The ISSCC/JSSC Journey of Chinese Chips
Latest update time：2020-06-02
Reads：
Source: This article is reprinted from ICAC Workshop by Semiconductor Industry Observer, thank you!
The integrated circuit industry has never been as widely concerned and has never triggered large-scale discussions as it is today. This is due to both the national strategic background of the current Sino-US competition and the stage of technological development. When the Shockley Laboratory was established in California, it was already the mid-1950s. It can be said that China has no historical baggage in this field. During the Cold War, the US-Soviet hegemony, integrated circuits did not become the focus, because the integration of electronic systems was far from the current level. In fact, until the turn of the century, there were still quite a lot of discrete modules and devices on the single standard 2G mobile phone motherboard. Now looking back at that time, paper letters still carried a lot of daily communication, Moore's Law planned the step-by-step evolution of semiconductor technology, and the entire integrated circuit industry continued to move towards higher, faster and stronger under the guidance of the Olympic spirit. At that time, I had just entered the industry, and the industry legends I read were mostly geniuses with personality, who only cared about making the best chips and did not care too much about the needs of applications and systems. Today, the integration of electronic systems is close to the point where there is nothing to do except chips. For example, for mobile phones, the cost can be calculated by counting chips and important components. Therefore, technically speaking, the barriers in the hardware field have been concentrated on integrated circuits. In recent years, the United States has been stepping up its crackdown on Huawei and even the domestic high-tech industry by blocking chips and their industrial chains. Firstly, it has the successful experience of strangling Japan in the 1980s and 1990s, and secondly, it does not have many other options.
Global system and industry-academia symbiosis
The current situation is certainly not good, but if you think about it calmly, it is not too bad. Integrated circuits follow Moore's Law and develop rapidly, and have high capital, talent and technology thresholds. Since the 1990s, the country has actually tried to invest in the development of integrated circuit manufacturing many times, but most of them fell behind as soon as they were built, and have not caught up to this day. But in recent years, Moore's Law has gradually slowed down, and the entire industry has slowed down, and our efficiency and chances of catching up have greatly increased. In addition, the country's demand and investment have also attracted a large amount of capital and talent to invest in this industry. There is not only national capital such as the Big Fund, but also a large amount of private capital. There are not only graduates from microelectronics colleges in various places, but also continuous overseas talent returning. Then the remaining key is to complete the accumulation of technology, including the localization of cutting-edge technology and leading research. But this is actually very, very difficult, because the integrated circuit industry is a giant ecosystem covering multiple links from materials to devices, from design to manufacturing, from systems to applications. At present, no country has the ability to build such an industrial system alone. For example, the EUV lithography machine, which is always mentioned as a bottleneck, can only be manufactured by ASML of the Netherlands. Among them, the key component of the optical system, the multi-layer film reflector, can only be manufactured by Zeiss of Germany. Its flatness requirement is equivalent to the elevation fluctuation of no more than 1 mm within the entire area of Germany. Many students who pay attention to photography will notice that the lens seems simple, but in fact, even the copycats are not well made now, and this is just the tip of the iceberg of the integrated circuit industry system. For this kind of technology that is not well made, the daily mentality of practitioners is actually to be cautious and tread on thin ice, so that many senior engineers finally grow into a delicate and steady professional character. Therefore, it is good not to overturn on the curve, and the possibility of overtaking on the curve is very low. It is more realistic for the entire industrial system to perform its duties and continue to work hard for several years.
Back to technology, it is much more pure. Organize a group of competent engineers in a reasonable way, give them enough resources and patience, and technical success is not a puzzle. HiSilicon is a good example. In fact, in the history of the development of integrated circuit technology, Chinese have been making great contributions. Professor Sa Zhitang, one of the proposers of CMOS technology, has participated in the germination and development of the industry since the Shockley Laboratory and Fairchild period; Professor Shi Min proposed NVM; FinFET and FDSOI proposed by Professor Hu Zhengming are the mainstream technologies in the nano era. In the field of design, Professor Huang Qiuting was selected as one of the top ten contributors of the 50th anniversary of ISSCC on behalf of Chinese in 2003. It is very interesting that several of them were born in mainland China, spanning academia and industry, and finally returned to academia. This is also one of the characteristics of the integrated circuit industry. The relationship between academia and industry is very close. Judging from the development history of FinFET, even the highly resource-intensive manufacturing industry is still constantly looking for new technologies in academia. In the field of integrated circuit design, this situation is even more common. Represented by Qualcomm and Broadcom, many famous fabless design companies have a deep academic background, and even a considerable number of them are directly started by university professors. In China, the industrialization efforts of outstanding scholars from Tsinghua University and Fudan University have also achieved remarkable results. Due to this natural symbiosis, in the design industry, the level of academic research actually directly reflects the development status of the overall industry.
