Tag: AI processing

  • Asus Launches PE3000N Mini PC with Nvidia Jetson IGX Thor T5000

    Asus Launches PE3000N Mini PC with Nvidia Jetson IGX Thor T5000

    Key Takeaways

    1. Nvidia launched the Jetson IGX Thor platform, with development kits and modules now available for OEM partners to create devices like Asus IoT’s PE3000N mini PC.
    2. The Jetson IGX Thor T5000 module features advanced specs, including a Blackwell GPU, 14-core ARM processor, and 128 GB RAM, offering 2,070 FP4 TFLOPS for AI tasks.
    3. The PE3000N mini PC is designed for high-bandwidth sensor fusion, supporting up to 16 GMSL cameras and various I/O options, including USB-A, USB-C, and HDMI.
    4. The device excels in AI capabilities, capable of running generative AI models and facilitating real-time video analytics and autonomous control.
    5. Pricing for the PE3000N is not yet available, but it is expected to be similar or higher than Nvidia’s development kit, with availability anticipated in Q1 2026.

    Nvidia launched the Jetson IGX Thor platform in August, but its release was slow, similar to the DGX Spark platform. Now that the development kit and the standalone module are finally up for grabs on the official website, Nvidia’s OEM partners are gearing up to roll out their devices that run on the Jetson IGX Thor platform. Asus IoT is among the first to announce their device, the PE3000N mini PC, which is designed to meet the high standards of next-gen robotics and smart automation.

    Performance Specifications

    The Jetson IGX Thor T5000 module features a Blackwell GPU, a 14-core ARM processor, and 128 GB LPDDR5X RAM, offering 2,070 FP4 TFLOPS for AI tasks. The compact design of the PE3000N mini PC allows developers and integrators to reach new heights in autonomy, sensor fusion, and AI-driven control for industrial, commercial, and smart infrastructure applications. Built for industrial environments, this device boasts a rugged chassis that meets MIL-STD-810H standards and can function in temperatures from -20°C to 60°C.

    Configurability and Connectivity

    Asus provides excellent configurability with the PE3000N, allowing it to come with up to 4x 25 GbE NICs and support for 16 GMSL cameras, making it ideal for high-bandwidth sensor fusion and advanced machine vision. On the I/O front, there are standard desktop ports, including 4x USB-A, 2x USB-C, and HDMI video out along with audio jacks. Many industrial applications can also be catered to through an optional second stack, which adds interfaces like PoE, GMSL, CAN, and QSFP28. Other supported standards include PTP/PPS for accurate sensor synchronization, onboard TPM 2.0 for enhanced security, plus LTE/5G/GNSS modules for broader connectivity.

    AI Capabilities

    With the powerful AI-processing abilities of the Jetson IGX Thor T5000 module, the PE3000N is capable of running generative AI models, such as visual language models (VLMs) and large language models (LLMs), as well as facilitating real-time video analytics and autonomous control tasks. Nvidia’s Jetson IGX Thor is also compatible with various physical AI applications, including Nvidia Isaac for robotics, Nvidia Holoscan for real-time sensor processing, and Nvidia Blueprint for AI agents focused on video analytics.

    At this moment, there’s no pricing information available, but considering that Nvidia’s own development kit is priced at around ~3,000 EUR, it’s likely that the price of the PE3000N will be similar or even higher with all the extra I/O options included. The device is expected to be available in Q1 2026.

    Source:
    Link

     

    Tags: , ,
  • ASUS NUC 14 Pro AI: Copilot+, Intel Arc GPU, Wi-Fi 7, Thunderbolt 4

    ASUS NUC 14 Pro AI: Copilot+, Intel Arc GPU, Wi-Fi 7, Thunderbolt 4

    ASUS has revealed the ASUS NUC 14 Pro AI, a compact PC made for activities that need AI processing, gaming, and edge computing. This device runs on Intel Core Ultra processors (Series 2) and comes with Microsoft Copilot+, merging high performance with efficiency in a small form factor.

    Features of the ASUS NUC 14 Pro AI

    The ASUS NUC 14 Pro AI is powered by Intel Core Ultra 9 processors that integrate CPU, GPU, and NPU technologies. This setup can achieve up to 120 platform TOPS (Tera Operations Per Second) and 48 NPU TOPS, which is three times more AI performance than previous versions. Its Intel Arc GPU uses Xe2 Architecture, offering up to 67 TOPS, and includes XeSS and XMX AI engines, which improve gaming experiences and content creation.

    Design and Connectivity

    Built for ongoing usage, the device includes advanced thermal management and security features like Secure Boot, TPM, and fingerprint recognition. With a 0.6-liter size and a height of 34mm, it’s perfect for Internet of Things (IoT) and edge computing applications. The design allows for tool-free access, making upgrades quite simple.

    Connectivity is robust, featuring Wi-Fi 7, Bluetooth 5.4, and dual Thunderbolt 4 ports that ensure fast data transfer and compatibility. Plus, the ASUS Control Center and Windows Autopilot make it easy to deploy and integrate into IT setups.

    Specifications and Storage Options

    The NUC 14 Pro AI offers various configurations with Intel Core Ultra 5multiple I/O ports

    Pricing and Availability

    Details on pricing and availability will be provided at a later date, and customers interested can reach out to ASUS representatives for additional information.

