DeepSeek 开源面向华为昇腾算力平台的 TileLang 编译工具、计算库和分布式通信库，与此前面向英伟达平台的开源组件一一对应。DeepSeek 称，其训练所用的每个 TileLang 算子在昇腾上都有对应的高性能实现；此次开源还涵盖矩阵运算、跨设备通信等组件。
Coverage of the same occurrence shares a node; subsequent developments have their own nodes.
DeepSeek released Ascend components including DeepGEMM and FlashMLA, along with the DeepEP distributed communication library. Huawei shared CANN community guides covering inference deployment, large-scale training and Agentic RL. The source reports measured DeepEP interconnect bandwidth of 375 GB/s for Dispatch and 347 GB/s for Combine.
By original publication time, with each report's bookmarks and feedback preserved.
DeepSeek released Ascend components including DeepGEMM and FlashMLA, along with the DeepEP distributed communication library. Huawei shared CANN community guides covering inference deployment, large-scale training and Agentic RL. The source reports measured DeepEP interconnect bandwidth of 375 GB/s for Dispatch and 347 GB/s for Combine.
同一事件，精选展示《DeepSeek 开源华为昇腾平台基础设施组件》DeepSeek has released open-source TileLang compilation tools, compute libraries and distributed communication libraries for Huawei Ascend, corresponding to components it previously released for Nvidia platforms. DeepSeek says every TileLang operator used in its training has a high-performance Ascend implementation; the release also includes components for matrix operations and cross-device communication.