Glenfly Tech Co., Ltd. ("Glenfly") officially launched GF800, its first high-performance ARM-based SoC.

 

As the pivotal product marking Glenfly's evolution from a GPU chip and graphics card supplier into a full-stack SoC platform provider, GF800 translates the company's long-accumulated expertise in graphics rendering, parallel computing and AI acceleration into platform-level, productized capabilities — delivering a highly integrated, highly scalable and independently controllable system-level computing platform for edge-side products, edge computing and industry intelligence applications.

 

Today, GPU solutions in the ARM SoC market mainly include ARM Mali/Immortalis, Qualcomm Adreno, Samsung Xclipse and Imagination PowerVR. GF800 integrates Glenfly's continuously self-developed GPU IP into an SoC product for the first time, extending its proprietary GPU capabilities from standalone chips to system-level platforms and offering customers a new, differentiated and independently controllable choice.

 

Backed by a highly integrated SoC design, multi-OS support and a unified software platform, GF800 addresses a broad range of computing terminals — including but not limited to AI terminals, edge servers, industrial computing equipment, robotics and embodied intelligence — further expanding the platform footprint of domestic SoCs across edge-side products, edge computing and industry intelligence scenarios.

 

Eight Core Highlights, Unlocking Full Platform Value

  • Heterogeneous Integration for Efficient Collaboration | GF800 adopts a heterogeneous architecture combining an ARM CPU, an in-house GPU and a dedicated NPU with a unified memory architecture, consolidating general-purpose computing, graphics rendering and AI acceleration on a single platform to provide downstream terminal manufacturers and industry partners with a complete and flexible one-stop underlying platform. Across different workload types, GF800 fully leverages the respective strengths of the CPU, GPU and NPU, providing a system-level computing foundation for multitasking, graphics applications, AI inference and multimedia processing.
  • In-House GPU with Powerful Graphics Processing | GF800 integrates Glenfly's proprietary GPU, supporting DirectX 12, OpenGL 4.6, OpenGL ES 3.2, OpenCL 3.0 and Vulkan 1.3, with the corresponding CTS tests passed, and features a dedicated built-in 2D hardware engine. Balancing display performance, full graphics rendering and general-purpose compute acceleration, GF800 delivers a stable, efficient and consistent experience for high-quality graphics processing.
  • AI Computing with Efficient Deployment | GF800 is built on Glenfly's proprietary Glenfly HMUA hardware architecture and Glenfly HUCA software platform, driving in-depth optimization and sustainable evolution around graphics capabilities, the AI ecosystem and general-purpose computing at the architectural level. The full-stack hardware-software synergy between Glenfly HUCA and Glenfly HMUA meets requirements such as rapid AI model migration, mainstream deep learning framework adaptation, large model inference acceleration and multimodal model deployment. Following an all-big-core design approach, the chip is equipped with eight Arm Cortex-A78 high-performance general-purpose cores, providing stable and efficient general computing power for a wide range of terminal applications and comfortably handling multitasking and high-load computing. GF800 not only provides a high-performance, hardware-software co-designed computing platform for edge-side AI applications, but also retains the inherent power-efficiency advantages of the ARM architecture.
  • Broad Ecosystem, Lower Migration Costs | GF800 is deeply optimized and adapted for domestic operating systems such as UOS and Kylin, supports mainstream operating systems including Linux and Android, and comes with a comprehensive driver and software stack — supporting the DRI architecture, following the DRM/KMS framework and remaining compatible with Windows. Targeting edge-side products, mobile intelligent terminals and customized dedicated systems, GF800 helps partners improve efficiency and reduce secondary development costs across multi-terminal adaptation, cross-platform migration and rapid deployment.
  • High-Specification Display with Multi-Channel Independent Output | GF800 supports mainstream display interfaces including HDMI 2.1, DP 1.4/eDP 1.4b and DSI, enabling up to four independent display outputs via HDMI, DP/eDP and DSI, with support for up to 8K display. The chip integrates HDCP content protection and has passed conformance tests including HDMI CTS, DP CTS and HDCP CTS, safeguarding the reliability, compatibility and content security of display output.
  • Full-Stack Multimedia for Enhanced Perception and Interaction | GF800 features a next-generation ISP supporting image input of up to 48 MP, with integrated hardware-accelerated algorithms including HDR, 3A, LSC, 3DNR, 2DNR, sharpening, dehazing, fisheye correction and gamma correction. Combining video codec and ISP imaging capabilities, GF800 delivers image perception, video encoding/decoding and audio processing in one, providing complete multimedia support for machine vision, video processing, voice interaction and multimodal intelligent applications.
  • Rich Interfaces, Flexible Expansion | GF800 offers an extensive set of I/O interfaces including PCIe 4.0, USB 3.2, SATA, eMMC, GMAC, CSI, SDIO, I²S, PDM, UART, SPI and I²C, meeting the diverse needs of industry partners in terminal form factors, functional expansion and product customization.
  • High Integration for Cost Reduction and Efficiency Gains | Leveraging a highly integrated SoC design, hardware-software co-optimization, rich I/O expansion capabilities and multimedia processing performance geared to 8K scenarios, GF800 reduces system complexity and peripheral component costs, shortens product development cycles, and provides partners with a flexible, reliable and continuously evolving platform solution.

Strengthening Full-Stack Capabilities to Empower Domestic Intelligent Computing

 

The launch of GF800 marks the formal realization of Glenfly's full-stack SoC platform capabilities. Building on its proprietary core GPU technology, the company has further established system-level product capabilities spanning CPU, GPU, NPU, multimedia, display, I/O, drivers and the software ecosystem. Looking ahead, Glenfly will continue to strengthen core technology R&D and product iteration, refine its product portfolio for edge-side products, edge computing and industry intelligence, and work with industry partners to build an open, thriving and sustainable domestic intelligent computing ecosystem.

 

Wenjie Wu, Deputy General Manager of Glenfly: "GF800 is a high-performance, high-efficiency platform built for next-generation intelligent terminals and edge computing applications. Powered by high-performance Arm processors, and underpinned by our core team's 20-plus years of GPU R&D experience together with our continuously evolving Glenfly HMUA graphics and computing architecture and Glenfly HUCA computing ecosystem, it will provide partners with a complete, open and highly competitive intelligent computing platform."

Yuanfeng Wang, Chief Technology Strategy Officer of Glenfly: "The significance of GF800 lies not merely in launching a brand-new SoC, but in deeply integrating more than two decades of Glenfly's GPU architecture innovation, driver software development and system engineering practice into a scalable, sustainably evolving platform system. Our proprietary GPU is the technological cornerstone of GF800, while full-stack synergy is its core competitive advantage. Building on this platform, we will continue to deliver stable, efficient and ever-evolving product value for edge computing and industry intelligence applications."