooner, but also isolated a small set of high-coverage stimuli suitable for regression tests. Program-based stimuli, as one possible stimulus representation in that verification context, can therefore be part of a reduced high-coverage regression stimulus set when produced by such a flow.

CITATIONS

4 sources
4 citations
[1] Program-based stimuli are programs loaded directly into processor program memory, contrasted with bit-vector stimuli applied to input ports. Automation of Processor Verification Using Recurrent Neural Networks
[2] Simulation-based processor verification commonly generates stimuli using pseudorandom generators, applies them to processor inputs, and monitors functional coverage to determine verification completeness. Automation of Processor Verification Using Recurrent Neural Networks
[3] The cited work proposes dynamically altering pseudorandom-generator constraints using a recurrent neural network that receives coverage feedback from simulation of the design under verification. Automation of Processor Verification Using Recurrent Neural Networks
[4] The experiments reported earlier coverage closure and isolation of a small high-coverage stimulus set usable for regression tests. Automation of Processor Verification Using Recurrent Neural Networks