Cann
PulseAugur coverage of Cann — every cluster mentioning Cann across labs, papers, and developer communities, ranked by signal.
4 day(s) with sentiment data
-
DeepSeek open-sources DeepInfer for efficient LLM inference
DeepSeek has open-sourced DeepInfer, an inference framework designed for efficient large-scale model deployment. The toolkit supports offline inference for models like DeepSeek-V4.1-Flash on hardware such as Nvidia SM10…
-
DeepSeek open-sources production inference engine, DeepInfer
DeepSeek has open-sourced its production inference engine, DeepInfer, which powers its large-scale services. The released code includes a substantial scheduler, MoE expert-parallel load balancing, speculative decoding, …
-
DeepSeek open-sources DeepInfer inference engine for large-scale AI models
DeepSeek has open-sourced its high-performance inference engine, DeepInfer, which is designed to power large-scale production services for their models. This release includes optimized kernels, support for various paral…
-
Neural networks compared for indoor environment prediction
A new research paper compares two types of geometry-conditioned neural networks, Canvas-based neural networks (CaNN) and coordinate-based neural networks (CoNN), for predicting indoor airflow and temperature. The study …
-
Chinese AI firms buy 1M Nvidia GPUs amid domestic chip shortfall · 2 sources tracked
Chinese AI companies have reportedly purchased over one million Nvidia GPUs within a year to meet the escalating demands of artificial intelligence workloads. Despite advancements in domestic chip production, a signific…
-
DeepSeek, Huawei Open-Source Tools to Challenge Nvidia's AI Dominance · 10 sources tracked
DeepSeek and Huawei have jointly released a suite of open-source tools designed to enhance the software ecosystem for Huawei's Ascend AI chips. This initiative aims to provide a viable alternative to Nvidia's dominant C…
-
Huawei Ascend CANN community leads China with 6,230+ active developers
Huawei's Ascend CANN open-source community has become a leader in China, boasting over 6,230 monthly active developers. The community has achieved more than 60 million repository downloads, with a significant portion of…
-
Huawei open-sources AscendNPU-IR for AI processor optimization
Huawei has released AscendNPU-IR, an open-source intermediate representation tool built on MLIR. This tool is designed to enhance the compilation and optimization of operators for Ascend AI processors, aiming to improve…
-
Alibaba open-sources AI software stack to challenge Nvidia's CUDA
Alibaba's chip design unit, T-Head, is open-sourcing its proprietary software stack to challenge Nvidia's CUDA ecosystem. This move aims to simplify the adoption of Alibaba's Zhenwu AI computing architectures for develo…
-
China's Kunpeng and Ascend platforms enhance developer experience
Chinese computing platforms, Kunpeng and Ascend, are transitioning from basic usability to a more developer-friendly ecosystem. This shift addresses past challenges where developers spent significant time on environment…
-
DeepSeek V4 release signals China's chip-model synergy in AI
DeepSeek's V4 model release highlights a significant shift in China's domestic computing power ecosystem, moving from passive model adaptation to active "chip-model synergy." This collaboration, exemplified by the V4's …
-
Triton language now runs efficiently on Huawei Ascend NPUs
A new compilation framework, Triton-Ascend 3.2.0, has been released to enable the Triton programming language to run efficiently on Huawei's Ascend hardware. This framework simplifies operator development by automating …
-
DeepSeek V4 First Release Adaptation Behind: Why does Ascend insist on not doing a CUDA compatibility layer?
Huawei's Ascend AI accelerators are forging a unique path by eschewing CUDA compatibility to build an independent ecosystem. This strategy focuses on deep architectural changes in their latest Ascend 950 chips to addres…