长沙晚报全媒体记者 刘攀 2025世界计算大会期间，华为聚焦算力基础设施创新突破进行了两场（基石篇与行业篇）专题活动，与 会嘉宾围绕算力前沿技术、创新突破与生态共建等展开深入交流，共同探讨如何以领先的AI算力基础 设施，为各行各业的智能化升级注入强劲动力。 连日来，记者跟随这场盛会，走进各大典型应用现场及算力创新成果展区，在这场算力基础设施迭代、 产业升级、人才培养齐头并进的鲜活实践中，探访人工智能时代的算力"脉动"。 携手湖南共推算力基础设施迭代 近年来，湖南牢牢抓住企业、产业、产业链、产业生态四个着力点，持续强化"三力支撑"，一体推 进"四算布局"，打造了一个国家级先进制造业集群、形成了一条具有影响力的技术路线，推出了一批国 产市场占有率高的优势产品。 "我们始终聚焦计算创新，计算创新是算力产业发展的核心引擎；生态开源开放，是产业共荣共生的必 由之路；安全可靠的算力基础，则是实现数字科技自立自强、产业持续高质量发展的根本前提。"2025 世界计算大会期间，华为湖南总经理詹德生表示，下一步，华为将依托湖南在先进计算、音视频、人工 智能等产业的基础和优势，与湖南同心同力，推动算力基础设施持续创新迭代，让算力之 ...

长沙晚报掌上长沙11月20讯（全媒体记者 邹麟）11月20日，以"计算万物 湘约未来——智算驱动新质生 产力"为主题的2025世界计算大会在湖南长沙北辰国际会议中心隆重开幕。 大会同期举办的2025世界计算大会创新成果展示，特设湖南省"人工智能+"展台，集中推出湖南在实 施"人工智能+"行动中的十大重点企业、十大示范场景、十大未来场景、十大重点科技创新平台（首 批）和十大支持措施等"五个十大"清单。宇树科技、强脑科技、华为、中兴通讯、中国电子、曙光信 息、六岳微电子等头部企业，还带来具身智能人形机器人及四足机器人、BrainCo 智能仿生手、昇腾超 节点、折叠屏笔记本电脑、全球首个单机柜级640卡超节点等重磅展品，展现中国计算产业的前沿实力 与创新活力，全面呈现从芯片、硬件到软件、应用的全栈技术突破。 1 1.74 Com 2 44 9 > 超大带宽 te 超低时追 ########## 1 - 4 - 4 - 4 0 THE WINING AND ANY et. T comp Re 国家 级计 1-13 bot 題 ittt MILL 1 2 r 900 ATTACT POOK 12 P 1111 m 19 ...

Group 1 - The article discusses the importance of building a modern infrastructure system, emphasizing the need for integrated planning and sustainable operations [1] - It highlights the necessity of advancing new infrastructure, particularly in information communication networks and major technological facilities [1] - The article also mentions the enhancement of transportation systems and energy infrastructure to ensure resilience and coverage in weak areas [1] Group 2 - The concept of "Supernode" is introduced as a new generation of computing infrastructure, with leading companies like NVIDIA and Huawei launching their own Supernode products [2] - Supernodes are designed to support parallel computing tasks and accelerate data exchange between GPUs, significantly reducing training cycles for large models [4] - The article notes that the shift to Supernodes increases power consumption and communication complexity, leading to investment opportunities in related sectors such as AI chip manufacturers and server producers [4] Group 3 - The article identifies key companies involved in the Supernode ecosystem, including domestic AI chip leaders and high-end CPU manufacturers [4][5] - It emphasizes the potential for growth in companies producing Supernode servers and those providing supporting technologies like PCB and liquid cooling solutions [4]

Core Insights - The conference highlighted advancements in AI capabilities, particularly through the launch of the Ascend Super Node and CANN ecosystem, which aims to enhance computational power and efficiency in AI applications [1] Group 1: AI City Solutions - Longgang District introduced the AI CITY benchmark solution, leveraging Ascend computing power to integrate government and public service data, focusing on urban management, public services, and industrial empowerment [1] - Traffic efficiency on main roads improved by 18%, and the accuracy of AI-assisted diagnoses in community hospitals exceeded 92% [1] - The "4T Digital Living Space" initiative centers on four digital rights, providing T-level storage, network, public resources, and intelligent computing support, transitioning from "urban governance" to "citizen services" [1] Group 2: Technological Advancements - The Ascend Super Node has overcome traditional cluster communication bottlenecks, achieving training performance three times that of conventional nodes, supporting the training of trillion-parameter models [1] - CANN has been fully open-sourced, promoting deep adaptation of AI frameworks and significantly enhancing development efficiency through automatic operator generation technology [1] - The plan to cultivate 2 million Ascend developers over the next two years aims to strengthen the talent foundation in the industry [1] Group 3: Collaboration and Applications - The conference also announced the second batch of "City + AI" application scenarios in Longgang District, covering 21 fields with a total of 424 scenarios [1] - Longgang Urban Investment signed a cooperation agreement with Huawei to jointly build a local and cloud-based computing resource pool [1]

