tensorflow

Posted on by Jean-Luc Aufranc (CNXSoft) - No Comments on Cadence Tensilica HiFi iQ DSP to power next-generation Voice AI and Audio applications

Cadence Tensilica HiFi iQ DSP to power next-generation Voice AI and Audio applications

Tensilica HiFi iQ DSP block diagram

Cadence Tensilica HiFi iQ DSP IP is the sixth generation of the company’s HiFi DSP family found in countless SoCs with audio features, and is based on a new architecture designed for next-generation voice AI and immersive audio applications. Compared to the Tensilica HiFi 5s DSP, the HiFi iQ DSP offers twice the compute performance, 8x higher AI performance, and more than 25% energy savings for most workloads, while delivering over 40% performance uplift on several audio codecs. The company claims it can act as an all-in-one AI processor for voice AI applications running popular SLMs and LLMs on the DSP itself. Tensilica HiFi iQ DSP key specifications: Load Units – 2 VLIW Slots – 5 Vector operations (SIMD) – 256-bit Accumulator Width – 80-bit Fixed Point MACs MACs for AI – Up to 128 16×8, up to 256 16×4 Optional instructions for AI (improved from HiFi 5s) Optional Vector […]

Posted on by Debashis Das - 2 Comments on Firefly CAM-3576 series – Tiny Rockchip RK3576 SBCs for commercial, industrial, and automotive applications

Firefly CAM-3576 series – Tiny Rockchip RK3576 SBCs for commercial, industrial, and automotive applications

CAM 3576Q38 Mini AI SBC

Firefly Technology has introduced the CAM-3576 series of tiny (38 × 38 mm) SBCs based on the Rockchip RK3576 processor with a 6 TOPS NPU for AIoT, edge AI, smart vision, industrial, and automotive applications. It comes in three variants, which include the CAM-3576Q38 (commercial), the CAM-3576JQ38 (industrial), and the CAM-3576MQ38 (automotive) modules designed for smart cameras, intelligent security systems, dash cams, and private on-device AI model deployment. The CAM-3576 series supports up to 16GB of LPDDR5 RAM, up to 256GB eMMC flash, and also includes a microSD card for expansion. Additionally, the boards feature a MIPI CSI input for up to 16MP camera sensors with HDR support, Fast Ethernet, Wi-Fi 6, USB 2.0, USB-C (device), RS-485, UART, I²C, ADC, GPIOs, audio input/output, and RTC support. Firefly CAM-3576Q38 specifications: SoM – ICORE-3576Q38 SoC – Rockchip RK3576 (Q38 – Commercial) or Rockchip RK3576J (JQ38 – Industrial) or Rockchip RK3576M (MQ38 – Automotive) […]

Posted on by Jean-Luc Aufranc (CNXSoft) - 1 Comment on Toradex Luna SL1680 – A “Pro Consumer” Raspberry Pi lookalike based on Synaptics SL1680 Edge AI SoC

Toradex Luna SL1680 – A “Pro Consumer” Raspberry Pi lookalike based on Synaptics SL1680 Edge AI SoC

Toradex Luna SL1680 SBC

Toradex Luna SL1680 is a credit card-sized single board computer heavily inspired by the Raspberry Pi 5 design, but powered by a Synaptics SL1680 quad-core Cortex-A73 SoC suitable for Edge AI applications thanks to a built-in 7.9 TOPS NPU. Designed for pro consumer and light industrial Edge AI applications, the Luna SL1680 ships with up to 4GB RAM and 256GB eMMC flash, and offers many of the same ports as the Pi 5, including a Gigabit Ethernet RJ45 port with optional PoE support, four USB 3.0 ports, two micro HDMI ports (but one Tx, one Rx), MIPI DSI and CSI connectors, and a 40-pin GPIO header. The 16-pin PCIe FFC connector is gone, making way for an M.2 Key-E socket, mostly useful for adding wireless connectivity. Toradex Luna SL1680 specifications: SoC – Synaptics SL1680 CPU – Quad-core Arm Cortex-A73 processor @ 2.1 GHz GPU – Imagination PowerVR Series9XE GE9920 GPU […]

Posted on by Jean-Luc Aufranc (CNXSoft) - 2 Comments on Quarky Intellio – A LEGO-compatible AI, Augmented Reality, and IoT learning platform (Crowdfunding)

Quarky Intellio – A LEGO-compatible AI, Augmented Reality, and IoT learning platform (Crowdfunding)

Quarky Intellio rover car robot

Quarky Intellio is an ESP32-S3-based development kit designed as an educational platform for learning AI, Augmented Reality (AR), and IoT concepts. It is compatible with LEGO bricks and targets users aged 10 and older. The core AI-AR module features an SPI TFT display interface, a 5MP camera, a speaker and microphone, a microSD card slot for data storage, a USB-C port for programming and charging, servo and GPIO expansion ports, and a 1,000 mAh battery. The company offers a discovery kit with the core module only as well as a rover car kit, and since it’s compatible with LEGO, users can easily create their own robot. Quarky Intellio specifications: Core module – Espressif Systems ESP32-S3-WROOM-1-N16R8 SoC – ESP32-S3 CPU – Dual-core LX7 processor with up to 240MHz Memory – 512KB SRAM, 8MB PSRAM (SiP) Storage – 384KB ROM Wireless – WiFi 4 802.11b/g/n and Bluetooth 5.0 LE Storage – 16MB flash […]

