加勒比海盗

China, Russia launch fourth joint Arctic scientific expedition, process lasts about 62 days: report_我的网站

A |     

A Russian scientific expedition vessel docks at the port in Russia’s Far Eastern city of Vladivostok. Photo: Xinhua
The fourth China-Russia joint Arctic expedition was launched on Thursday in Russia’s Far Eastern city of Vladivostok. Scientists from both countries will conduct major frontier basic research on the rapid climate changes in the Arctic region, focusing on the area from the Northwest Pacific to the Arctic Ocean, Xinhua News Agency reported on Thursday. Nine people from China will participate in the voyage. 
At the launch ceremony, Wei Zexun, director of the First Institute of Oceanography, Ministry of Natural Resources (MNR), said the joint expedition will focus on the waters of the Northwest Pacific and the Arctic Ocean, further deepening scientific understanding of ocean dynamic processes, climate change patterns, and marine ecosystems in the study areas. It is not only a concrete measure to implement the cooperation consensus between China and Russia in the marine field, but also a vivid example of the two countries’ comprehensive strategic partnership of coordination in marine science and technology in the new era.
According to Du Libin, executive dean of the College of Ocean Science and Engineering, Shandong University of Science and Technology, the Arctic is highly sensitive to global climate change, and a frontier for ocean science, playing a key role in the “Polar Silk Road” initiative and China-Russia polar marine cooperation.
The joint expedition will highlight deeper marine research cooperation and innovation, while advancing polar scientific exploration and practical collaboration between the two countries, Du said. 
According to Zou Jianjun, chief Chinese scientist of the scientific expedition team and a researcher at the First Institute of Oceanography, MNR, the expedition will last about 62 days. The scientists will conduct multidisciplinary surveys in the Chukchi Sea, East Siberian Sea, Lomonosov Ridge, Makarov Basin, and Amundsen Basin to study changes in the marine environment of these regions amid rapid climate change and reveal their response processes and feedback mechanisms in relation to global climate change. 
A representative from the Pacific Oceanological Institute (POI), Far Eastern Branch of the Russian Academy of Sciences, said during the ceremony that the joint expedition is expected to strengthen Russia-China ties and expand cooperation in the Arctic and Pacific regions. 
As freight volumes along the Arctic shipping route are steadily increasing, and the role of Arctic shipping route in the global economy is becoming increasingly important, comprehensive scientific research of the region has become necessary.
Global Times
。    《科创板日报》7月22日讯 近日，英伟达正式发布新一代Spectrum-6以太网交换芯片，用于支持基于Vera Rubin平台的超大规模AI基础设施。         英伟达表示，Spectrum-6交换芯片将作为下一代Spectrum-X Ethernet交换机平台核心组件，通过高带宽、智能流量管理和故障恢复能力，提升AI工厂整体计算效率。包括CoreWeave 、微软、Nebius、SpaceXAI和特斯拉在内的AI基础设施建设者，将成为首批引入Spectrum-6来加速其AI工厂建设的公司。

B |          产品性能方面，Spectrum-6单交换容量达102.4Tbps，是上一代产品的两倍，旨在连接数十万级GPU和CPU，为大规模AI训练、智能体应用及推理任务提供网络支持。基于此，Spectrum-X以太网交换机能够实现比普通以太网高1.6倍的AI网络性能，并且在超过10万个GPU的部署中保持高达95%的网络效率。         此外，Spectrum-6交换芯片同时支持可插拔式光模块和共封装光学（CPO），并支持液冷散热，可为整个AI工厂提供全面的端到端散热方案，同时提高网络电源效率。         截至目前，基于Spectrum-6的交换机已经进入全面生产阶段，且在全球千兆级AI工厂中普及。         中国银河指出，Spectrum-X Ethernet Photonics是全球首款200G/lane大规模量产的CPO交换机。Vera Rubin量产Spectrum-X Ethernet Photonics进一步验证光互连在下一代AI数据中心中的价值。硅光子技术渗透率的提升也将同步带动上游光芯片需求的高速增长。         ▌交换机成Token生成倍增器          交换机是一种用于电光信号转发的网络设备，而在英伟达看来，AI性能本质上是网络问题。

C |          其指出，全球最先进的人工智能工厂汇集了数十万个GPU和CPU，用于训练前沿模型、驱动智能体人工智能，并以前所未有的规模生成智能。在这个层面上，网络成为推动Token生成的关键计算能力倍增器。         对交换机领域的重视，也让这家科技巨头收获了客户与收入增长。         作为Spectrum-6交换芯片的首批用户，CoreWeave表示：“将英伟达 Spectrum-6和液冷式Spectrum-X以太网基础设施引入AI工厂，有助于我们为客户提供训练前沿模型和更快部署推理所需的带宽、弹性和效率。”          据IDC最新数据，英伟达在2026财年第一季度首次成为全球数据中心以太网交换机市场的营收冠军。在过去，该领域由博通、思科等网络巨头主导。         开源证券表示，从交换芯片和GPU配比来看，传统架构交换机与GPU配比约为3:64，而英伟达NVL72包含72颗GPU和9个交换托盘，GBNVL72柜内交换托盘各配备2颗交换芯片，交换芯片与GPU配比达1:4，最新VRNVL72柜内交换托盘配备交换芯片数翻倍，交换芯片与GPU配比进一步提升至1:2，随着Scaleout和Scaleup双线演进，交换网络在计算集群中所占权重或逐步提升。         该机构强调，交换网络正算力配套设施配角，跃升为集群核心主角。随着AI模型参数扩张，算力需求已从单纯GPU堆叠，转向全维度系统架构重构。受单芯片物理功耗密度、互连带宽及内存容量瓶颈制约，算力增长边际效益持续递减，交换网络成为制约算力的天花板，当前系统级协同架构（如高带宽域互联）是突破单芯片性能上限的主要技术路径。

D |          （科创板日报 张真）。

Current article:http://lyupqv.cuozubeishenjingli.cfd/list_mwzfzcw/9uce5yn.html

Published on:15:20:36