马上注册,结交更多好友,享用更多功能,让你轻松玩转社区。
您需要 登录 才可以下载或查看,没有账号?注册
x
本帖最后由 movit 于 2023-8-28 22:51 编辑
https://semiwiki.com/artificial-intelligence/333623-amd-puts-synopsys-ai-verification-tools-to-the-test/原文The various algorithms that comprise artificial intelligence (AI) are finding their way into the chip design flow. What is driving a lot of this work is the complexity explosion of new chip designs required to accelerate advanced AI algorithms. It turns out AI is both the problem and the solution in this case. AI can be used to cut the AI chip design problem down to size. Synopsys has been developing AI-assisted design capabilities for quite a while, beginning with the release of a design space optimization capability (DSO.ai) in 2020. Since then, several new capabilities have been announced, significantly expanding its AI-assisted footprint. You can get a good overview of what Synopsys is working on here. One of the capabilities in the Synopsys portfolio focuses on verification space optimization (VSO.ai). The real test of any new capability is its use by a real customer on a real design, and that is the topic of this post. Read on to see how AMD puts Synopsys AI verification tools to the test. 翻译结果:人工智能(AI)组成的各种算法正在进入芯片设计流程。推动这项工作的原因是新芯片设计的复杂性爆炸,这些设计需要加快先进的AI算法。事实证明,在这种情况下,AI既是问题也是解决方案。AI可以用于将AI芯片设计问题缩小到合理的范围。Synopsys在很长时间内一直在开发AI辅助设计能力,从2020年发布设计空间优化能力(DSO.ai)开始。自那以后,它宣布了几个新的能力,显著扩大了其支持AI的范围。您可以在这里获得对Synopsys正在开展的工作的很好概述。Synopsys的某些能力集中在验证空间优化(VSO.ai)上。任何新功能的真正测试都是由真实客户在真实设计中使用,这就是本文的主题。继续阅读,了解AMD如何对Synopsys的AI验证工具进行测试。 简化版英文:The complexity of new chip designs needed for advanced AI algorithms is driving the integration of various AI algorithms into the chip design flow. AI is both the problem and the solution in this case. Synopsys has been developing AI-assisted design capabilities for a long time, starting with the release of DSO.ai in 2020. Since then, they have announced several new capabilities, expanding their AI-assisted footprint. One of the focuses of the Synopsys portfolio is VSO.ai for verification space optimization. This post discusses how AMD tests Synopsys AI verification tools in real designs. 固定搭配解释:find one's way into: 进入,融入
drive: 推动,驱动
cut down to size: 缩小到合理的范围
focus on: 专注于
put something to the test: 对某事进行测试
单词提取:comprise: 包括
chip: 芯片
complexity: 复杂性
accelerate: 加快
optimization: 优化
原文VSO.ai – What it Does Test coverage of a design is the core issue in semiconductor verification. The battle cry is, “if you haven’t exercised it, you haven’t verified it.” Stimulus vectors are generated using a variety of techniques, with constrained random being a popular approach. Those vectors are then used in simulation runs on the design, looking for test results that don’t match expected results. 翻译结果VSO.ai - 它是做什么的? 半导体验证中,设计的测试覆盖率是核心问题。人们高呼:“如果你没有对其进行测试,那你就没有验证它。” 通过使用各种技术生成刺激向量,其中受限随机是一种流行的方法。然后,这些向量被用于设计的模拟运行中,寻找与预期结果不匹配的测试结果。 简化版英文The test coverage of a design is the main issue in verifying semiconductor devices. The mantra is, "if you haven't tested it, you haven't verified it." Stimulus vectors are created using different methods, with constrained randomization being a popular approach. These vectors are then used in simulations, checking for test results that deviate from the expected outcomes. 固定搭配解释单词提取test coverage: 测试覆盖率
design: 设计
semiconductor: 半导体
verification: 验证
stimulus: 刺激
vector: 向量
constrained: 受限的
random: 随机的
原文By exercising more of the circuit, the chance of finding functional design flaws is increased. 翻译结果:通过更多地运行电路,增加了发现功能设计缺陷的机会。 简化版英文:Running the circuit more increases the chance of finding design flaws. 固定搭配解释:暂无固定搭配。 单词提取:exercise: 运行,执行
