Skip to content
STIMSMITH

Processor-level Input Stimuli Generation

Concept

Processor-level input stimuli generation is the creation of instruction-level test inputs for processor verification. Evidence from Bruns et al. (GLSVLSI 2022) describes traditional model-, constraint-, coverage-, machine-learning-, and symbolic-execution-based approaches, and a coverage-guided fuzzing approach that generates processor core instructions directly for RTL cross-level verification against an instruction set simulator.

First seen 5/28/2026
Last seen 5/28/2026
Evidence 3 chunks
Wiki v1

WIKI

Overview

Processor-level input stimuli generation is the generation of instruction-level test inputs used to verify processor designs. In simulation-based processor verification, these stimuli are intended to exercise processor behavior thoroughly enough to expose mismatches or bugs in the implementation.

Bruns et al. describe this area as a long-standing part of simulation-based processor verification, with multiple families of approaches proposed to improve the generation of processor-level stimuli. Reported directions include model-based test generators driven by input-format specifications, constraint-based generation using CSP/SMT solvers, optimizations for propagating constraints across multiple instructions, mining processor manuals to obtain input models, coverage-guided generation based on Bayesian networks, other machine-learning-based methods, and symbolic-execution-based formal techniques.

READ FULL ARTICLE →

NEIGHBORHOOD

No graph connections found for this entity yet. It may appear in future ingestion runs.

explore full graph →

RELATIONSHIPS

2 connections
The paper generates processor-level input stimuli using coverage-guided fuzzing techniques.
Endless Instruction Stream Generation ← implements 85% 1e
Endless instruction stream generation is a technique for generating processor-level stimuli.

CITATIONS

8 sources
8 citations — click to expand
[1] Simulation-based processor verification has a long history of test generation techniques for processor-level input stimuli, including model-based, constraint-based, coverage-guided, machine-learning, and symbolic-execution approaches. Efficient Cross-Level Processor Verification using Coverage-guided Fuzzing
[2] Bruns et al. propose applying state-of-the-art coverage-guided fuzzing to RTL processor verification using an ISS as reference model in cross-level co-simulation. Efficient Cross-Level Processor Verification using Coverage-guided Fuzzing
[3] The fuzzing process in the cited approach is guided by code coverage from both the reference ISS and the RTL core under test. Efficient Cross-Level Processor Verification using Coverage-guided Fuzzing
[4] The co-simulation was designed to feed the same instruction sequences to both models and to support arbitrary instruction sequences, including potential infinite loops. Efficient Cross-Level Processor Verification using Coverage-guided Fuzzing
[5] The cited fuzzing approach supports arbitrary control flows including self-loops, as well as load/store instructions. Efficient Cross-Level Processor Verification using Coverage-guided Fuzzing
[6] The approach uses custom mutation procedures tailored to generate common instruction patterns. Efficient Cross-Level Processor Verification using Coverage-guided Fuzzing
[7] The authors contrast their test-case-by-test-case generation flow with a custom generator that produces a single endless instruction stream, which they describe as architecture-specific and manually intensive. Efficient Cross-Level Processor Verification using Coverage-guided Fuzzing
[8] The authors report that their fuzzer optimizations improved verification results statistically significantly over a baseline state-of-the-art coverage-guided fuzzer and found numerous intricate bugs in VexRiscv. Efficient Cross-Level Processor Verification using Coverage-guided Fuzzing