Microcontroller Tips

Synopsys has introduced its ESUN IP solution, a Layer 1 and Layer 2 platform for Ethernet for Scale-up Networking aimed at AI accelerators, switches and HPC systems. The solution includes a 1.6T multi-rate Ethernet MAC with configurations up to 4 x 400G, a 1.6T PCS with RS-FEC, a UE/ESUN Link Layer Controller with link-level retry and credit-based flow control, a silicon-proven 224G PHY and verification IP to support integration and validation.

Infineon Technologies AG has introduced the AIROC™ UWB TSL100, a first-generation ultra-wideband product family for precise positioning and smart presence detection in automotive, consumer and industrial applications.

The GD24CL series from GigaDevice is an I²C EEPROM family for long-term storage of configuration and other critical system data in industrial, energy, IoT, data center and networking applications. The devices support 100 kHz, 400 kHz and 1 MHz I²C operation, offer endurance up to 4 million program/erase cycles, data retention up to 100 years, integrated ECC and operation from -40°C to 125°C.

The NSC128L42 from Nuvoton Technology Corporation is a 32-bit Arm® Cortex®-M23 microcontroller for high-precision measurement that integrates a 24-bit Sigma-Delta ADC, 12-bit SAR ADC, LCD driver and voice support in a single package. The device runs at up to 49.152 MHz, supports 2.0 V to 5.5 V operation, includes 320 KB Flash and 18 KB SRAM, and targets portable and industrial designs such as blood pressure monitors, precision scales and other smart measurement systems.

Like many aspects of artificial intelligence (AI), there’s not a single approach when calculating efficiency across the AI power chain. It depends on where you look. Are the calculations related to the facility layer, the compute and hardware layer, or the workload and algorithm layer? At the facility level, efficiency is discussed in terms of power usage effectiveness (PUE) and its inverse, datacenter infrastructure efficiency (DCiE).

Trans inductor voltage regulators, occasionally referred to as transient load voltage regulators, (TLVRs) use coupled inductors with an additional secondary winding to create a series-connected secondary loop, allowing all phases to react simultaneously to load changes. That enables TLVRs to support the extreme power demands of GPUs in AI datacenters, and as detailed below, they cost less than conventional multi-phase designs.

By Pietro Antonio Ciclese, Senior Technical Marketing Engineer, Ambarella The workloads that generate the most commercial value in edge AI scenarios run under power and thermal constraints that are, to put it mildly, very far from those found in data centers. Take, for example, a vision camera used for surveillance or inspection. These often draw power over Ethernet, and are therefore capable only of delivering 15 to 30 watts in total.

by Sean Toohey Edge AI has transformed how we use and manage embedded systems by shifting data processing from the cloud to the device. With less reliance on remote servers for processing insights and handling requests, operators benefit from faster responses and swifter workflows. However, as data processing priorities shift from cloud to edge-based, the way we approach embedded system security must also evolve.

A new family of FPGAs from Efinix targets constrained edge AI deployments with a combination of power reduction, integrated packaging, high-speed I/O, reliability features, and built-in cryptographic security. The Titanium Edge devices reduce static power consumption by 50% compared to the previous Titanium generation, which is designed to support always-on inference in thermally limited systems.

The EPIC-PTH9 from AAEON is a 4-inch EPIC single-board computer built on Intel Core Ultra Series 3 processors, offering up to 16 cores and 16 threads, configurable TDP from 25W to 65W and up to 180 TOPS from the CPU, Intel Arc B390 GPU and NPU for AI inference workloads.