AI COMPUTE•NVIDIA•GPU CLOUD•DATA CENTERS•SEMICONDUCTORS•ENERGY•FUNDING•M&A•AI COMPUTE•NVIDIA•GPU CLOUD•DATA CENTERS•SEMICONDUCTORS•ENERGY•FUNDING•M&A•
GPU·DATA HUB
Newsletter
Semiconductors

Security-Aware Layer Mapping Reduces Fault Vulnerability In 3D Compute-In-Memory (ASU, U. of Maryland)

Security framework targets multi-tenant vulnerabilities in 3D compute-in-memory chips

By GPU Data Hub Desk · GPU Data Hub·Editorial standards··United States

Academic researchers from Arizona State University and the University of Maryland have introduced a layer-mapping scheme called SEAL to mitigate security risks in heterogeneous 3D compute-in-memory architectures. The method seeks to prevent fault vulnerabilities arising from shared power delivery networks during multi-tenant deep neural network inference workloads.

Original source

This summary was written by the GPU Data Hub desk from reporting published by Semiconductor Engineering on 25 Oct 2026, 20:30. Read the full article at the original publisher.

Read Original Source →

By subscribing you consent to receive the daily briefing. Unsubscribe at any time. See our privacy policy.

LinkedInX

More from GPU Data Hub

Latest →

Pt-W Alloy Shows Potential For Thinner High-NA EUV Mask Absorbers (Chinese Academy of Sciences et al.)

Researchers at the Chinese Academy of Sciences and affiliated institutions have published technical work evaluating a platinum-tungsten alloy for high-NA extreme ultraviolet mask absorbers. The research suggests the alloy could allow for thinner absorber layers, supporting improved imaging performance during lithography. Advancements in absorber materials remain vital for scaling optical performance in future semiconductor fabrication nodes.

Semiconductors··Semiconductor Engineering

Dry-Deposited Resist Shows Different Tone Behavior Under EUV And E-Beam Exposure (KU Leuven, Kunsan, VUB, Imec)

A research consortium including Imec and KU Leuven has analysed the radiation response of dry-deposited metal-organic resists under extreme ultraviolet and electron-beam lithography. The study reveals distinct tone conversion behaviour depending on the exposure source employed during the patterning cycle. These findings offer useful insights for developing new resist chemistries required for advanced semiconductor manufacturing.

Semiconductors··Semiconductor Engineering

Silicon Photonic Modulator Reaches 200 Gb/s PAM4 With Reduced DSP (Science Tokyo, Yokohama, AIST, PETRA)

Researchers in Japan have created a low-voltage silicon photonic modulator that achieves 200 Gb/s PAM4 transmission while lowering digital signal processing overheads. The compact device integrates an electro-optic equaliser to support high-speed optical transceivers, catering to the increasing bandwidth requirements of artificial intelligence data centres.

Semiconductors··Semiconductor Engineering

Neural Networks Speed 3nm Nanosheet SPICE Simulation (TU Munich)

Researchers at the Technical University of Munich have deployed neural network models to accelerate SPICE simulations for 3nm nanosheet devices. The study investigates deep learning techniques to expedite transistor characterisation and standard cell library development for advanced semiconductor processes.

Semiconductors··Semiconductor Engineering

Commerce Dept. Plans Equity Stakes in 4 Companies Under $389M CHIPS R&D Proposal

The US Department of Commerce has issued letters of intent to allocate up to $389 million in federal research subsidies across four enterprises, with the government intending to take equity positions. Distributed through the CHIPS R&D Office, the capital aims to accelerate domestic development of semiconductor materials vital to data centres and energy networks.

Semiconductors··HPCwire