    Source: Link

    Image 1
    Tags: ,
  • MediaTek unveils Dimensity 9300+ supporting Google Gemini Nano, Meta Llama

    MediaTek unveils Dimensity 9300+ supporting Google Gemini Nano, Meta Llama

    MediaTek recently unveiled the Dimensity 9300+, a cutting-edge chipset tailored for top-tier smartphones. This new chip, although slightly enhanced in specifications compared to its predecessor, the Dimensity 9300, primarily emphasizes advancements in on-device AI processing and extended support for large language models (LLMs).

    Enhanced AI Capabilities and LLM Compatibility

    The Dimensity 9300+ is engineered to support various LLMs, such as Google’s Gemini Nano and Meta’s Llama 2 and 3. MediaTek highlights the chip’s broad LLM compatibility and its on-device LoRA Fusion technology, which is expected to cultivate a diverse range of generative AI applications.

    Industry-Leading Performance

    MediaTek asserts that the Dimensity 9300+ is the first in the industry to achieve rapid operation speeds for specific LLMs. Notably, it boasts a speed of 22 tokens per second for Meta’s Llama 2 7B model.

    Power Efficiency and Connectivity Enhancements

    Apart from AI advancements, the Dimensity 9300+ promises notable enhancements in power efficiency, reportedly reducing power consumption by 20% in certain gaming scenarios. The chipset also integrates a “star speed engine” to optimize network connectivity and facilitates dual network concurrency.

    Technical Specifications
    Constructed using a 4nm process, the Dimensity 9300+ houses an octa-core CPU, featuring one high-performance Arm Cortex-X4 core clocked at 3.4 GHz, three additional Cortex-X4 cores clocked at 2.85 GHz, and four Cortex-A720 cores running at 2.0 GHz.

    The chipset incorporates MediaTek’s innovative NeuroPilot technology to elevate AI processing capabilities. The APU 790 AI engine enables the execution of LLMs with diverse parameter sizes, ranging from 1 billion to 33 billion, with double the speed of competing solutions for 7 billion parameter LLMs.

    For gaming enthusiasts, the Dimensity 9300+ integrates a second-generation hardware raytracing engine alongside an Arm Immortalis-G720 GPU. Additionally, it incorporates MediaTek’s Imagiq 990 ISP for enhanced camera functionality and MiraVision 990 for improved visual content processing.

    In terms of connectivity, the Dimensity 9300+ features a 5G R16 modem supporting Sub-6GHz networks and downlink speeds of up to 7Gbps.

    Availability

    The Dimensity 9300+ will make its debut on the Vivo X100S, becoming the first device to integrate this advanced chipset. Furthermore, the iQOO Neo9S Pro, set for release this month, will also be equipped with the Dimensity 9300+.

    MediaTek unveils Dimensity 9300+ supporting Google Gemini Nano, Meta Llama
    Tags: ,
  • Qualcomm Snapdragon X Elite Chip Appears on Geekbench

    Qualcomm Snapdragon X Elite Chip Appears on Geekbench

    Recently, a new Qualcomm chip, identified as model number Qualcomm ZH-WXX and thought to be the Snapdragon X Elite, made an appearance on the Geekbench benchmarking platform. This chip showcased impressive performance metrics with a single-core score of 2574 and a multi-core score of 12562. It was tested on a device operating on Microsoft Windows 11, equipped with a 12-core CPU at 4.01 GHz, and 32GB of RAM.

    In October 2023, Qualcomm introduced the Snapdragon X Elite, which features the custom-designed Qualcomm Oryon CPU. The chip is built on a TSMC’s 4nm process node, featuring 12 cores and 64-bit architecture. It boasts a base clock speed of 3.8GHz, with the capability to reach up to 4.3GHz in dual-core turbo frequency.

    Performance Boost with Snapdragon X Elite

    The company positions this CPU as a trailblazer in mobile computing, claiming it delivers up to twice the CPU performance of its competitors. It achieves this heightened level of performance at peak power while consuming only a third of the energy.

    AI Processing Capabilities of Snapdragon X Elite

    The standout feature of the Snapdragon X Elite is its AI processing capabilities. It is said to be 4.5 times faster in AI processing compared to its rivals, capable of handling complex AI models with over 13 billion parameters directly on the device. This capability enables a smoother user experience, enhancing multitasking and offering intuitive features like AI-powered assistants and advanced video conferencing tools.

    Market Impact and Future Expectations

    The company has made strong claims about the X Elite’s performance, stating it surpasses the Apple M2 Max and Intel Core i7-1355U in single-threaded performance metrics. According to Qualcomm, the X Elite not only outperforms the M2 Max in single-threaded tasks but also achieves peak performance with 30% lower power consumption. In multi-core tasks, it’s said to be 50% faster than the M2.

    Major PC manufacturers such as Dell, Acer, Honor, HP, Lenovo, Samsung, and ASUS are anticipated to release laptops featuring the Snapdragon X Elite starting in mid-2024. Even Xiaomi is reportedly preparing to unveil a new slim laptop, likely from the Redmi G Pro 2024 series, featuring this chip.

    Qualcomm Snapdragon X Elite Chip Appears on Geekbench
    Tags: , ,