ISSCC and JSSC
As we all know, the highest level of results in the field of integrated circuit design are usually published in ISSCC (International Solid-State Circuits Conference) and JSSC (IEEE Journal of Solid-State Circuits), which are known as the "Olympics of Chips". ISSCC is only two pages long and does not need to reveal too many technical details. Therefore, it is traditionally a stage for both academia and industry to showcase the latest results. Work published in ISSCC usually requires the best and/or the first. Some milestone products in the history of integrated circuits were first announced at ISSCC. JSSC was originally established to publish extended long articles of ISSCC. The review process is more rigorous, and authors are allowed to have technical discussions with reviewers. Later, several major conferences that are slightly inferior to ISSCC, such as VLSI Circuits, CICC, ESSCirC and ASSCC, will also publish excellent papers in JSSC. In addition, JSSC also accepts direct submissions. Due to the strict review, articles published in JSSC are usually more solid. It can be said that JSSC is a detailed and serious technical accumulation and archive in this field. There is no doubt that the number of papers published in ISSCC and JSSC is a direct measure of academic level and also indirectly reflects the overall development level of the design industry.
Here I recall a memorable event. In early October 2017, I visited MIT with a school delegation and met with Professor Anantha Chandrakasan, the newly appointed Dean of the MIT School of Engineering. He previously served as the head of the EECS Department and is also a member of the American Academy of Sciences and the National Academy of Engineering. After we introduced the glorious history and great achievements of Chengdu University of Electronic Science and Technology, especially when we mentioned that electronic science and technology ranked first in the country, he asked a question: How many ISSCCs do you have? There was a silence. Perhaps thinking that we didn't hear clearly, he explained again, saying that he knew that Tsinghua and Fudan had a small number of ISSCCs, and wanted to know how many papers the best electronic engineering majors in China had. As a contact person and a junior colleague, I could only give that embarrassing answer truthfully. Even now, I am still embarrassed when I think about it. Scholars and managers of this level certainly know CNS, and MIT certainly has no shortage of CNS, but it is obvious that he still regards ISSCC/JSSC as a yardstick for measuring integrated circuits and even the entire electronic engineering.
The Road to ISSCC/JSSC in China
Back to the topic, ISSCC/JSSC mentioned above also reflects the development level of the entire design industry. Correspondingly, mainland my country rarely published ISSCC or JSSC before 2000. The first ISSCC paper should be the PLL published by Dr. Chang Zhongyuan and others from IDT Xintao in 2005. The following year, the team of Professor Wu Nanjian from the Institute of Semiconductors published the first ISSCC in mainland academia, titled PLL frequency synthesis. In 2007, Dr. Li Zhenbiao and others from Dingxin published the third ISSCC, titled TD-SCDMA transceiver of my country's independent 3G standard. Later, Dr. Yu Xueyi and others from the team of Professor Li Yugen and Professor Wang Zhihua of Tsinghua University published two consecutive ISSCCs in 2008 and 2009. Dr. Lu Lei and others from the team of Professor Tang Changwen of Fudan University also published ISSCC in 2009, and the topics were all related to frequency synthesis. In addition, the team of Professor Hong Zhiliang of Fudan University published ISSCC papers on power management in 2011 and 2014, the Loongson team of the Institute of Computing Technology of the Chinese Academy of Sciences published ISSCC papers on processors in 2011 and 2013, and the teams of Professor Zeng Xiaoyang and Professor Yu Zhiyi of Fudan University published ISSCC papers on processors in 2012 and 2013. However, it was not until 2016 that mainland China began to publish ISSCC papers every year.
In terms of JSSC, early domestic work includes the full adder published by Ningbo Normal University and Hangzhou University in 1992, the III/V process op amp published by the Thirteenth Institute in 1995, and the PLL frequency synthesizer published by Dr. Yang Chonghe of Shanghai Belling in 1997. In 2004, Professor Liu Leibo of Tsinghua University and others published papers on JPEG2000 encoder, in 2006 Professor Huang Fengyi of Southeast University and others published papers on modeling, in 2008 Professor Wu Nanjian's team published papers on visual chips, and Professor Zeng Xiaoyang's team published papers on ADC. However, it was not until 2013 that JSSC papers began to be published every year in mainland China. Since 2018, the number of ISSCC and JSSC papers in mainland China has shown a rapid growth trend, in sync with industrial development.