circuit: 电路
chance: 机会
原文Verification teams choose structural code coverage metrics (line, expression, block, etc.) of interest and automatically add them to simulation runs. As each test iteration generates constrained-random stimulus conforming to the rules, the simulator collects metrics for all the forms of coverage included. The results are monitored, with the goal of tweaking the constraints to try to improve the coverage. At some point, the team decides that they have done the best that they can within the schedule and resource constraints of the project, and they tape out. 翻译结果:验证团队选择感兴趣的结构代码覆盖度量(行、表达式、块等),并自动将它们添加到仿真运行中。每个测试迭代按照规则生成受限随机刺激,仿真器收集所有包含的覆盖度量。结果被监测,目标是调整约束条件,试图提高覆盖率。在某个阶段,团队决定在项目的时间表和资源限制内已经尽力而为,然后停止开发。 简化版英文:The verification team selects structural code coverage metrics of interest and adds them to simulation runs. Each test iteration generates constrained-random stimulus that follows certain rules, and the simulator collects metrics for all forms of coverage. The results are monitored, and the team tries to improve the coverage by adjusting the constraints. Eventually, the team decides to stop development when they have done their best within the project's schedule and resource constraints. 固定搭配解释:add to: 将...添加到
simulation run: 仿真运行
constrained-random stimulus: 受限随机刺激
collect metrics: 收集度量
monitor the results: 监测结果
adjust the constraints: 调整约束条件
stop development: 停止开发
schedule and resource constraints: 时间表和资源限制
单词提取:verification: 验证
structural: 结构的
code coverage: 代码覆盖度
metric: 度量
iteration: 迭代
conform to: 符合
simulator: 仿真器
tweak: 调整
constraint: 约束条件
tape out: 停止开发
原文Code coverage does not reflect the intended functionality of the design, so user-defined coverage is important. This is typically a manual effort, spanning only a limited percentage of the design’s behavior. Closing coverage and achieving verification goals is quite difficult. 翻译结果:代码覆盖率不能反映设计的预期功能,因此用户定义的覆盖率非常重要。这通常是一项手动工作,只涵盖设计行为的有限百分比。完成覆盖率和实现验证目标是相当困难的。 简化版英文:User-defined code coverage is important because it reflects the intended functionality of the design. However, it is a manual effort and only covers a limited percentage of the design's behavior. Closing coverage and achieving verification goals can be challenging. 固定搭配解释:code coverage: 代码覆盖率
user-defined: 用户定义的
verification goals: 验证目标
单词提取:coverage: 覆盖率
functionality: 功能性
design: 设计
manual: 手动的
behavior: 行为
closing: 完成
verification: 验证
原文A typical chip project runs many thousands of constrained-random simulation tests with a great deal of repetitive activity in the design. So, the rate of new coverage slows, and the benefit of each new test reduces over time. 翻译结果:一个典型的芯片项目运行着成千上万次约束随机模拟测试,在设计过程中有很多重复的活动。因此,新覆盖率的速度变慢,每个新测试所带来的益处随着时间的推移会减少。 简化版英文:A typical chip project runs thousands of constrained-random simulation tests with repetitive design activity. As a result, the rate of new coverage slows down and the benefit of each new test decreases over time. 固定搭配解释:单词提取:原文At some point, the curve flattens out, often before goals are met. The team must try to figure out what is going on and improve coverage as much as possible within time and resource constraints. This “last mile” of the process is quite challenging. The amount of data collected is overwhelming and trying to analyze it and determine the root cause of a coverage hole is difficult and labor-intensive. Is it an illegal bin for this configuration or a true coverage hole? The design of complex chips contains many problems that look like this – the requirement to analyze vast amounts of data and identify the best path forward. The good news is that AI techniques can be applied to this class of problems quite successfully. 