News 

Follow us in real time, and we will constantly change with the world

 

Current location:
Homepage
/
/
EDA Cloud Tools Achieve Commercialization Node, Can New Models Shape the Future of IC "Core"

EDA Cloud Tools Achieve Commercialization Node, Can New Models Shape the Future of IC "Core"

  • Categories:News
  • Author:
  • Origin:
  • Time of issue:2022-08-12
  • Views:0

(Summary description)Process miniaturization is one of the most important features of the development of integrated circuit manufacturing technology. With the improvement of process capability, more devices can be integrated on chips of the same area, thereby improving chip performance and reducing unit manufacturing costs. It can be said that the progress of integrated circuit technology is aimed at improving the cost-effectiveness of integrated circuits.

EDA Cloud Tools Achieve Commercialization Node, Can New Models Shape the Future of IC "Core"

(Summary description)Process miniaturization is one of the most important features of the development of integrated circuit manufacturing technology. With the improvement of process capability, more devices can be integrated on chips of the same area, thereby improving chip performance and reducing unit manufacturing costs. It can be said that the progress of integrated circuit technology is aimed at improving the cost-effectiveness of integrated circuits.

  • Categories:News
  • Author:
  • Origin:
  • Time of issue:2022-08-12
  • Views:0
Information

Process miniaturization is one of the most important features of the development of integrated circuit manufacturing technology. With the improvement of process capability, more devices can be integrated on chips of the same area, thereby improving chip performance and reducing unit manufacturing costs. It can be said that the progress of integrated circuit technology is aimed at improving the cost-effectiveness of integrated circuits.

However, the increasing complexity of IC design has also brought corresponding challenges. For example, during the development process, the IC design team almost always faces a surge in computing resource demand, difficulty in meeting EDA peak performance requirements, consumption costs of deep process data migration, resource grabbing caused by multiple projects in parallel, and efficiency impacts caused by office location limitations. These issues can directly affect the chip's research and development cycle, and even lead to low yield and inability to mass produce the chip. In addition, a large number of servers consume significant electricity bills, and cost budgeting is also a major issue.

Therefore, IC design companies are increasingly hoping to utilize "cloud capabilities" in advanced process design to shorten turnaround time. In the view of EDA manufacturers, by transferring some or all of EDA computing to the cloud, design companies will be able to obtain flexible computing resources and economies of scale. Therefore, under the dual wheel drive of seeking higher design convergence rate and cost-effectiveness, EDA is also exploring an emerging operational model - "cloud computing+EDA".

According to Gartner's data, the global cloud computing market size in 2018 was $64 billion, and it is expected to reach $246.1 billion by the end of 2020, with a compound annual growth rate of 18% from 2019 to 2023. As the key to supporting the Internet, artificial intelligence, big data, and more, the importance of cloud computing is becoming increasingly prominent.

More importantly, cloud based IC design is becoming an inevitable trend for future development.

The combination of epidemic acceleration EDA tools and cloud computing

The EDA industry has roughly gone through three stages of development: the computer-aided design (CAD) era before the 1980s, the computer-aided engineering (CAED) era in the 1980s, and the electronic system design automation (EDA) era after the 1990s. In recent years, with the deepening penetration of cloud computing in various industries, the integration of EDA and cloud computing is also deepening. Especially in China, there are a large number of newly established small and medium-sized microchip design enterprises, which have a deeper demand for cloud based EDA tools.

According to the data of IC Design Branch of China Semiconductor Industry Association, there will be 2810 chip design enterprises in Chinese Mainland in 2021, with a year-on-year growth of 26.7%, which are widely distributed in consumer electronics, automobile, smart city and other industries. These enterprises are mostly small and medium-sized enterprises, and most of them face problems such as shortage of manpower and design capabilities. Especially when design teams conduct simulation and verification, they often lack large-scale computing power cluster support.

