How AI helps Siemens EDA realize accelerated creation of "core"?

If the global economic problems make every link of the semiconductor industry face challenges, then as the top of the industry chain - EDA software must have undertaken the most pressure and hope. how is the market environment in the AI era, and what are the challenges that the semiconductor industry will face?

There have been four waves of downward nodes in the industry in 20 years: the economic crisis, the U.S. financial crisis, the new crown epidemic, and today's downward pressure on the economy in the post-epidemic era.



The first wave was the U-shaped recession, which was the bursting of the Internet bubble, and it took several years for the industry to slowly (2002/2003) recover from the U-shaped recession. The second wave was the U.S. financial crisis, when the EDA industry wondered if it would lead to a prolonged winter in the semiconductor industry. Later, it was proved that this V-shaped, although the decline was very fierce, but also quickly recovered, due to the state of the world economy, the supply chain situation, China's booming industry, the United States re-capitalization, high-tech "flywheel effect", and soon found a new direction of investment, and so on.



The arrival of the epidemic at the end of 2019, so that the entire semiconductor industry to undertake an unprecedented pressure. Ling Lin said that this is an epidemic affecting all of humanity, resulting in a lot of inventory squeeze, but also affects the judgment of the supply chain, everyone is stocking up and rush to buy, the industry as a whole shows a U-shaped trend. Now that the epidemic is over, the semiconductor industry has not gone all the way up as expected. Ling Lin believes that this fourth round of recession, more because the overall supply chain is particularly affected by geopolitical and other factors, or because of the need to reshape the supply chain, but also includes the previous stocking up too much, stocking up for the wrong reasons.



However, because of the first three "recession", the entire semiconductor industry in fact does not present a pessimistic stance. Ling Lin said: "Predicting the future need to look at history. Chinese people are more intelligent, the Chinese people look at the future is to look at the longer history. I think there is some truth to this, we must look back." He cited the example that the current overall market is embracing intelligent transformation, or industrial upgrading. Many data show that the current automotive and industrial growth is rapid, and practitioners should not worry about whether the future growth, but which points will grow in the future. It can be expected that the future of cloud computing, data centers, wireless communications, industrial electronics and consumer electronics have a relatively large growth, and then the overall expected by 2030 the entire semiconductor industry will exceed 1 trillion. So stay hopeful that there is more we can do in the future. Ling Lin also analyzed the market trends in the semiconductor industry from an EDA perspective. Siemens EDA analyzed the types of industries that startups are investing in, and concluded that "edge computing," "cloud computing," and "vision/facial recognition" are highly probable to be the hot technologies of the future. Hot technologies.



To be launched, in order to accumulate momentum. Ling Lin said that Siemens EDA software also maintains confidence in the industry, but also to keep their own development in three directions: 

(1) more advanced process technology. Even if Moore's Law slows down, Siemens EDA still think new technology will still come out, just may be a little slower. Also making up for this with heterogeneous integration will give rise to many new technologies. 

(2) Design scale up. 

(3) Chip-system-product. Siemens EDA not only look at the chip inside, SoC, but also see the chip connected to other chips, including the flow of signals on the CPV board, put it in the system of systems, or even a super system, the performance can not reach the goal. The whole from the chip to the system to the product digital twin, Siemens EDA to provide a more complete and comprehensive vision to look at the problem, there is a unique solution, which goes beyond the general definition of EDA, and it is in the whole pervasive, global AI under the guidance. Siemens EDA will not be in a particular, but in the various algorithms inside the AI to do this tool, try to use machine learning methods to do rapid deployment and get results.


AI empowers the whole industry chain

At present, Siemens EDA already has many AI technologies, including "yield improvement", "library building", "digital verification", etc. Building library is to reduce this "library" by 100 times-1000 times through machine learning, and quickly run through this "library". Building library that is through machine learning to reduce this 100 times - 1000 times, very quickly to the "library" run through.Lincoln Lee said, with the convenient AI tools, the previous run, do not need to re-run, we do not need to waste time, just run the changed place is good. These are the methods of machine learning, to improve our current tool capabilities.



At the same time, Siemens EDA has maintained a momentum of acquiring excellent machine learning companies around the world. Ling Lin said that since Mentor joined Siemens in 2017, Siemens has done 11 mergers and acquisitions in the EDA field, which are distributed globally and encompass all aspects of the scope. Last year Siemens launched the platform Siemens Xcelerator, Siemens as a technology company, as the name suggests, is to help enterprises that now want to digital transformation to accelerate the process of digitalization and intelligence, and EDA is a very core component of this platform. From IC design tools such as micro, to software, hardware, such as macro, Siemens is trying to achieve complete coverage.



In fact, the pillar of the digital economy is to rely on the evolution of semiconductors, integrated circuits, that is, rely on Moore's law + super Moore's law to do, EDA is a very important "point", which is a very important link. Ling Lin said: "As Siemens EDA, we will focus on our own business, to do a good job, feeding the entire Siemens solution closed loop robustness."


Multi-pronged approach to promote the development of the semiconductor industry



Horizontally, in the field of EDA software, Siemens EDA's biggest advantage is that it has a huge digital industrialization software group as a "backing", and provides customers with full-process services through systematic solutions. In this conference, Siemens also brings more practical tools for the whole semiconductor industry.



(1) Calibre DesignEnhancer



Siemens Digital Industry Software has launched the innovative Calibre® DesignEnhancer software, which enables integrated circuit (IC), layout and routing (P&R) and full-customization design teams to automate the "Calibre DesignEnhancer" early in the IC design and verification process. layout modification optimization early in the IC design and verification process, dramatically increasing productivity, improving design quality and reducing time to market. The tool offers three usage models, including Vias Correction, Power/Ground Enhancement, and Fill Cell Insertion.



(2) Solido Design Environment Software



Officials said that the Solido Design Environment uses artificial intelligence (AI) technology to support cloud-based integrated circuit (IC) design and verification, which can help design teams to cope with increasingly stringent power, performance, yield and reliability requirements, and accelerate the speed to market. It is understood that some Siemens EDA customers have already started to use the Solido design environment software, SK hynix Inc. has used the Solido design environment software to significantly reduce the overall production time.



(3) Questa Verification IQ software



Official information shows that this software provides groundbreaking solutions for the complex design challenges of next-generation high-performance integrated circuits (ICs). It is a team-based, cloud-based software that is data-driven and employs artificial intelligence (AI) technology to help accelerate verification convergence, simplify traceability, optimize resources, and facilitate speed-to-market. Many leading chip design and verification companies are utilizing big data to drive innovation, streamline operations and improve efficiency," said Abhi Kolpekwar, vice president and general manager, Design Verification Technology, Siemens EDA. Siemens' new Questa Verification IQ, a data-driven verification solution that leverages analytics, collaboration and traceability to accelerate and streamline the verification process, will achieve significant results when used with Siemens' Polarion REQUIREMENTS software."



EDA, the most fundamental and front-end area of the semiconductor industry, can provide downstream customers with a more easily utilized platform and broader range of tools. Siemens EDA can reach customers' goals more easily by combining it with its own digital industrial software.