翻译结果:在某个阶段,曲线趋平,通常是在目标实现之前。团队必须努力弄清楚发生了什么,并在时间和资源限制内尽可能改善覆盖范围。这个过程的“最后一英里”非常具有挑战性。收集到的数据量令人难以应对,试图分析数据并确定覆盖漏洞的根本原因是困难且充满工作量。这是合法配置不当还是真正的覆盖漏洞? 复杂芯片的设计存在许多类似的问题-分析大量数据并确定最佳前进路径的要求。好消息是,AI技术可以成功地应用于这类问题。 简化版英文:At some point, the team needs to figure out what is happening and improve coverage as much as possible within time and resource limitations. This is a challenging task because the amount of data collected is overwhelming and analyzing it to determine the root cause of a coverage hole is difficult and labor-intensive. It is important to distinguish whether it is an incorrect configuration or an actual coverage hole. Complex chip designs often require analyzing vast amounts of data to find the best solution, but AI techniques can be utilized successfully in solving these types of problems. 固定搭配解释:figure out: 弄清楚,找出
improve coverage: 改善覆盖范围
within time and resource constraints: 在时间和资源限制内
last mile: 最后一英里
overwhelming: 令人难以应对的
analyze data: 分析数据
determine the root cause: 确定根本原因
coverage hole: 覆盖漏洞
incorrect configuration: 不正确的配置
actual: 真正的,实际的
chip design: 芯片设计
best solution: 最佳解决方案
utilize: 利用
单词提取:curve: 曲线
flatten out: 趋平,变平缓
team: 团队
figure out: 弄清楚,找出
improve: 改善
coverage: 覆盖范围
possible: 可能的
time: 时间
resource: 资源
constraint: 限制
last mile: 最后一英里
process: 过程
challenging: 具有挑战性的
amount: 数量
data: 数据
overwhelming: 令人难以应对的
analyze: 分析
determine: 确定
root cause: 根本原因
hole: 漏洞
illegal: 不合法的
bin: 容器
configuration: 配置
true: 真正的,真实的
design: 设计
complex: 复杂的
chip: 芯片
contain: 包含
problem: 问题
vast: 巨大的
best: 最佳的
path: 路径
forward: 前进
good: 好的
news: 新闻
techniques: 技术
apply: 应用
class: 类别
successful: 成功的
原文For coverage definition, Synopsys VSO.ai infers some types of coverage beyond traditional code coverage to complement user-specified coverage. Machine learning (ML) can learn from experience and intelligently reuse coverage when appropriate. Even during a single project, learnings from earlier coverage results can help to improve coverage models.VSO.ai works at the coarse-grained test level and provides automated, adaptive test optimization that learns as the results change. Running the tests with highest ROI first while eliminating redundant tests accelerates coverage closure and saves compute resources. 翻译结果:对于覆盖率的定义,Synopsys VSO.ai通过推断除了传统的代码覆盖率之外的一些覆盖类型,来补充用户指定的覆盖范围。机器学习(ML)可以从经验中学习,并在适当时智能地重用覆盖范围。即使在一个项目中,来自之前覆盖结果的经验也可以帮助提高覆盖模型,VSO.ai在粗粒度测试级别上工作,并提供自动化、自适应的测试优化,随着结果的变化而学习。先运行具有最高投资回报率的测试,同时消除冗余测试,加快覆盖范围的完成并节省计算资源。 简化版英文:Synopsys VSO.ai uses machine learning to optimize test coverage by inferring different types of coverage beyond traditional code coverage. By learning from experience, VSO.ai can intelligently reuse coverage to complement user-specified coverage. VSO.ai works at the test level and provides automated test optimization that adapts as results change. Running the most impactful tests first while eliminating redundant tests helps to speed up coverage closure and save resources. 固定搭配解释:infers some types of coverage: 推断一些覆盖的类型
complement user-specified coverage: 补充用户指定的覆盖范围
learn from experience: 从经验中学习
reuse coverage: 重用覆盖范围
test optimization: 测试优化
coverage closure: 覆盖范围的完成
save compute resources: 节省计算资源
