9/26/2024 Cassandra Smith 4 min read
Electrical and Computer Engineering professor Naresh R Shanbhag received national recognition for technology his team patented a few years ago that has had major impact on the world of semiconductors. The Semiconductor Research Corporation (SRC) presented Shanbhag with the 2024 SRC Innovation Award.
Written by Cassandra Smith
Electrical and Computer Engineering and Coordinated Science Laboratory professor Naresh R Shanbhag received national recognition for technology his team patented a few years ago that has had major impact on the world of semiconductors.
The Semiconductor Research Corporation (SRC) presented Shanbhag with the 2024 SRC Innovation Award. This recognition recognizes the lead faculty whose SRC research led to patents and IP that have significantly impacted the semiconductor industry. This year's award is for the patent US 9687877B2: “Compute Memory” by Professor Naresh Shanbhag and his doctoral student Mingu Kang, presently a faculty at the University of California at San Diego.
Recognition from the SRC holds much value, according to Shanbhag. “It brings together academia and a consortium of semiconductor companies to facilitate collaborative research and provides research funding for faculty members here in the USA as well as globally.”
Shanbhag researches the design of energy efficient semiconductor integrated circuits and systems for machine learning and communications. At the time of patent filing in 2015, he was director of the Systems on Nanoscale Information FabriCs (SONIC) Center where SRAM-based in-memory computing was first proposed by Shanbhag and Kang, and on which the Compute Memory patent is based.
SONIC was funded by the SRC and the Defense Advanced Research Projects Agency (DARPA). SONIC’s mission was to build energy-efficient integrated circuits and systems for artificial intelligence. This multi-university research center comprised nearly 30 professors and more than 100 graduate students from all over the United States. The center concluded its mission in 2017.
“The chips that are built using in-memory computing technology achieve the state-of-the art in terms of energy efficiency and compute density, meaning you can pack many computations in a small area. In terms of density of computation and energy efficiency, there is no question that this technology shines. When we proposed the idea in 2014 paper, it was very clear to me that in-memory computing was going to be an important technology because of its relevance to artificial intelligence, so we filed a patent,” said Shanbhag.
The patent, which was awarded in 2017, was based on Mingu’s doctoral research. Mingu received the CSL Outstanding Dissertation Award in 2018 after he graduated.
“CSL was very forward-looking in that respect,” said Shanbhag. “They could see already the impact of his thesis. Since 2017, industry has been actively trying to integrate this technology into their products. There are multiple start-up companies that are looking to commercialize this technology.”
According to Shanbhag, the number of researchers, globally, working on this technology has grown substantially based on in-memory computing articles published in all the major circuit conferences. The patent was cited more than 20 times in subsequent filings by other researchers’ who have conceived of variants of the idea. Three doctoral students in Shanbhag’s group who followed Mingu – Sujan Gonugondla, Hassan Dbouk, and Saion Roy – also contributed to developing in-memory computing further.
Seeing the success of this project brings Shanbhag great pride: “I’m proud of what my students have accomplished. They are amazing.” In his acceptance speech, Shanbhag also highlighted the honor of sharing the moment with another former doctoral student, Yingyan Lin, now a faculty at Georgia Tech, who was present to accept the 2024 SRC Young Faculty Award.
Shanbhag said he is happy to see their work being recognized not only for themselves, but those who invested in them. “The SRC and DARPA are seeing the benefits; one of the outcomes of their funding is this unique technology that has spread like wildfire,” he said. “I really appreciate their strong support through SONIC and subsequent programs such as the DARPA Foundations of Novel Compute (FRANC), SRC AI Hardware (AIHW), and through the C-BRIC, COCOSYS, and CUBIC research centers.”