DeepSeek open-sources Huawei Ascend toolkit with TileLang support, challenging Nvidia's CUDA
kimmonismus · x · 2026-09-30
Per Bloomberg, DeepSeek has released an open-source toolkit for Huawei's Ascend accelerators, including TileLang support optimized for Ascend 950. The tools simplify writing compute kernels — with the compiler handling scheduling and synchronization — potentially making Huawei hardware far more practical for developers outside Nvidia's ecosystem. The author argues the Chinese chip industry matters more than the models: China still depends on Nvidia GPUs but is pushing hard to break free and build a CUDA alternative.
Related event: DeepSeek Open-Sources Full Ascend Infrastructure Stack for Huawei Chips(8 posts)→
More from Infra
- 5 ways to cut LLM costs without changing models: optimize tokens, caching and calls — goyalshaliniuk · 2026-09-30
- DeepSeek Now Training on Huawei Ascend 950 Chips — WebAssemblyMan · 2026-09-30
- FreeToken: open-source engine runs 290B+ MoE models locally on consumer hardware — tom_doerr · 2026-09-30
- EnerTune at SOSP'26 cuts LLM serving energy 1.4-2.3x vs SOTA systems — tianyin_xu · 2026-09-30
- SOSP'26 paper proposes energy-conscious GPU sharing for inference serving, beyond utilization — tianyin_xu · 2026-09-30
- SOSP'26 papers: AgileLog for isolated AI-agent data streams, CXL-LSM for CXL shared memory — tianyin_xu · 2026-09-30