单词提取:原文The tool also works at the fine-grained level within the simulator to improve the test quality of results (QoR) by adapting the constrained-random stimulus to better target unexercised coverage points. This not only accelerates coverage closure, but also drives convergence to a higher percentage value. 翻译结果:该工具还在模拟器的细粒度层面上工作,通过调整约束随机刺激以更好地定位未使用的覆盖点来提高测试结果(QoR)的质量。这不仅加速了覆盖点的覆盖,还推动了收敛到更高百分比的数值。 简化版英文:The tool improves the quality of test results by adapting the stimulus to target unexercised coverage points, accelerating coverage closure and convergence to a higher percentage value. 固定搭配解释:fine-grained level: 细粒度层面
improve the quality of: 提高...的质量
target coverage points: 定位覆盖点
accelerate coverage closure: 加速覆盖点覆盖
drive convergence: 推动收敛
higher percentage value: 更高百分比的数值
单词提取:原文The last mile closure challenge is addressed by automated, AI-driven analysis of coverage results. VSO.ai performs root cause analysis (RCA) to determine why specific coverage points are not being reached. If the tool can resolve the situation itself, it will. Otherwise, it presents the team with actionable results, such as identifying conflicting constraints. 翻译结果:通过自动化和人工智能驱动的覆盖分析来解决最后一英里的闭环难题。VSO.ai进行根本原因分析来确定为什么没有达到特定的覆盖点。如果工具可以自行解决问题,它会这样做。否则,它会向团队展示可操作的结果,如识别冲突约束条件。 简化版英文:Automated analysis of coverage results is used to address the challenge of the last mile closure. VSO.ai performs root cause analysis (RCA) to determine why specific coverage points are not reached. It provides actionable results to the team, such as identifying conflicting constraints, if it cannot resolve the situation itself. 固定搭配解释:address the challenge: 解决挑战
root cause analysis: 根本原因分析
reach coverage points: 达到覆盖点
resolve the situation: 解决情况
provide actionable results: 提供可操作的结果
identify conflicting constraints: 识别冲突约束条件
原文The figure below summarizes the benefits VSO.ai can deliver. A top-level benefit of these approaches is the achievement of superior results in less time with less designer effort. We will re-visit this statement in a moment. 翻译结果下图总结了VSO.ai可以提供的好处。这些方法的顶层好处是在更少的时间和更少的设计师努力下实现更好的结果。我们将在一会儿重新审视这个表述。 简化版英文The figure below shows the benefits of using VSO.ai. One major benefit is achieving better results in less time with less effort from designers. We will discuss this further later. 固定搭配解释单词提取benefit: 好处
approach: 方法
achievement: 实现
superior: 更好的
effort: 努力
原文What AMD Found At the recent Synopsys Users Group (SNUG) held in Silicon Valley, AMD presented a paper entitled, “Drop the Blindfold: Coverage-Regression Optimization in Constrained-Random Simulations using VSO.ai (Verification Space Optimization) .” The paper detailed AMD’s experiences using VSO.ai on several designs. AMD had substantial goals and expectations for this work:Reach 100% coverage consistently with small RTL changes and design variants, but in an optimized, automated way. 翻译结果AMD发现了什么 在最近在硅谷举办的Synopsys User Group(SNUG)上,AMD提出了一份题为“脱下眼罩:使用VSO.ai(验证空间优化)在有约束随机模拟中进行覆盖回归优化”的论文。该论文详细介绍了AMD在几个设计中使用VSO.ai的经验。AMD对这项工作有着重要的目标和期望:通过小型RTL更改和设计变体以优化自动化方式,达到 100% 的覆盖率。 简化版英文AMD found that using VSO.ai in constrained-random simulations can optimize and automate the process of reaching consistent 100% coverage with small RTL changes and design variants. 固定搭配解释单词提取constrained: 有约束的
random: 随机的
simulation: 模拟
optimize: 优化
automated: 自动化的
原文AMD applied a well-documented methodology using VSO.ai across regression samples for four different designs. The figure below summarizes these four experiments. 翻译结果AMD采用了一个经过充分记录的方法,在四个不同设计的回归样本中使用了VSO.ai。下图总结了这四个实验。 简化版英文AMD used a well-documented methodology with VSO.ai on regression samples for four designs. The figure below summarizes these four experiments. 固定搭配解释单词提取methodology: 方法论
regression: 回归
design: 设计
