Arm’s TCS23; AI partnerships; nickel aluminide; DVFS vulnerability; Cadence acquires Pulsic

Open-source software for an open-source ISA: the Risc-V Software Ecosystem (RISE) Project is a new initiative dedicated to enabling a software ecosystem for application processors that includes software development tools, virtualization support, language runtimes, Linux distribution integration, and system firmware, working upstream first with existing open-source communities in accordance with open-source best practices. The RISE Governing Board includes Andes, Google, Intel, Imagination, MediaTek, Nvidia, Qualcomm, Red Hat, Rivos, Samsung, SiFive, T-Head, and Ventana.

New Arm mobile computing platform

At the recent Computex event in Taiwan, Arm announced its Total Compute Solutions 2023 (TCS23), a new platform for mobile computing aimed at premium smartphones. TCS23 IP includes a new Arm Immortalis GPU based on the 5th Generation GPU architecture, a new cluster of Armv9 CPUs supporting artificial intelligence, and other enhancements. A new graphics feature introduced in the 5th Gen GPU architecture is Deferred Vertex Shading (DVS), a technique that redefines the dataflow and enables partners to scale for larger core counts and higher performance points. A key part of the CPU cluster is the new Arm Cortex-X4 – “the fastest CPU that we have ever built,” Arm stated in a blog post – bringing 15 percent more performance compared to the Cortex-X3 while consuming 40 percent less power on the same process. Arm is taping out the Cortex-X4 on the TSMC N3E process.

Cadence will support customers using the new Arm TCS23 through RTL-to-GDS digital flow Rapid Adoption Kits (RAKs) for 3nm and 5nm nodes. The company has fine-tuned its RAKs for Arm Cortex-X4, Cortex-A720 and Cortex-A520 CPUs and Immortalis-G720, Mali-G720 and Mali-G620 GPUs. Cadence verification flow has also been optimized for the newest Arm CPUs and GPUs.

Synopsys, too, has announced support for TCS23 through QuickStart Implementation Kits (QIKs) that are tuned for the latest 5, 4 and 3nm process technologies.

MediaTek and Nvidia to collaborate on automotive solutions

MediaTek and Nvidia will collaborate to deliver a complete range of in-vehicle AI cabin solutions. MediaTek will develop automotive SoCs integrating a new Nvidia GPU chiplet with Nvidia AI and graphics IP. The solutions will run Nvidia Drive OS, Drive IX, CUDA and TensorRT software technologies — delivering a range of AI cabin and cockpit functions. This offer will complement MediaTek’s Dimensity Auto platform, which leverages the company’s expertise in mobile computing, high-speed connectivity, entertainment and Android ecosystem.

Tenstorrent and LG collaborating on AI, video codec

Canada-headquartered Tenstorrent and LG Electronics are collaborating to build a new generation of Risc-V, AI, and video codec chiplets to potentially power LG’s premium TV and automotive products of the future and Tenstorrent’s data center products. Through this collaboration, LG will receive from Tenstorrent artificial intelligence and Risc-V CPU technology to drive AI-enhanced features and high-performance computing in LG’s future premium TVs, automotive chips and other products. Leveraging LG’s expertise in video codec technology, the collaboration at chiplet-level and/or IP-level will also deliver video processing capability into Tenstorrent’s future data center products.

ASICs with embedded MEMS sensors

Nanusens has come up with a solution to build CMOS ASICs with embedded MEMS pressure sensors which can be shrunk to whatever process node as it does not require any analog circuitry. Building on its previous achievement – creating MEMS on regular CMOS fabs – the company has now engineered a fully digital circuit design to measure the capacitance of its nanosensors.

New integration and packaging technologies for LiDARs

French research institute CEA-Leti presented new integration and packaging technologies for next-generation LiDAR devices in autonomous vehicles at the recent Electronic Components and Technology Conference in Orlando, Fla. The paper describes a system where light beam steering was developed using wafer-level silicon technology, including optical phase arrays. The system employs 10µm-diameter mid-process through-silicon vias (TSVs), significantly improving interconnect density by distributing them on the whole backside surface of the devices. Combining 40µm fine-pitch, lead-free solder flip chip on a silicon interposer, this collective integration increases the performance and compactness of LiDARs, while lowering their cost for use in autonomous vehicles. CEA-Leti presented other six papers at ECTC, illustrating its 3D interconnection expertise, primarily on its semiconductor wafer-level platform.

Nickel aluminide, a lower resistivity alternative to copper

Belgian research center imec has provided the first experimental evidence that the resistivity of a 7.7-nanometer film of NiAl (nickel aluminide) is as low as 11.5 µOhm cm, which is 23% lower than copper. Alternatives to copper or ruthenium – the metals which are currently used in interconnect metallization schemes – are needed to keep pace with the continuing device scaling, since the width of the most critical interconnect lines in state-of-the-art logic and memory chips soon approaches 10-nanometer. This result was achieved on a 300-millimeter silicon wafer after depositing a 50-nanometer thick large-grain NiAl film on a germanium epi-layer at back-end-of-line compatible temperatures, followed by thinning experiments.

DVFS-linked security vulnerabilities

A team of security researchers at Georgia Tech, the University of Michigan and Ruhr University Bochum in Germany has investigated form of side-channel attack that takes advantage of Dynamic Voltage and Frequency Scaling (DVFS) techniques used by SoCs. The team demonstrated instruction- and data-dependent behavior of Arm CPUs, iGPUs and dGPUs, leaking information via frequency, power and, temperature; presented browser-based ‘pixel stealing’ and ‘history sniffing’ attacks against recent versions of Chrome and Safari, with all side channel countermeasures enabled; showed ‘website fingerprinting’ attacks on Apple devices using internal power and frequency measurements from unprivileged software.

Acquisitions

Cadence has acquired UK-based Pulsic, a longtime provider of technology for floorplanning, placement, and routing of custom ICs. Pulsic offering includes analog, custom design, and memory layout and routing solutions.

Further reading

A blog post from Zvi Or-Bach, President and CEO of MonolithIC 3D, focuses on the hybrid bonding technologies used by Chinese companies as an alternative to dimensional scaling, achieving significant performance despite export restrictions on EUV lithography.

Atomic Semi and Anysphere are two of the startups on which OpenAI – the company behind ChatGPT – is investing, according to market intelligence firm CB Insights. San Francisco-based Atomic Semi describes itself on LinkedIn saying it’s “building a small, fast semiconductor fab. We’ll build the tools ourselves then quickly push them to more advanced geometries”. Anysphere is building Cursor, a code editor based on GPT-4.

This entry was posted on Friday, June 2nd, 2023 at 7:20 am. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.