Group 1 - The conference highlighted the importance of "AI computing power" and introduced the AI CITY benchmark solution, which aims to connect government data, public service data, and citizen needs [1] - The "4T for Home" project envisions a future where ordinary residents can benefit from digital technology, emphasizing the practical application of digital solutions in daily life [1] - The initiative aims to make digital cities more relatable and beneficial to residents, moving away from the perception of smart cities as complex systems [1] Group 2 - The "4T for Home" project promotes the concept of "digital equality," allowing all residents, regardless of their technical knowledge, to access public resources and AI tools [2] - The construction of AI CITY and the "4T for Home" project is expected to face challenges, but the value of AI technology will ultimately be measured by its impact on people's lives [2] - The initiative is supported by significant government investment and aims to create a positive ecosystem where government guidance, technological support, and resident benefits coexist [2]

Core Insights - Huawei's Ascend AI Summit 2025 will take place in Shanghai on September 18, 2025, focusing on the latest technological achievements and innovations in Ascend's hardware, software, and industry solutions, along with sharing successful business cases and insights from partners [1] Group 1: New Technologies - The summit will present significant advancements in Ascend's foundational hardware, software, and industry solutions, showcasing breakthroughs in ultra-node technology, comprehensive open-source software, and innovations across the entire technology chain including pre-training, post-training, and inference solutions [2] Group 2: New Directions - The summit will introduce the "Ultra-Node Development Report," which defines ultra-nodes and outlines their core characteristics, analyzing global industry evolution, stability challenges, and technological ecosystem patterns to provide clear directions for industry development [3] - Practical applications of Ascend ultra-nodes in various sectors such as internet, telecommunications, finance, energy, and manufacturing will be shared, with guest speakers from Midea discussing real-world applications and value [3] Group 3: New Value - The summit will highlight large-scale expert parallel practices across six major industries including telecommunications, government, education, finance, large models, and electricity, with a focus on innovative practices in digital finance by Postal Savings Bank [4] - New products including a large model application integrated machine and 2025 Ascend hardware partner products will be showcased, demonstrating the latest achievements and innovative directions in the Ascend ecosystem [4] - An Ascend MVP appointment ceremony will take place, inviting top industry experts and scholars to provide professional guidance and advice for the development of Ascend AI [4] Group 4: Event Highlights - The Huawei Connect 2025 will feature three major Ascend-themed forums focusing on open-source frameworks and intelligent cluster operations, showcasing how Ascend technology drives AI computing towards greater efficiency and openness [5] - The exhibition area will present the latest Ascend products and solutions, interactive demos, industry case studies, and open speeches, allowing attendees to experience advanced AI technologies and their business applications firsthand [5]

Group 1 - The fourth Global Digital Trade Expo will take place in Hangzhou from September 25 to 29, showcasing cutting-edge technologies and immersive experiences, heavily featuring artificial intelligence [1] - The expo will include a main digital trade exhibition area and seven specialized industry exhibition areas, focusing on sectors such as e-commerce, artificial intelligence, smart transportation, digital entertainment, digital healthcare, smart spaces, and smart logistics [1] - Leading companies and institutions like Alibaba, Huawei, and Zhijiang Laboratory will present their AI innovations, demonstrating capabilities from foundational models to advanced applications [1][2] Group 2 - The AI exhibition area has doubled in size this year, covering a broader industry chain and introducing a dedicated intelligent agent section, featuring the world's first intelligent agent capable of perception, interaction, and reach [2] - Over 80 robotic products will be showcased at this year's expo, significantly increasing from the previous year, with a dedicated robot show area featuring various performances [2] - The "Artificial Intelligence +" concept will be integrated throughout the exhibition, with other specialized areas presenting the latest AI products and applications across various industries, including cross-border e-commerce, green transportation, medical diagnostics, cultural entertainment, and furniture architecture [3]