原文Regression Characteristics Across Four Designs AMD then presented a detailed overview of these designs, their challenges and the results achieved by using VSO.ai, compared to the original effort without VSO.ai. Recall one of the hallmark benefits of applying AI to the design process: Achievement of superior results in less time with less designer effort. In its SNUG presentation, awarded one of the Top 10 Best Presentations at the event, AMD summarized the observed benefits as follows: 1.5 – 16X reduction in the number of tests being run across the four designs to achieve the same coverage. Quick, on-demand regression qualifier. Can be used to gauge how well the test distribution of a regression is if user is uncertain on iterations needed. Potentially target more bins under same budget. If default regression(s) do not achieve 100% coverage, VSO.ai can potentially exceed this (i.e., experiment #1). Testcase(s) removal in coverage regressions if not contributing. More reliable test grading for constrained random testsURG (Unified report generator): seed-based v/s VSO.ai: probability-based. Debug. Uncover coverage items that have a lower probability of being hit than expected. This presentation put VSO.ai to the test and the positive impact of the tool was documented. As mentioned, this kind of user application to real designs is the real test of a new technology. And that’s how AMD puts Synopsys AI verification tools to the test.. 翻译结果:四种设计的回归特征 随后,AMD提供了对这些设计的详细概述,以及使用VSO.ai相比未使用VSO.ai的原始工作的挑战和实现的结果。回想一下,将人工智能应用于设计过程的标志性好处之一:在更少的时间和更少的设计师工作的情况下实现更优秀的结果。在其SNUG演示中,AMD总结了观察到的好处如下:在四种设计中运行的测试数量减少了1.5至16倍,以实现相同的覆盖范围。快速的按需回归验证。如果用户对所需迭代次数不确定,可以用来评估回归的测试分布情况。在同一预算下,可能针对更多的箱体。如果默认的回归测试未达到100%的覆盖率,VSO.ai有可能超越这一点(即实验#1)。覆盖回归中的测试用例如果没有作出贡献,则移除。对于约束随机测试,具有更可靠的测试评分。URG(统一报告生成器):基于种子和VSO.ai:基于概率。调试。揭示比预期更低概率被触发的覆盖项。这次演示对VSO.ai进行了测试,并记录了该工具的积极影响。如前所述,这种用户将新技术应用于真实设计,这是对新技术的真实测试。这就是AMD如何对Synopsys人工智能验证工具进行测试。。 简化版英文:AMD presented an overview of four designs, their challenges, and the results achieved by using VSO.ai compared to without it. The benefits of applying AI to the design process are superior results achieved in less time with less effort. AMD summarized the observed benefits as follows: reduction in the number of tests needed for the same coverage across the four designs, quick regression qualification, the ability to gauge the test distribution of a regression if unsure about the number of iterations needed, the potential to target more bins within the same budget, the possibility of achieving higher coverage than default regressions, the removal of test cases that do not contribute to coverage regressions, more reliable test grading for constrained random tests, and the use of probability-based VSO.ai instead of seed-based URG (Unified Report Generator) for debugging and identifying coverage items with lower than expected hit probability. The positive impact of VSO.ai was demonstrated through real design applications, which is the true test of a new technology for AMD. 固定搭配解释:apply AI to: 将人工智能应用于
achieve superior results: 获得更优秀的结果
reduction in: 减少
run tests: 运行测试
achieve coverage: 实现覆盖率
gauge how well: 评估...好坏程度
target bins: 针对箱体
exceed 100%: 超过100%
remove test cases: 移除测试用例
reliable test grading: 可靠的测试评分
debug: 调试
uncover coverage items: 发现覆盖项
单词提取:overview: 概述
achieve: 实现
coverage: 覆盖率
regression: 回归
reduction: 减少
tests: 测试
quick: 快速的
on-demand: 随需应变的
qualifier: 限定符
gauge: 评估
distribute: 分布
uncertain: 不确定的
potentially: 可能地
target: 针对
exceed: 超过
contribute: 贡献
reliable: 可靠的
grade: 评分
debug: 调试
uncover: 发现
|