Computing Technology
Search documents
寅谱计算发布KLEENE主板智控技术
Zheng Quan Ri Bao Wang· 2025-12-23 08:37
Core Insights - Infplane Computing held a technology and product launch event on December 19, introducing its foundational technology products, including the KLEENE motherboard control technology and AI inference acceleration technology combined with SSDs [1][2]. Group 1: KLEENE Technology - The KLEENE technology focuses on the underlying control layer of computers, optimizing power consumption, performance, and workload distribution without altering hardware design or software compatibility [1]. - The CTO emphasized that as AI moves into inference and engineering phases, computational power is becoming a new foundational resource, with system-level scheduling and energy efficiency management being critical for user experience [1]. Group 2: Hilbert Product Launch - Infplane Computing launched its first product in the "desktop supercomputer" series, the Infplane Personal Workstation: Hilbert, developed in collaboration with AMD and other partners [2]. - Hilbert aims to bring server-level computing capabilities to personal desktop environments, supporting local large model inference, scientific computing, and generative AI workflows, while also functioning as a high-performance gaming machine [2]. - The product is set to launch in North America in 2026, with a second-generation product expected to integrate workstation, NAS, gateway, and gaming capabilities [2]. Group 3: AI PC Collaboration - The CEO discussed advancements in AI PC collaboration, highlighting the need for system-level hardware and software synergy to redefine personal computer capabilities as large models transition from cloud to local environments [3]. - Infplane Computing is collaborating with Mechanical Revolution to explore new AI PC product forms based on KLEENE technology, aiming for stable and efficient operation of multimodal applications on personal devices [3]. - The company is also working on integrated terminal forms that enhance computing and storage collaboration, addressing the challenges posed by growing model sizes and rising memory costs [3].
寅谱计算举办技术与产品发布会
Zheng Quan Ri Bao Wang· 2025-12-22 07:15
Core Insights - Infplane Computing has launched a series of foundational technology products, including the KLEENE motherboard control technology and AI inference acceleration technology combined with SSDs, aimed at enhancing computing capabilities in personal desktop environments [1][2] Group 1: KLEENE Technology - The KLEENE technology focuses on the underlying control layer of computers, optimizing power consumption, performance, and workload distribution without altering hardware design or software compatibility [1] - The CTO emphasized that as AI moves into inference and engineering phases, computational power is becoming a new foundational resource, with system-level scheduling and energy efficiency management being critical for user experience [1] Group 2: Hilbert Workstation - The first product from the "desktop supercomputing" series, named Hilbert, is designed to bring server-level computing capabilities to personal desktops, supporting local large model inference, scientific computing, and generative AI workflows [2] - Hilbert will launch in North America in 2026, serving as a precursor to a second-generation product based on the AMD platform, which will integrate workstation, NAS, gateway, and gaming console functionalities [2] - The product emphasizes stability, reproducibility, and long-term user experience over single performance metrics [2] Group 3: AI PC Collaboration - The CEO highlighted the collaboration with Mechanical Revolution to explore AI PC capabilities, focusing on system-level hardware and software synergy for stable and efficient operation of multimodal applications on personal devices [3] - The company is advancing integrated terminal development that combines computing and storage, addressing the challenges posed by growing model sizes and rising memory costs [3] - Ongoing technical reserves are being developed in areas such as efficient offloading for large model inference and SSD control protocols, with plans to disclose details of the next-generation products in the future [3]