Meta 近日披露了其最新的两个24k GPU集群的详尽信息，这些集群将被用于训练其先进的Llama3大语言模型。在一篇名为“构建Meta的GenAI基础设施”的博文中，该公司详述了这两个数据中心规模集群的硬件、网络、存储、设计、性能和软件的细节。这两个新集群是基于2022年推出的Meta AI Research SuperCluster (RSC)进行构建的。
据Meta介绍，这两个新集群各自包含24,576个Nvidia Tensor Core H100 GPU，相较于之前的拥有16,000个Nvidia A100 GPU的集群，这是一个显著的扩展。这一升级使得新集群能够处理更大、更复杂的AI模型，特别是在自然语言处理、语音识别和图像生成等领域，从而推动了生成式AI产品开发的边界。
Meta强调，这些增强的计算能力将为AI研究和开发提供更强大的支持，为未来的AI创新打开新的可能性。随着GPU数量的增加，Meta的基础设施现在能够支持的模型规模更甚以往，为人工智能技术的持续进步铺平道路，预示着公司在AI领域的前沿探索将更加深入。这一进展对于全球AI研究社区而言，无疑是一个重要的里程碑，也彰显了Meta在构建下一代人工智能基础设施上的决心和实力。
英语如下：
News Title: “Meta Constructs 24k GPU Superclusters to Power Llama3 Language Model Training”
Keywords: Meta, GPU clusters, Llama3
News Content: Meta recently unveiled detailed information about its two new 24k GPU clusters, which will be utilized to train its advanced Llama3 large language model. In a blog post titled “Building Meta’s GenAI Infrastructure,” the company outlined the specifics of the hardware, network, storage, design, performance, and software of these data center-scale clusters. These new clusters are built upon the Meta AI Research SuperCluster (RSC) introduced in 2022.
According to Meta, each of these new clusters comprises 24,576 Nvidia Tensor Core H100 GPUs, a substantial expansion from the previous clusters with 16,000 Nvidia A100 GPUs. This upgrade enables the new clusters to handle larger and more complex AI models, particularly in natural language processing, speech recognition, and image generation, pushing the boundaries of generative AI product development.
Meta emphasizes that these enhanced computational capabilities will provide stronger support for AI research and development, opening up new possibilities for future AI innovations. With the increased GPU count, Meta’s infrastructure is now capable of supporting even larger model scales, paving the way for continued advancements in artificial intelligence technology. This progress signifies a significant milestone for the global AI research community and underscores Meta’s commitment and capacity to build the next generation of AI infrastructure.
Undoubtedly, this development deepens Meta’s foray into the forefront of AI exploration, indicating a dedication to pushing the limits of what’s possible in the realm of artificial intelligence.
【来源】https://www.datacenterdynamics.com/en/news/meta-reveals-details-of-two-new-24k-gpu-ai-clusters/
Views: 1