登录查看更多内容

Last updated on 2024年10月8日

What are some effective functional coverage techniques for verifying complex designs?

由人工智能和领英社区提供技术支持

Functional verification is the process of ensuring that a design meets its specifications and requirements before it is implemented in hardware. It involves creating testbenches, stimuli, checkers, and monitors to simulate and verify the design's behavior and functionality. However, verifying complex designs can be challenging, as they may have many features, modes, configurations, and corner cases that need to be tested. How can you ensure that your verification plan covers all the relevant scenarios and conditions? How can you measure and improve the quality and completeness of your verification effort? In this article, we will discuss some effective functional coverage techniques that can help you answer these questions and achieve your verification goals.

此文章中的业界达人

由社区从 14 条内容中精选。了解更多

Sweety P.

Formal Verification Engineer | Certified Keynote Speaker | Married to VLSI | Insatiable burning desire for…
Negar Motamen

ASIC/ FPGA Designer Engineer

1 What is functional coverage?

Functional coverage is a metric that quantifies how well your testbench exercises the design under test (DUT). It measures how many of the predefined functional scenarios, features, or behaviors of the design have been tested by your stimuli. Functional coverage can help you identify gaps, redundancies, and inefficiencies in your verification plan, as well as guide you to create more effective and targeted test cases. Functional coverage can be collected and analyzed using various methods, such as code coverage, assertion coverage, covergroups, coverpoints, and cross coverage.

添加您的观点

Sweety P.

Formal Verification Engineer | Certified Keynote Speaker | Married to VLSI | Insatiable burning desire for Semiconductors | "You can't burn a woman made of FIRE. She is ME!!!" | @SwitiSpeaksOfficial
举报内容
The planning of functional coverage should start with verification plan. You need to start documenting various design functionalities & the corresponding cover groups, cover points, cross cover, etc. This document needs to be reviewed with architect & design & based on their inputs modify or add more coverage. Now, comes the coding part. Code the functional coverage as planned & run the test cases with coverage enabled. You will find the coverage holes & either add more test cases or modify existing ones till you achieve your coverage goal, ideally 100% If you are doing FPV, formal tools have an app for coverage. You should run that app to ensure you are not missing any valid design scenario. This gives good confidence in verification.

已翻译

赞
Negar Motamen

ASIC/ FPGA Designer Engineer
举报内容
In some cases, in order to have 100% coverage, Formal Verification could be the most effective and reliable solution. However for some cases it might seem overpriced but if reliability and availability of the system, it might be the best way.

已翻译

赞
Lotem Dallal

Hardware Engineer @intel | Getting Thing Done Expert ?? | gen-AI Enthusiast | M.Sc. Bio-Medical Engineer
举报内容
I would add that there is a verification methodology that suggests to use coverage for finding bugs. Is is called "coverage driven verification" and it uses coverage gaps found in DUT to address possible bugs.

已翻译

赞
Jo?o Bosco Diniz, MSc, SSM

Verification Lead at Koninklijke Philips N.V
举报内容
The ideal world is starting with design requirements analysis and then defining the verification tests accordingly. After that creating the Verification traceability matrix and test plan. Other inputs are usually included as field failure complains, DFMEA, risk analysis, safety analysis, etc.

已翻译

赞
Hardik Trivedi

"Passionate Pre Silicon Verification Engineer at Intel Corporation | Constantly Evolving | Unleashing Innovations in Complex Problem Solving in the World of Semiconductors"
举报内容
Effective functional coverage techniques for verifying complex designs: 1. Develop coverage models based on specifications 2. Employ assertion coverage 3. Implement protocol compliance coverage 4. Develop scenario-based coverage models 5. Implement power-aware coverage metrics 6. Employ coverage closure analysis tools These techniques ensure comprehensive verification of complex design functionalities.

已翻译

赞

加载更多内容

2 Why is functional coverage important?

Functional coverage is important because it provides feedback and insight into your verification progress and quality. It can help you answer questions such as: How much of the design functionality have you tested? Are there any untested or under-tested features or scenarios? How confident are you that your design meets its specifications and requirements? How can you optimize your verification resources and time? By using functional coverage, you can monitor and improve your verification plan, as well as communicate and demonstrate your verification results to stakeholders and customers.

添加您的观点

SUNODH KUMAR GARLAPATI

VLSI Design Verification Engineer
举报内容
leveraging functional coverage data can inform decision-making processes, such as prioritizing test development and resource allocation.

已翻译

赞

3 How to define functional coverage goals?

Before you can collect and analyze functional coverage, you need to define your functional coverage goals. These should be based on the design specifications and requirements, as well as the verification strategy and methodology. Examples of functional coverage goals include testing features and modes of the design, valid and invalid inputs and outputs, possible transitions and sequences, error and exception handling mechanisms, and performance and timing constraints. It is important to document your functional coverage goals in a verification plan, and review them regularly to ensure that they remain aligned with design changes and updates.

