Skip to content
STIMSMITH

Jenkins

Tool WIKI v2 · 5/27/2026

Jenkins is an open-source CI server and automation server. In the provided RISC-V vector accelerator verification evidence, Jenkins is the basis of a CI infrastructure that generates and runs tests, records coverage metrics, retries failed tests, selects regression suites by coverage, and runs merge-gate and weekly regressions.

Overview

Jenkins is described in the evidence as an open-source CI server. Its public upstream repository describes it as the “Jenkins automation server” and identifies the implementation language as Java. [jenkins-public-repo] [jenkins-ci-server]

Role in the RISC-V vector accelerator verification flow

In the cited verification environment, the Continuous Integration infrastructure was built using Jenkins. The authors created a set of Jenkins pipelines that interact with each other with the goal of improving design correctness. [jenkins-ci-server]

The CI flow used Jenkins to generate and run tests and to collect metrics from simulations. The recorded metrics included functional coverage, assertion usage such as active and passed assertions, and code coverage. [jenkins-coverage-metrics]

Jenkins pipeline structure

The Jenkins-based infrastructure implemented four pipelines: [jenkins-pipelines]

  1. New tests — Generates random tests with RISCV-DV, compiles the DUT, executes the binaries, and classifies tests as passed or failed. Passed tests are used to create a regression set, while failed tests are retained for debugging and rechecking until the error is fixed. [new-tests-pipeline]
  2. Retry — For each change in the main branch of the DUT repository, re-executes the set of failed tests and classifies them again as passed or failed. [retry-pipeline]
  3. Selection — Runs every day at midnight. If the number of passed tests is above a threshold, tests are ranked by collected coverage, and two regression sets are created: a large set and a small set. [selection-pipeline]
  4. Regressions — When a DUT change is a candidate to be merged, the small regression set is executed to check correctness. Once per week, the large regression set is executed to ensure recent changes do not break known-good tests. [regressions-pipeline]

Technical significance

Within this verification infrastructure, Jenkins serves as the orchestration layer for random test generation, DUT compilation, binary execution, test-result classification, failed-test retry, coverage-driven regression selection, and scheduled or merge-triggered regression execution. [jenkins-pipelines]

This role connects Jenkins directly to the broader CI/CD infrastructure used by the verification team: the Jenkins pipelines automated recurring validation activities and provided coverage and assertion data used to evaluate simulation results. [jenkins-ci-server] [jenkins-coverage-metrics]

LINKED ENTITIES

2 links
Functional Verification of a RISC-V Vector Accelerator USES
CI/CD Infrastructure IMPLEMENTS

CITATIONS

8 sources
8 citations
[1] Jenkins public repository describes it as the Jenkins automation server and identifies Java as the repository language. jenkinsci/jenkins
[2] The verification infrastructure was built using the open-source CI server Jenkins, with a set of interacting pipelines. Functional Verification of a RISC-V Vector Accelerator
[3] The CI flow generated and ran tests and collected functional coverage, assertion usage, and code coverage metrics from simulations. Functional Verification of a RISC-V Vector Accelerator
[4] The Jenkins infrastructure implemented four pipelines: New tests, Retry, Selection, and Regressions. Functional Verification of a RISC-V Vector Accelerator
[5] The New tests pipeline generated random tests with RISCV-DV, compiled the DUT, executed binaries, and classified tests as passed or failed. Functional Verification of a RISC-V Vector Accelerator
[6] The Retry pipeline re-executed failed tests for each change in the main branch of the DUT repository and reclassified them. Functional Verification of a RISC-V Vector Accelerator
[7] The Selection pipeline ran daily at midnight, ranked passed tests by collected coverage when a threshold was met, and created large and small regression sets. Functional Verification of a RISC-V Vector Accelerator
[8] The Regressions pipeline ran the small regression set for candidate DUT changes and the large regression set weekly to check that recent changes did not break known-good tests. Functional Verification of a RISC-V Vector Accelerator

VERSION HISTORY

v2 · 5/27/2026 · gpt-5.5 (current)
v1 · 5/27/2026 · gpt-5.5