NVIDIA同时发布了 全新Vera CPU机架设计，单个机柜集成256颗液冷散热CPU，总计45056个线程，同时还有74颗Bluefield-4 DPU、ConnectX SuperNIC网卡，配备最多400TB LPDDR5内存，带宽300TB/s。

英伟达宣称，相较 x86 竞品，其单沙箱性能提升 1.5 倍、每核心内存带宽提升 3 倍、能效翻倍。为实现这些目标，该公司打造了88 核 144 线程CPU，较第一代 Grace 的 72 核有所提升。英伟达还表示，新内核IPC 吞吐量提升 1.5 倍—— 相较于其他竞争架构每代仅个位数或百分之十几的增幅，这是一次巨大的代际飞跃。上一代 Grace 采用公版 Arm Neoverse 内核，而 ...

Intel versus AMD used to feel like an easy call, but building a PC nowadays means looking past old assumptions and comparing what each brand does best.

人工智能的蓬勃发展一直导致各种硬件短缺。最初，GPU的短缺尤为突出，因为数据中心和超大规模数据中心一直在大量采购这些组件，随着显卡供应在2025年年中左右开始趋于正常，由于人工智能数据中心对 高带宽内存 和 企业级存储 ...

The cubicals of the agentic AI age are cores GTC Intel and AMD take notice. At GTC on Monday, Nvidia unveiled its latest ...

As if NVIDIA wasn’t enjoying a lion’s share of the AI chips market, it just unveiled details about its Vera CPU to turn up the heat further on CPU chip makers AMD and Intel trying to shore up the ...

Both AMD and Intel said during the 2026 Morgan Stanley Technology, Media & Telecom Conference that demand for CPUs is seeing an uptick due to artificial in ...

Intel’s market share has been steadily declining as AMD continues to make significant gains in the laptop processor market. According to Stocklytics.com, Team Blue’s market share dropped by 20%, while ...

快科技3月15日消息，近日，俄罗斯微电子企业Tramplin Electronics（跳板电子）宣布，已获得基于中国龙芯LoongArch指令集架构研发的 Irtysh（额尔齐斯河）系列处理器首批样品。这款专为自主数据中心、高性能计算场景打造的CPU ...

Advanced Micro Devices offers significant upside at current levels; the AI accelerator business presents major optionality.

AMD Ryzen CPUs are riding a wave of popularity right now, and the latest best seller charts show them completely annihilating the competition from Intel. What's more, most of the Intel chips that are ...

Intel Corporation INTC has announced that the latest Intel Xeon 6 processors will power the new NVIDIA DGX Rubin NVL8 systems developed by NVIDIA Corporation NVDA. This collaboration highlights the ...