Nvidia has announced the Jetson Orin module, the perfect solution for next generation AI and robotics applications. The module features an Nvidia Ampere architecture GPU together with deep learning and vision accelerators. Its fast memory bandwidth and high-speed IO can feed multiple concurrent AI application pipelines. This translates into being able to develop solutions using complex AI models to solve problems such as natural language understanding, 3D perception, and multi-sensor fusion.
The Jetson Orin series of system-on-modules (SOMs) deliver up to 80x the AI performance of Nvidia’s Jetson Nano. The Jetson family now boasts the Orin-based modules that span from entry-level Orin Nano to the highest performance Jetson AGX Orin. This allows customers the flexibility to easily scale applications.
Figure 1. Block diagram of Jetson Orin Nano.
The Orin Nano is the smallest Jetson form-factor, with power options from 5 to 15 W. Two versions of the Orin Nano are available, one with 4 GB of RAM and the other doubles this to have 8 GB integrated.
The Ampere Architecture GPU has up to eight streaming processors (SMs) composed of 1024 CUDA cores and up to 32 Tensor Cores used for AI processing. This delivers better performance per watt than the previous generation and allows applications to take advantage of the fine-grained structured sparsity in deep learning networks to double the throughput for Tensor Core operations.
The Jetson Orin Nano includes a six-core Arm Cortex-A78AE CPU, video decode engine, ISP, audio processing engine, and video input block. Despite its small footprint of 70 x 45 mm, the module includes various high-speed interfaces including PCIe Gen3, three 10 Gbps USB 3.2 ports, eight lanes of MIPI CSI-2 camera ports, and various sensor connections.
Altron celebrates 60th birthday with a call to rebuild Johannesburg Altron Arrow
News
Altron is celebrating its 60th birthday by honouring Johannesburg’s heritage and encouraging business, government and civil society to come together and respond to our President’s call to rebuild Johannesburg.
Read more...KIOXIA pioneer new 3D Flash technology EBV Electrolink
DSP, Micros & Memory
KIOXIA Corporation and Sandisk Corporation have pioneered a state-of-the-art 3D flash memory technology, setting the industry benchmark with a 4,8 Gb/s NAND interface speed, superior power efficiency, and heightened density.
Read more...Artificial intelligence meets embedded development Avnet Silica
DSP, Micros & Memory
Microchip Technology is leveraging the power of artificial intelligence (AI) to assist software developers and embedded engineers in writing and debugging code with the launch of its MPLAB AI Coding Assistant.
Read more...Embedded module for AI vision applications Rugged Interconnect Technologies
DSP, Micros & Memory
The TQMa95xxSA supports AI/ML with vision applications through an integrated NPU and an image processor unit.
Read more...High-speed Flash for SoC applications NuVision Electronics
DSP, Micros & Memory
GigaDevice has unveiled the GD25NE series of dual-power supply SPI NOR Flash chips, designed specifically for 1,2 V system-on-chip (SoC) applications.
Read more...Super-fast H.264 encoder FPGA core EBV Electrolink
DSP, Micros & Memory
An ITAR-compliant H.264 core designed for AMD FPGAs provides baseline H.264 support and is currently the smallest and fastest FPGA core in the industry.
Read more...Reference board for cardio monitoring Altron Arrow
Telecoms, Datacoms, Wireless, IoT
The STDES-ESP01 reference board from STMicroelectronics demonstrates the capability of the ST1VAFE6AX and ST1VAFE3BX biosensors to detect ECG and SCG signals.
Read more...ST MCUs extend ultra-low power innovation Altron Arrow
DSP, Micros & Memory
STMicroelectronics has introduced new STM32U3 microcontrollers with cutting-edge power-saving innovations that ease deployment of smart connected tech, especially in remote locations.
Read more...Multicell battery monitoring Altron Arrow
Power Electronics / Power Management
The LTC6811 from Analog Devices is a multicell battery stack monitor that measures up to 12 series connected battery cells with a total measurement error of less than 1,2 mV.
Read more...Chipset enables ultra-wide signal capture RFiber Solutions
DSP, Micros & Memory
Jariet Technologies has developed Electra, a chipset that enables ultra-wide, multi-function and multi-band signal capture and generation from a single component.
printer friendly version