What's new

China makes breakthrough in system-integrated memristor computing-in-memory chips

beijingwalker

ELITE MEMBER
Joined
Nov 4, 2011
Messages
65,195
Reaction score
-55
Country
China
Location
China
China makes breakthrough in system-integrated memristor computing-in-memory chips

By Global Times
Published: Oct 10, 2023 05:45 PM Updated: Oct 10, 2023 05:40 PM

Photo: Courtesy of the research team from the Tsinghua University

Photo: Courtesy of the research team from the Tsinghua University

China's efforts to ramp up semiconductor innovation seem to bear more fruits as the Tsinghua University has successfully developing the world's first fully system-integrated memristor computing-in-memory chip that supports efficient on-chip learning, which is also energy efficient.

The chip, though still in the laboratory phase, is expected to promote development in artificial intelligence (AI), autonomous driving, wearable devices and other high-tech fields.

Developed by a team in the School of Integrated Circuits in Tsinghua University, the advanced chip has an cutting-edge advantage as it consumes energy that is only 3 percent of an application-specific integrated circuit (ASIC) system for on-chip learning, which will greatly meet demand for increased computing efficiency in the era of AI, said the research team from the university.

The research paper has been published in the September edition of "Science."

It took the team about 11 years to complete the job starting from the research on using memristors for storage application. Memristor is the fourth basic circuit component after resistors, capacitors, and inductors. It can "memorize" the passing charge after the power is turned off, and is used as a new type of nano-electronic synaptic device.

Based on nanometer-scale memristor, the computing-in-memory architecture is like a new working model of "working from home", which completely eliminates the energy consumption in commuting for data, and greatly enhance the operating speed, Yao Peng, the co-first author of the recently published paper, told the Global Times on Tuesday.

"Such advantages could have broad application in future edge computing and cloud computing," Yao said.

In addition to lower latency and smaller energy consumption, the chip can effectively protect user privacy and data security, according to Yao.

The memristor storage and computing chip is another example of China's unwavering efforts over recent years to strengthen semiconductor research and development, which is facing suppression from the US government.

China has rolled out a number of favorable policies including tax reductions and funding support to accelerate the development of its semiconductor sector, shoring up self-reliance in chips innovation and tech advancement.

In June, a subsidiary of China Electronics Technology Group Corporation (CETC) achieved the 28-nanometer full manufacturing process of ion implanter -- a leading equipment in chip manufacturing. At present, the 28-nanometer manufacturing process is a mature procedure with the widest coverage in the field of chip application.

The outline of the 14th Five-Year Plan (2021-25) for National Economic and Social Development and the Long-Range Objectives through the Year 2035 stated China will speed up the development of high-end chips.

 
.

China’s computational power gains new strength with 255-detected-photon quantum computer​

By Xinhua News Agency

Chinese scientists unveiled a quantum computer prototype named “Jiuzhang 3.0” with 255 detected photons on Wednesday, once again pushing the boundaries of photonics quantum computing technology.

Led by the renowned Chinese quantum physicist Pan Jianwei, the research team has successfully accomplished this quantum computing feat, achieving a speed that is 10 quadrillion times faster in solving Gaussian boson sampling (GBS) problems compared to the world’s existing fastest supercomputers.

Gaussian boson sampling, a classically intractable problem, was employed in this study to provide a highly efficient way of demonstrating quantum computational speedup in solving some well-defined tasks.

The study was published online in the journal Physical Review Letters on Wednesday Beijing Time.

Lu Chaoyang, a member of the research team and professor at the University of Science and Technology of China, said that a series of innovations, including a newly developed superconducting nanowire single-photon detection scheme with fiber loop-based configuration, increased the number of detected photons for “Jiuzhang 3.0” to 255, greatly improving the complexity of photonics quantum computing.

“By demultiplexing photons into time bins through delays, we’ve achieved capabilities of pseudo photon number resolving,” Lu added.

According to the state-of-the-art exact classical simulation algorithm, “Jiuzhang 3.0” is a million times faster at solving GBS problems than its predecessor, “Jiuzhang 2.0.”

Moreover, the most complex samples of GBS that “Jiuzhang 3.0” can calculate in just one microsecond would take the world’s fastest supercomputer, “Frontier,” more than 20 billion years to complete.

In 2021, the team led by Pan developed the “Jiuzhang 2.0” with 113 detected photons and a 66-qubit programmable superconducting quantum computing system named “Zuchongzhi 2.1,” making China the only country to achieve a quantum computational advantage in two mainstream technical routes — one via photonics quantum computing technology and the other via superconducting quantum computing technology.

Establishing quantum computational advantage requires great endeavor, with long-term competition between classical algorithms and quantum computing hardware, the team noted.

They anticipate that this work will, on one hand, stimulate more research on classical simulation algorithms, and on the other hand, through diligent efforts, gradually address various scientific and engineering challenges in quantum computing research. Ultimately, quantum computers will achieve computational power beyond the reach of classical computers, driving the advancement of science and technology.

 
.
Back
Top Bottom