Analog in-memory compute; quantum AFM sensor.
Researchers from Politecnico di Milano, Peking University, and Hewlett Packard Labs developed a Closed-Loop In-Memory Computing (CL-IMC) chip to reduce data movement between memory and processor.
The fully integrated analog accelerator uses two 64×64 arrays of programmable SRAM cells along with integrated components including operational amplifiers and analog-to-digital converters.
“The integrated chip demonstrates the feasibility on an industrial scale of a revolutionary concept such as analogue computation in memory. We are already working on putting this innovation into use in real-world applications to reduce the energy costs of computation, especially in the field of artificial intelligence,” said Daniele Ielmini, a professor in the Department of Electronics, Information and Bioengineering – DEIB at the Politecnico di Milano, in a statement.
The ability to handle complex calculations directly in the structure of the memory enabled the test chip to achieve accuracy to conventional digital systems, but with lower power consumption, less computing latency, and a smaller silicon footprint. [1]
Researchers from TU Wien created an extremely compact parallel-plate capacitor for mechanical vibration sensing in atomic force microscopes. The quantum sensor uses a movable aluminum membrane separated from a fixed electrode by 32nm, creating an electrical resonant circuit.
“Our aluminum membrane forms a tiny capacitor together with an electrode. Combined with an inductor, this creates a resonant circuit whose resonance is extremely sensitive to any change in the mechanical vibration,” said Daniel Platz from the Institute of Sensor and Actuator Systems at TU Wien, in a press release.
The approach has better noise performance compared to optical measurement methods and eliminates the need for bulky optical components. The team also demonstrated that purely mechanical systems integrated on a chip can also be used to enable room temperature operation. [2]
[1] P. Mannocci, C. Zucchelli, I. Andreoli, et al. A fully integrated analogue closed-loop in-memory computing accelerator based on static random-access memory. Nat Electron (2026). https://doi.org/10.1038/s41928-025-01549-1
[2] I. Ignat, D. Platz, U. Schmid. High Aspect Ratio, Superconducting Vacuum Gap Capacitor NEMS with Plate Distances Down to 32 nm. Adv. Mater. Technol. 10, no. 18 (2025): 10, e01909. https://doi.org/10.1002/admt.202401909
 |
 |
 |
 |
 |  |
 |  |
Leave a Reply