Never miss a story from Cadence Academic Network. Subscribe for in-depth analysis and articles.
The International Symposium on Physical Design (ISPD) contest is a well-known competition in the EDA field, where the main idea is to have EDA companies share the industrial problems they are facing with the academic community to drive practical research. At the beginning of the contest, the hosting company proposes a problem and releases a set of benchmarks associated with the problem, and then researchers from all over the world can register for the contest and work on the problem. After a couple of months, the teams who come up with the best solutions are invited to present at the International Symposium on Physical Design.
The 2018 and 2019 ISPD contests were organized by Cadence researchers and developers: Gracieli Posser, Wen-Hao Liu, Stefanus Mantik, William Chow, Yixiao Ding, and Amin Farshidi. The proposed problem for the contest was the initial detailed routing, which is a critical problem for advanced-node enablement, like 7nm, 5nm, and 3nm technology nodes. Therefore, one of the objectives of the contest was to drive practical detailed routing researches to consider real design rules, memory scalability, and runtime scalability. The 2019 ISPD contest added more challenging rules, bigger benchmarks, and harder runtime and memory constraints compared to the 2018 contest, making the problem even closer to the real-world industry challenges. These contests give future engineers insight into what challenges they’ll be facing in their careers and give us the opportunity to attract those talented students to work in the EDA field.
This year, the contest had 33 teams, which is one of the highest rates of participation since the first ISPD contest starting in 2005. It’s exciting to see the number of interested students, especially since this is the second year the contest covered the same topic. Highlights:
The organizers greatly appreciated the huge support provided by the Cadence Academic Network. As the contest results were evaluated using Cadence's Innovus Implementation System, all participants needed Cadence licenses, which were provided to them through Cadence Academic Network. If any student or professor wants to continue their research using the benchmarks released by this contest, they need to use Innovus as well; Cadence Academic Network can help them with the licenses. If a university already has Cadence licenses, contest benchmarking will allow for the students to continue to work on the problem and grow as an Innovus user in academia.
Cadence and the Academic Network were also supportive with promoting the ISPD contest and ISPD symposium on social media. It helped a lot to increase the number of participants and bring more interested people to the EDA domain.
Cadence presented the contest problem and benchmarks used in the contest through a published paper. The contest paper usually largely impacts the EDA community and will drive the next couple of years of research in this area. For example, below there is a list of the citation numbers of the previous contest papers. Note that, in the EDA field, a paper is considered as a high-impact paper if it has more than 10 citations.
The ISPD conference was held April 14 - 18, 2019 in San Francisco, California. The paper was presented during the conference, by Gracieli Posser, William Chow, and Wen-Hao Liu
During the presentation, the contest problem, benchmarks, and evaluation were introduced and discussed. After that, the top 5 ranking teams were announced and each of them had a short presentation to introduce their innovative solution to the contest problem:
The Chinese University of Hong Kong (Hong Kong)
Gengjie Chen, Haocheng Li, Bentian Jiang, Jingsong Chen, Evangeline Young
National Taiwan University (Taiwan)
Chen-Chia Chang, Chia-Ming Chang, Wei-Kai Liu, Chen-Hao Hsu, Yao-Wen Chan
Kim & Lee
Daeyeon Kim, Sung-Yun Lee, Minhyuk Kweon, Seokhyeong Kang
Fuzhou University (China), National Tsing Hua University (Taiwan)
Zhen Zhuang, Chien-Hao Tsou, Weida Zhu, Chao-Yuan Huang, Genggeng Liu, Wenzhong Guo, Ting-Chi Wang
Universidade Federal de Pelotas (Brazil)
Stephano Gonçalves, Felipe Marques
Gracieli Posser (contest chair) and Ismail Bustani (ISPD general chair) with the first place group
Gracieli Posser (contest chair) and Ismail Bustani (ISPD general chair) with the second place group
Gracieli Posser (contest chair), Wen-Hao Liu (past contest chair) and Ismail Bustani (ISPD general chair) with the third place group
Congratulations to all teams for the great work!
This contest presents one of the critical problems Cadence is researching with academia, so all the contestants are working on a relevant problem that Cadence cares about. It gives those future engineers hands-on experience tackling a real-world problem and shows them what it would be like to be a Cadence employee. We have been fortunate enough to attract some of the top talent that participated in previous ISPD contests. Participating in the ISPD contest is a fun way for Cadence to have industry problems solved and engage with the next generation of engineers. We look forward to playing a role in the ISPD contest in the years to come.
Learn more about the ISPD contest.