As pointed out by Lin Jiapeng, Senior Vice President of Guowei Sierxin, EDA cloud is the most direct solution for enterprises to solve computing power problems. Whether it's design or verification, IC design companies have a huge demand for computing power, and many small companies cannot afford it and can only trade time for money. If there is a better solution in the cloud, it will be very helpful for them.

Specifically, for most new IC companies, achieving chip casting as early as possible is a key step for their survival and development, and timely delivery of products to customers is crucial for design efficiency. In the entire design and development process, simulation and verification have become increasingly important. When chip design teams conduct simulation and validation, they often need to call large-scale computing power clusters. In such a large computing environment, the management and scheduling of the entire cluster's computing power, as well as the interaction between the computing power cluster and the storage system, also require a professional IT team to operate. EDA cloud can effectively solve these challenges.

In recent years, the outbreak of COVID-19 has had a great impact on people's working and living patterns, and home office has become very common in all walks of life. This is a promoting effect for EDA to go cloud. Many chip design companies have turned to working from home, and for chip design engineers, EDA tools are indispensable in their work. The cloud platform EDA tool is precisely an important supplement to its home office and home device design chip.

Disadvantages of EDA on cloud

One of the biggest and most obvious risks when considering transferring EDA and chip design to the cloud is security.

Chip design related documents are the intellectual property of a company, and in many cases, they can be said to be the most valuable assets of the company. Considering this, the company has always hosted all its EDA and other design files exclusively on local machines, making it inaccessible to anyone outside of its network. Cloud computing is inherently less secure than local computing. Moving to the cloud means that the company will trust other entities to host intellectual property that may be worth billions of dollars.

In addition, hosting the cloud can allow thousands of users to access the same hardware resources simultaneously. This feature may increase the risk of information being stolen or hacked by malicious opponents.

If companies can confidently migrate their designs to the cloud, they must have cloud resources that have verified end-to-end security. In the past, concerns about semiconductor intellectual property (IP) and data security have hindered EDA's migration to the cloud.

The advantages of EDA on the cloud

Despite the main drawbacks of security, migrating EDA to the cloud still has benefits.

One of the most significant advantages of cloud computing for EDA is that it allows companies to expand their computing resources up and down as needed, providing significant advantages from efficiency, economy, and results perspectives.

From an efficiency perspective, migrating EDA to the cloud is very beneficial as it provides designers with unlimited computing resources. The simulation and validation of IC design can be computationally expensive, and in some cases, it may take a whole day to complete the simulation. Local data centers are unable to quickly adapt to constantly changing design and validation workloads. By migrating EDA to the cloud, cloud computing can help improve throughput by accessing more and flexible computing resources. Designers can obtain more computing resources than other methods, allowing them to simulate and validate designs faster.

In terms of economic benefits, cloud computing for EDA is also very meaningful. Different stages of the IC design lifecycle require different amounts of calculations. Some stages, such as validation and simulation, may require a large amount of computation, while others may not require computation. This becomes an economic challenge as companies may purchase the computing required for the validation phase, but then during the less computationally intensive phase, the hardware is mostly idle. This process can be seen as a waste of resources and funds, as their computing hardware is not being used, which means their investment is not being fully utilized. Cloud computing with a "pay as needed" structure may solve this problem by allowing companies to expand or shrink their computing resources as needed. This flexibility enables companies to only pay and use the required resources at a given time, saving money by minimizing indirect costs. Companies like startups may choose to go all the way to the cloud.

On the other hand, enterprise semiconductor companies may use hybrid cloud models to enhance their local data centers. With cloud service providers providing flexible pricing models, such as on-demand pricing, design teams can expand and reduce computing resources as needed at each stage of the IP design and development cycle. Cloud computing resources can help reduce capital expenditures and reduce the cost of upgrading and maintaining data centers throughout the year. Most importantly, cloud computing can allow access to the latest available hardware, otherwise the cost of continuously upgrading local machines with the latest hardware is too high.