Group 1 - Huawei announced the open-source release of the Pangu model with 70 billion parameters and the Pro MoE 72B model with 720 billion parameters, enhancing the development of large model technology on domestic computing platforms [1] - The Pro MoE 72B model achieves superior performance comparable to trillion-parameter models by dynamically activating a network of experts, with only 160 billion parameters activated during operation [1] - The latest Super CLUE ranking places Huawei's large model as the top domestic model within the trillion-parameter category, indicating significant advancements in the field [1] Group 2 - Huawei's Ascend chips and CANN heterogeneous computing architecture are part of a fully autonomous and optimized closed-loop solution, marking a shift in global AI computing competition towards large-scale system efficiency and ecosystem development [2] - The Ascend super node has been commercially deployed in data centers such as China Telecom, contributing to the growth of the domestic supply chain [2] Group 3 - Huasen Tiancheng has collaborated with Huawei on Ascend chips and AICC intelligent contact center initiatives, indicating a strategic partnership in AI technology [3] - Softcom Power has launched the Softcom Ascend AI workstation to enhance local AI inference and production efficiency, reflecting the industry's push towards localized AI solutions [4]

Core Insights - The Guangdong-Hong Kong-Macao Greater Bay Area is experiencing strong growth in the computing power industry, with Shaoguan becoming the largest smart computing cluster in South China [1][2] - The total computing power scale of the eight major hubs in China's "East Data West Computing" initiative reached 215.5 EFLOPS, accounting for over 70% of the national total [1] - Shaoguan has established a capacity of 120,000 standard racks and attracted 22 smart computing center projects with a total investment exceeding 60 billion [2] Industry Development - A series of technology innovation platforms were launched at the conference, including the Shaoguan Data Investment & Huawei Joint Innovation Laboratory, which will focus on data elements, talent cultivation, and adaptation centers [3] - The AI Edge Alliance was officially established to promote the integration of AI and edge computing, aiming to inject new vitality into the smart computing industry's upgrade [3] - Shaoguan's network latency has improved significantly, with data transmission to Guangzhou taking only 1.3 ms and to Shenzhen 1.66 ms, meeting most computing power business needs in the Greater Bay Area [2] Investment and Projects - The Greater Bay Area Computing Scheduling Center has been completed, and the Guangzhou Data Exchange (Shaoguan) service base is operational [2] - China Telecom plans to invest over 2.4 billion in Shaoguan to build 38,500 standard racks, while Softcom Power Information Technology will develop a computing power scheduling platform [3]

Core Viewpoint - The article discusses the advancements in AI, particularly focusing on Huawei's innovations in the MoE (Mixture of Experts) architecture and the introduction of RL (Reinforcement Learning) post-training techniques, which aim to enhance the efficiency and performance of large language models (LLMs) in the competitive AI landscape [1][3]. Group 1: MoE Architecture and Huawei's Innovations - The MoE model, originally proposed by Canadian scholars, has evolved significantly, with Huawei introducing the MoGE architecture that addresses inefficiencies in the traditional MoE model, leading to cost reduction and improved training and deployment [1]. - Huawei's approach emphasizes the importance of creating a collaborative ecosystem to foster the growth of the Ascend ecosystem in China [1]. Group 2: RL Post-Training Techniques - RL post-training has emerged as a critical pathway to enhance LLM performance, with models like OpenAI's o1 and DeepSeek-R1 leveraging this technique to improve reasoning capabilities in complex tasks [3][5]. - The RL post-training phase currently consumes 20% of the total computational resources, projected to rise to 50%, significantly impacting model performance and costs [3]. Group 3: Challenges in RL Post-Training - The traditional On-Policy algorithms create a "computational black hole" due to the alternating execution of training and inference tasks, leading to underutilization of resources [6][7]. - The complexity of task scheduling in large-scale clusters, exacerbated by the adoption of various parallel strategies, poses significant challenges for efficient resource utilization [8]. Group 4: Innovations in Resource Utilization - Huawei's RL Fusion technology allows a single card to handle both training and inference tasks simultaneously, effectively doubling resource utilization and throughput [9][10]. - The StaleSync mechanism enables near-asynchronous execution of tasks, achieving over 90% efficiency in horizontal scaling across CloudMatrix 384 super nodes [16][20]. Group 5: Performance Metrics and Results - The combination of RL Fusion and StaleSync has led to a significant increase in efficiency, with single-node throughput improving by 78.5% and overall performance enhancement of 1.5 times [30][31]. - StaleSync's implementation in cluster scaling shows a linear throughput increase from 35k tokens/s to 127k tokens/s as the number of super nodes increases, demonstrating its effectiveness in enhancing scalability [32]. Group 6: Conclusion - The advancements in RL post-training techniques by Huawei represent a significant leap in AI efficiency, positioning the company as a key player in the next generation of AI technology [33].