华为将发布AI领域突破性技术 有望解决算力资源利用效率难题

Core Insights - Huawei is set to release a groundbreaking AI technology on November 21, aimed at improving the efficiency of computing resource utilization from the industry average of 30%-40% to 70% [1] - The upcoming technology will unify resource management and utilization of GPUs, NPUs, and other computing resources through software innovation, enhancing support for AI training and inference [1] - The technology shares commonalities with the core technology of Israeli AI startup Run:ai, which was acquired by Nvidia for $700 million by the end of 2024 [1] Group 1 - Huawei's new AI technology will significantly enhance the potential of computing hardware by increasing utilization rates [1] - The technology will mask hardware differences, providing a more efficient resource support for AI applications [1] - Run:ai, founded in 2018, focuses on GPU scheduling technology and aims to create a platform for parallel execution of AI models across various hardware environments [3] Group 2 - Run:ai's core product is a software platform built on Kubernetes, designed for optimizing GPU resource utilization through dynamic scheduling and pooling techniques [3] - The platform is intended to efficiently manage workloads for generative AI, recommendation systems, and search engines, requiring complex scheduling to optimize performance [3]