The above statistics do not include Taiwan, Hong Kong and Macau. Compared with the mainland, Taiwan and Hong Kong started much earlier in design, and Taiwan has unique foundry advantages. Therefore, the explosive curve of ISSCC and JSSC is much earlier than that of the mainland. For example, the number of ISSCC papers published by NTU ranked first among universities in the world for three consecutive years from 2005 to 2007; in 2007, 11 ISSCC papers even ranked first among all units in the world, and Professor Li Zhiyi won the ISSCC Beatrice Best Paper Award. Hong Kong actually lacked sufficient industrial support after the abortion of HKSTC that year, but thanks to its special open status, the huge investment in the establishment of HKUST, and the recruitment of several top North American doctoral students at that time, it also achieved a glorious period in ISSCC and JSSC. The most noteworthy is actually Macau. In an environment with almost no industrial soil and no talent base, Professor Ma Xuyuan, a Portuguese who does not understand Chinese, has worked hard for more than 20 years to build a world-class integrated circuit design team from scratch. In recent years, the number of ISSCC papers published by the University of Macau has been among the highest in the world, and the number of papers on analog and radio frequency is believed to be the highest in the world. Professor Lu Yan's team also won the 2017 ISSCC Far East Best Paper Award. These achievements are not only the pride of Macau and China, but also a miracle in the development of the world's integrated circuits. In a sense, the success of IC design at the University of Macau is as significant as the success of HiSilicon.
Judging from the ISSCC/JSSC papers, IC design in Hong Kong, Macao and Taiwan is already at the world-class level, while the mainland is still in the stage of key breakthroughs and steady growth. Some international research teams have begun to emerge, but the development levels in different directions are still quite different. For example, in the direction of low-power AI chips, all three papers at ISSCC this year (2020) came from mainland universities, including two papers from Professor Liu Yongpan's team at Tsinghua University and one paper from Professor Shan Weiwei's team at Southeast University, reflecting a clear advantage. In recent years, there have been continuous articles published in the RF direction at ISSCC. However, one of the most traditional directions in the field of integrated circuits, and also the current bottleneck problem in my country, ADC, the mainland did not publish the first ISSCC paper until this year, which was completed by Dr. Liu Jiaxin of Tsinghua University and Professor Sun Nan's team at UT-Austin. The amplifier published by our group this year is a basic analog circuit, which is also the first in the mainland academic community. In addition, due to the scarcity of IC engineers and concerns about technical confidentiality, the trend of publishing papers and technical exchanges in the domestic industry is far from being formed; the industry-university interaction and cooperation model that adapts to national conditions also needs to be explored. Referring to the development experience of foreign countries and Hong Kong, Macao and Taiwan, these are the inevitable requirements for the healthy growth of the IC design industry.
About the author:
Li Qiang, professor at the University of Electronic Science and Technology of China, started working in integrated circuit design after graduating from undergraduate school in 2001. He published the first JSSC analog RF paper in Singaporean universities for his doctoral work. He was a tenured associate professor at Aarhus University in Denmark and currently directs the Institute of Low Power Integrated Circuits and Systems at the University of Electronic Science and Technology of China. Since 2018, he has published 11 ISSCC/JSSC papers.
*Disclaimer: This article is originally written by the author. The content of the article is the author's personal opinion. Semiconductor Industry Observer reprints it only to convey a different point of view. It does not mean that Semiconductor Industry Observer agrees or supports this point of view. If you have any objections, please contact Semiconductor Industry Observer.
Today is the 2328th issue of content shared by "Semiconductor Industry Observer" for you, welcome to follow.
Recommended Reading
★ New opportunities for domestic EDA
★ Texas Instruments' growth path as an analog giant
★ Taiwan's semiconductor strength that you may not know
Semiconductor Industry Observation
" The first vertical media in semiconductor industry "
Real-time professional original depth
Scan the QR code , reply to the keywords below, and read more
Two Sessions｜TSMC｜RISC-V｜Automotive Chips｜AI｜EDA｜China and the US｜Wafer｜RF
Reply
Submit your article
and read "How to become a member of "Semiconductor Industry Observer""
Reply Search and you can easily find other articles that interest you!
京公网安备 11010802033920号