Posted on by Jean-Luc Aufranc (CNXSoft) - No Comments on Brainchip AKD1500 PCIe/SPI Edge AI co-processor to power battery-operated wearables and IoT devices

Brainchip AKD1500 PCIe/SPI Edge AI co-processor to power battery-operated wearables and IoT devices

BrainChip AKD1500 Edge AI co-processor with PCIe and SPI host interfaces

BrainChip AKD1500 is an ultra-efficient Edge AI co-processor delivering up to 800 GOPS while consuming just 300 mW of power, making it suitable for battery-powered wearables and IoT sensors. It’s based on the same event-based neuromorphic technology found on the earlier AKD1000, relying on spiking neural networks (SNN) to deliver real-time inference in a way that is much more efficient than traditional AI chips. What’s new here is that thanks to PCIe and SPI interfaces, the new ADK1500 co-processor can be paired to a wide range of hosts, ranging from Linux-capable application processors to resource-constrained microcontrollers with x86, Arm, or RISC-V architectures. BrainChip AKD1500 key features and specifications: Akida Neuron Fabric Clocked at 5 to 400 MHz Delivering up to 800 effective GOPS at <1mW/GOPS On-device learning capabilities to enable “secure application personalization, without the need for a Cloud connection or retraining” On-Chip Conversion Complex Memory/Storage 1MB on-chip local memory […]

Posted on by Debashis Das - No Comments on Firefly EC-AGXOrin – Jetson AGX Orin 64GB AI inference system supports up to 8 GMSL2 cameras

Firefly EC-AGXOrin – Jetson AGX Orin 64GB AI inference system supports up to 8 GMSL2 cameras

EC AGXOrin NVIDIA Jetson 275 TOPS Edge AI Computer

Firefly EC-AGXOrin is an NVIDIA Jetson AGX Orin 64GB-powered AI inference system, similar to the AAEON BOXER-8645AI and Vecow RAC-1000 rugged Edge AI systems, and designed for edge AI applications such as in-vehicle computing, robotic control, machine vision, intelligent video analytics, and mobile robots. The device features eight GMSL2 connectors (input via two 4-pin Mini FAKRA interfaces) and supports up to 22-channel 1080p or 8K, 30fps H.265 video decoding. It delivers 275 TOPS of AI performance through the NVIDIA module, and integrates 64GB LPDDR5 RAM, 64GB eMMC storage, and offers M.2 NVMe and MicroSD card storage options. Other ports include a 10GbE RJ45 jack, five GbE jacks, USB 3.0, HDMI 2.0, RS232, RS485, CAN, and the system also supports WiFi 6, Bluetooth 5.2, 4G/5G cellular connectivity, and GPS/GNSS. Firefly EC-AGXOrin specifications: SoM – Jetson AGX Orin module with CPU – 12-core Arm Cortex-A78AE v8.2 64-bit processor with 3MB L2 + […]

Posted on by Jean-Luc Aufranc (CNXSoft) - 3 Comments on Google’s open-source, RISC-V-based Coral NPU is integrated into Synaptics SL2610 Edge AI SoCs

Google’s open-source, RISC-V-based Coral NPU is integrated into Synaptics SL2610 Edge AI SoCs

Google Coral NPU

Google has very recently introduced Coral NPU full-stack, open-source RISC-V-based platform for always-on AI on low-power edge devices and wearables. The first chip to integrate the Coral NPU is the upcoming Synaptics Astra SL2610 family. Google Coral NPU The Coral NPU aims to address the software fragmentation on entry-level AI accelerators that makes them difficult to program. By releasing an open-source NPU and associated source code, Google hopes its design will be adopted by silicon vendors, reduce software fragmentation over time, and help machine learning (ML) developers bring products to market faster. Building on the works on the Coral platform, the new, open-source Coral NPU is comprised of three main components: A scalar core – A lightweight, C-programmable RISC-V core that manages data flow to the back-end cores. It uses a simple “run-to-completion” model for ultra-low power consumption and traditional CPU functions. A vector execution unit – A single instruction […]

Posted on by Debashis Das - 7 Comments on Raspberry Pi-like Allwinner A527/T527 industrial SBC features dual camera support and AI acceleration

Raspberry Pi-like Allwinner A527/T527 industrial SBC features dual camera support and AI acceleration

CDEBYTE ECB32 Series Multi core Heterogeneous Allwinner T527 SBC

EBYTE has recently released an Allwinner A527/T527-based Raspberry Pi-like industrial SBC with dual camera and AI features, with a design very similar to the Walnut Pi 2B, and to a lesser extent, the Orange Pi 4A. It is designed for embedded, IoT, and smart commercial applications. The board supports up to 4GB LPDDR4 RAM, 32GB eMMC storage, and MicroSD card expansion. Display options include HDMI 2.0 and a 4-lane MIPI DSI interface enabling dual 4K output, while two 4-lane MIPI CSI connectors allow simultaneous camera inputs. Connectivity options include Gigabit Ethernet, Wi-Fi 4, Bluetooth 4.2, and a PCIe 2.1 x1 slot for high-speed expansion. It also features four USB 2.0 ports, a Type-C power interface, and a 40-pin Raspberry Pi-compatible GPIO header. EBYTE Allwinner A527/T527 SBC specifications: SoC – Allwinner T527 (industrial) / Allwinner A527 (commercial) CPU Octa-core Arm Cortex-A55 processor with four cores @ 1.80 GHz (T527) or 2.0 […]

Exit mobile version