Next Nvidia automotive AI chip; Xilinx, Mavenir mMIMO; Dell spins off VMWare.
Pervasive computing — IoT, edge, cloud, data center, and back
PC maker Dell Technologies is spinning off 81% equity ownership of VMWare to two standalone public companies. VMWare, founded in 1998, provides software to manage networking, apps, and cloud/data center. VMWare will pay a special dividend to its investing, which will generate $9.3 – $9.7 billion for Dell to use toward cutting its debt load. Both companies plan to continue their commercial agreement to work together. The transaction will close in 4th quarter of 2021.
The U.S. telecom company Verizon is launching a business-focused 5G internet service in 21 cities in the United States. Called 5G Business Internet, the service is fixed wireless, offers 100, 200, 400 Mbps plans with no data limits, using Verizon’s 5G Ultra Wideband technology.
Mavenir and Xilinx collaborated on an Open RAN massive MIMO (mMIMO) product. Massive multiple-input multiple-output (mMIMO) is wireless base station with an antenna array that can efficiently serve multiple mobile users and uses 5G bands, such as sub 6 GHz and millimeter wave. The Mavenir/Xilinx mMIMO combines Xilinx’s Category B O-RAN Radio Unit with Mavenir’s software — a virtualized radio access network (vRAN) that supports mMIMO. The companies expect the first Massive MIMO 64TRX product in late 2021.
Codasip has upgraded its Studio IC design and CodeSpace software tools with a RISC-V tutorial, an improved SDK, and linker support for developing linker scripts that represent a system memory map.
Nvidia announced its new data processing unit (DPU), the BlueField-3 DPU for data centers. The DPU adds AI and accelerated computing to speed up software-defined networking, storage, and cybersecurity at data centers. The security feature isolates and authenticates users to free up resources. “A new type of processor, designed to process data center infrastructure software, is needed to offload and accelerate the tremendous compute load of virtualization, networking, storage, security and other cloud-native AI services,” said Jensen Huang, founder and CEO of NVIDIA in a press release.
Arm reports on its collaborations with Nvidia here, including accelerating the edge DPU by adding GPUs to Marvell’s Arm-based Octeon family of DPUs. The DPUs are used for enterprise computing, 5G infrastructure, and edge applications. The NVIDIA GPU will offload workloads such as image recognition and security in constrained edge environments.
Automotive
Nvidia unveiled its Nvidia Drive Atlan SoC for automotive AI. The SoC has Nvidia GPUs, Arm CPU cores, and Nvidia’s BlueField data processing unit (DPU). Nvidia says it runs 1,000 trillion operations per second (TOPS). The chip is newest of Nvidia’s three automotive SoCs (Nvidia Drive series), which started with Xavier (30 TOPS) now in production vehicles today, Orin (254 TOPS) due out in 2022 vehicles, and the new Atlan to appear in 2025 vehicles. Separately announced, Volvo Cars, Zoox, and SAIC will now be using Nvidia Drive, according to a press release.
Velodyne and Ansys are working on physics-based models of Velodyne’s lidar sensor to add to Ansys’ VRXPERIENCE drive simulator. Their goal is to improve hazard identification in automated vehicles.
Verizon announced that it will offer a hyper precise location SaaS (software as a service) available in 100 markets for advanced IoT and smart driving. The Hyper Precise Location (HPL) service corrects global navigation satellite systems (GNSS) data in real time, for possible use on 5G edge systems such as C-V2X (cellular vehicle-to-everything) for automotive, robotics, construction, agtech, and HD mapping. Accuracy is claimed to be 1 – 2 centimeters (0.78 of an inch) on a 5G or 4G network. This follows an announcement last week that Verizon and automotive manufacturer Honda are working together on testing 5G and mobile edge computing (MEC) networks to improve automotive safety. Honda SAFE SWARM includes Verizon’s 5G Ultra Wideband and 5G Mobile Edge Compute. The two companies are hope the tests prove that AI might not be needed as heavily on self-driving cars.
Cadence is acquiring Pointwise, known for its mesh generation use in computational fluid dynamics (CFD). The acquisition follows the recent announcement that Cadence is acquiring another CFD company called NUMECA. The Pointwise acquisition target aerospace, where CFD can simulate air flow over a wing. The terms of the Pointwise deal were not disclosed.
Renesas used Synopsys’ DSO.ai (Design Space Optimization AI) to design its automotive chips. The AI system ingests a large amount of data from chip design tools and finds PPA optimizations and workflow improvements as the design evolves. “In only its first year in the market, DSO.ai has already helped many customers achieve better PPA solutions in dozens of design projects,” said Stelios Diamantidis, senior director of Artificial Intelligence solutions at Synopsys in a statement. “AI is giving EDA a new dimension for addressing the increased complexity of silicon technologies, accelerating product timelines, and enabling engineering teams to scale.”
Security
H3C Semiconductor designers used Ansys’ multiphysics platform to create the ENGIANT 660 and pass stringent testing requirements, according to a press release. a network processor chip that enables routing, AI, 5G backhaul and cybersecurity applications.
Siemens Digital Industries Software has signed an agreement to acquire OneSpin Solutions, a formal verification software for ICs. Siemens will add OneSpin’s technology to its Xcelerator portfolio of verification products.
Milestones
The Accellera board approved the Portable Test and Stimulus Standard 2.0.
Job, Event and Webinar Boards: Find industry jobs and upcoming conferences and webinars all in one place on Semiconductor Engineering. Knowledge Center: Boost your semiconductor industry knowledge. Videos: See the latest Semiconductor Engineering videos.
Leave a Reply