技术档案

AI GPUAI GPU

人工智能并行计算芯片

每条都摘自下方编号的来源原文；只写事实，不写预测或建议。

技术简介

GPU可用数千个线程同时执行CPU串行执行的数据处理操作。5
Where the CPU executes them serially; GPU executes them across thousands of threads simultaneously.

英伟达GB300 NVL72液冷机柜集成72颗Blackwell Ultra GPU和36颗Grace CPU。1
The NVIDIA GB300 NVL72 features a fully liquid-cooled, rack-scale architecture that integrates 72 NVIDIA Blackwell Ultra GPUs and 36 Arm®-based NVIDIA Grace™ CPUs into a single platform.
据马斯克9月24日披露，Colossus数据中心运行23万颗英伟达AI芯片，Colossus 2运行55万颗。3
24 post on X that his company's Colossus data center in Memphis, Tenn., was running 230,000 Nvidia AI chips, while Colossus 2 had 550,000.
Instinct MI455X GPU是AMD Helios机架架构的关键构件。4
The MI455X GPU will serve as a key building block for the AMD Helios rack-scale architecture, designed to deliver the performance and scalability required for next-generation AI infrastructure.

英伟达生产的GB300 NVL72机柜采用其Blackwell Ultra GPU和Grace CPU。1
The NVIDIA GB300 NVL72 features a fully liquid-cooled, rack-scale architecture that integrates 72 NVIDIA Blackwell Ultra GPUs and 36 Arm®-based NVIDIA Grace™ CPUs into a single platform.
AMD生产面向Helios机架系统AI推理、训练与微调的Instinct MI455X GPU。2
Built on 5th Gen AMD CDNA™ architecture, the AMD Instinct™ MI455X GPU is designed specifically for the AMD Helios rackscale solution offering leadership performance and modern capability for AI inference, training, and fine-tuning.