Workflow
算力效率
icon
Search documents
广东:对在游戏科技领域取得显著突破的优质项目,给予最高500万元的一次性扶持奖励
news flash· 2025-05-22 06:47
广东出台《关于推动广东网络游戏产业高质量发展的若干政策措施》。鼓励企业围绕游戏领域"卡脖 子"技术开展研发,支持人工智能等前沿科技在游戏研发中的运用,支持游戏先进技术向其他领域转化 应用。重点对虚拟引擎研发、游戏大模型训练、算力效率提升、先进技术转化、数据安全保障等领域进 行支持。鼓励企业与高校、科研机构建立实验室开展联合攻关。省级统筹资金,经公开评选公示,对在 游戏科技领域取得显著突破的优质项目,给予最高500万元的一次性扶持奖励。 ...
联想集团召开2025创新科技大会 充分释放AI价值
Jing Ji Guan Cha Wang· 2025-05-09 01:57
Group 1: AI Infrastructure Innovations - Lenovo is addressing industry pain points by upgrading hybrid AI infrastructure across multiple dimensions, focusing on computing efficiency and energy efficiency [1] - The company introduced four innovative technologies in computing efficiency, including AI inference acceleration algorithms that outperform the best industry solutions by 20%, and an AI compiler optimizer that reduces training and inference costs by over 15% [1] - Lenovo's new version of the heterogeneous intelligent computing platform, Wanquan 3.0, has achieved leading industry results in various high-quality AI cluster scenarios [1] Group 2: Green Computing Solutions - Lenovo has achieved significant energy efficiency improvements through liquid cooling technology, with a new immersion cooling system that doubles the cooling capacity compared to traditional solutions and achieves a system PUE as low as 1.035 [2] - The company launched a dual-optimization operational system for computing services, which enhances cluster resource utilization by 13% and identifies 58% of ineffective instances [2] Group 3: Data Infrastructure and Services - Lenovo's subsidiary, Lenovo Lingtuo, announced a new storage combination aimed at providing high-performance, reliable, and scalable solutions for AI, high-performance computing, big data analytics, and various unstructured data applications [2] - The goal is to create a unified data foundation for enterprises, facilitating a flexible architecture with strong data protection and higher scalability [2] Group 4: Strategic Vision - Lenovo's Vice President emphasized the company's commitment to building a more powerful, efficient, stable, and green hybrid infrastructure to accelerate the implementation of hybrid AI across various industries [3]