From the perspective of result quality, in order to meet the standards and adapt to the challenges of smaller and more advanced technology nodes, while not affecting the quality, performance, and yield of the results, designers need to conduct extensive validation to consider all potential design changes. These are highly computationally intensive workloads that require much more computing resources than the local data center can accommodate. The migration of chip design to the cloud can achieve and exceed the quality of results with higher chip yield during production runtime.

EDA Cloud Tools Achieve Commercialization Node

According to research and markets data, the global EDA market size in 2020 was approximately $11.5 billion, and it is expected to reach $14.5 billion by 2025. Among them, the proportion of cloud platform EDA tools is rapidly increasing. As a result, international EDA giants such as Synopsys and Cadence are increasingly paying attention to the process of EDA going to the cloud. Intel, NVIDIA and other chip giants are also exploring the application of EDA cloud tools.

Recently, Synopsys announced that Amazon's cloud computing service platform has deployed Synopsys' VCS FGP technology. Running related technologies in the cloud can enable design teams to achieve higher efficiency, shorten verification convergence time, and achieve excellent hardware cost-effectiveness. Regarding this, Xu Wei, Deputy General Manager of Xinsi Technology in China, said, "EDA cloud computing is a development trend, and both computing power and big data have their own advantages. More and more design companies will transition from self built private clouds to public clouds

As the public cloud architecture gradually stabilizes, the data security system gradually matures. At present, EDA cloud platform tools and operating environments are gradually integrated, and products can be replicated to different industries on a large scale and provided to customers. The integration of cloud technology's computing power, storage capacity, and EDA technology can greatly solve the computing power gap faced by current IC design, providing developers with real-time available computing power, a more flexible and efficient development environment, more optimized costs, and shortening product launch time. It can be said that the EDA cloud platform industry has reached a critical point in commercial development.

The integration of artificial intelligence and EDA continues to deepen

With the support of cloud computing technology, the integration of artificial intelligence and EDA is also constantly deepening. Optimizing customer experience and improving efficiency through the application of AI technology is an important direction for the iterative development of EDA. Algorithms such as deep learning can improve the autonomy of EDA software, improve IC design efficiency, and shorten the chip development cycle.

The report shows that the application of machine learning in EDA can be divided into four aspects: fast data extraction models; Hotspot detection in layout; Layout and route; Circuit simulation model. At present, many EDA companies have conducted in-depth layout and development in artificial intelligence. Wang Xiaoyu, General Manager of Cadence Company in China, said, "Artificial intelligence has enormous potential in areas such as large-scale digital chip optimization, digital simulation verification, and PCB design synthesis. Taking simulation verification as an example, the computational resources and time spent by enterprises on simulation verification are exponentially increasing. By using machine learning, the efficiency of productivity improvement can even reach more than 10 times

Applying AI and algorithms to their own products to achieve innovative solutions in vertical fields is a common strategy among major EDA manufacturers. In 2020, Synopsys launched its autonomous artificial intelligence application DSO.ai for chip design, which can search for optimization goals in chip design solutions and utilize reinforcement learning to optimize power consumption, performance, and area. Cadence's Cerebrus is directly integrated into the Cadence toolchain, from System C definition to standard library units, macros, RTLs, and signOff, allowing an engineer to define specifications and optimization objects at any level. Siemens EDA's Solido products can utilize machine learning to quickly generate and extract feature vector libraries, achieving higher validation accuracy in less time, and presenting the resulting data in a visual manner.

Scan the QR code to read on your phone

Related News

Address: No. 155 Guangming Street, Fuxi Street, Deqing County, Zhejiang Province

Contact person: Miss Tang

Tel:13706824925 (same as WeChat)

Email: info@gigelectronic.com

 

  

Gaige Electronics
Gaige Electronics
Gaige Electronics
Gaige Electronics
Gaige Electronics
Online customer service
Customer service hotline
13706824925 13706824925
Service time:
9:00 - 18:00
Customer service group:
Online Service

© 2023 Deqing Gaige Electronics Co., Ltd.     Powered by www.300.cn   SEO  Business License

Gaige Electronics