添加您的观点

Negar Motamen

ASIC/ FPGA Designer Engineer
举报内容
It would be a great idea to prepare a table for all test scenarios, the advantages of making tables for documentation are: 1. It would be easier for test scenario generator to generate scenarios for test bench. 2. All test scenarios could be categorized ( input/output validity tests, time constraints, functions, corners cases,and etc). By doing so, the verification engineer can also have a good insight of the coverage and functionality of the system. 3. Table is more readable and understandable for design team, by taking a quick look they can realize the weak points of the designed system.

已翻译

赞

4 How to collect functional coverage data?

To collect functional coverage data, you need to instrument your testbench with code, assertions, covergroups, coverpoints, and cross coverage. These tools enable you to specify which aspects of the design functionality you want to measure and track, and generate reports and statistics that show how much of the functional coverage goals have been met by your test cases. Code coverage measures how much of the design code has been executed by your testbench. Assertion coverage tracks how many of the assertions in your testbench have been triggered. Covergroups group together related coverpoints and cross coverage, which define the functional coverage goals in terms of bins or ranges. This helps you capture and track the diversity and complexity of your design functionality. It can also help you identify untested or dead code, verify correctness and robustness, detect and debug errors and bugs, as well as optimize code quality and efficiency.

添加您的观点

5 How to analyze functional coverage results?

To evaluate your verification progress and quality, you need to use tools that can generate and display reports and statistics of your functional coverage goals. These tools may include simulators, debuggers, viewers, analyzers, and databases, and they should provide features such as filtering, sorting, grouping, merging, comparing, and visualizing the functional coverage data. By analyzing these results, you can identify gaps, redundancies, and inefficiencies in your verification plan. This will allow you to create more effective and targeted test cases, optimize your verification resources and time, and communicate your verification results to stakeholders and customers.

添加您的观点

chandrakanth betageri

Design Verification Engineer at Intel Corporation | Senior IEEE Member
举报内容
The converged coverage vdb from regression database helps to analyze the coverage results. It helps to measure the coverage database and provide the datapoint to see missing coverages and analyze and address the coverage gaps by adding more stimulus, more randomization. Coverage App from Jasper also provides good data points for analysis and manage coverage gaps for the design.

已翻译

赞

6 How to improve functional coverage?

To improve functional coverage, it is necessary to use techniques that can enhance your verification plan and testbench. This includes reviewing and updating your functional coverage goals according to design specifications and requirements, as well as utilizing a verification methodology and framework that supports functional coverage collection and analysis. Additionally, randomization, constraints, and distribution should be used to generate diverse and realistic stimuli. Feedback, adaptive, and directed testing should be used to adjust stimuli according to the functional coverage results. Furthermore, assertions, checkers, and monitors should be employed to verify the design behavior and functionality. Finally, code quality and style guidelines should be used to improve readability and maintainability, as well as best practices and standards to ensure your functional coverage is consistent and accurate.

添加您的观点

Sagar Arora

ASIC DV Staff Er @Synopsys | Ex - Intel , Samsung | BITS Pilani
举报内容
Functional Coverage needs to be monitored regularly. If the results are not in favor, we need to analyze the the verification plan and the stimulus first. Plan should be made to mitigate the low functional coverage by enhancing existing scenarios, working on their randomization and configurations to expand their reachability. Adding new scenarios can be another option if some uncovered functionalities are observed. Regression frequency and increase in test count per regression is another approach that helps. Analysis of the bins can be done to ignore certain bins that may not be very important to cover after stakeholders approval.

已翻译

赞
chandrakanth betageri

Design Verification Engineer at Intel Corporation | Senior IEEE Member
举报内容
To improve functional coverage, focus on comprehensive test planning that identifies and targets all possible scenarios, including edge cases and corner cases. Utilize constrained-random testing to explore a wide range of conditions, and implement coverage-driven verification to track progress and gaps. Enhance stimulus generation with intelligent testbench designs and use functional coverage metrics to guide test creation. Regularly review and refine your coverage model to ensure it aligns with the evolving design spec and requirements. In addition, Formal Coverage App provides good confidence boost for coverage and identify coverage holes in the verification through stimuli coverage, checker coverage and comprehensive formal coverage.

已翻译

赞

加载更多内容

7 Here’s what else to consider

This is a space to share examples, stories, or insights that don’t fit into any of the previous sections. What else would you like to add?

添加您的观点

Lotem Dallal

Hardware Engineer @intel | Getting Thing Done Expert ?? | gen-AI Enthusiast | M.Sc. Bio-Medical Engineer
举报内容
I would start by understanding how coverage scores are calculated in the organization, and also understanding how target scores are set.

已翻译

赞

加载更多内容

Functional Verification

+ 关注

给文章评分

我们借助人工智能创建了此文章。您认为这篇文章怎么样？

很棒不太好

举报此文章

查看全部

What are some effective functional coverage techniques for verifying complex designs?

1

2

3

4

5

6

7

1 What is functional coverage?

2 Why is functional coverage important?

3 How to define functional coverage goals?

4 How to collect functional coverage data?

5 How to analyze functional coverage results?

6 How to improve functional coverage?

7 Here’s what else to consider

Functional Verification

给文章评分

感谢您的反馈

更多Functional Verification相关文章

更多相关阅读内容