DeepSeek 开源一套昪腾（Ascend）工具组件，对应其 Nvidia 侧的 TileLang、DeepGEMM、DeepEP、TileKernels、FlashMLA 和 DeepSelect，训练中用到的每个 TileLang 算子在昪腾上都有高性能实现，TileLang 承担 V4 系列训练的大部分算子。
DeepSeek released an Ascend toolkit matching its Nvidia components: TileLang, DeepGEMM, DeepEP, TileKernels, FlashMLA and DeepSelect.
It says every TileLang kernel used in its training now has a high-performance Ascend implementation. TileLang powers most kernels in V4-series training.
DeepSeek and Huawei also jointly optimized a 128-card Ascend 950 supernode, with key compute and communication tests reportedly approaching hardware limits.
- TileLang: https://github.com/tile-ai/tilelang
- DeepGEMM Ascend: https://github.com/deepseek-ai/DeepGEMM-Ascend
- DeepEP Ascend: https://github.com/deepseek-ai/DeepEP-Ascend
- TileKernels: https://github.com/deepseek-ai/TileKernels
- FlashMLA: https://github.com/deepseek-ai/FlashMLA
- DeepSelect: https://github.com/deepseek-ai/DeepSelect
来源：X.PIN · x.com