DeepSeek is training its models using Huawei's Ascend 950 hardware. This shift follows a statement from Liang Wenfeng made 26 months ago regarding the need for someone to step onto the frontier.
SLAI introduces T-Rex, an end-to-end optimization framework for the full-parameter post-training of trillion-parameter-scale MoE models on Ascend NPU SuperPOD. Using the DeepSeek-V4 model family as the target workload, the system addresses memory pressure and communication overhead through hierarchical parallelism and kernel execution optimizations.
SLAI introduces T-Rex, an end-to-end optimization framework for the full-parameter post-training of trillion-parameter-scale MoE models on Ascend NPU SuperPOD. Using the DeepSeek-V4 model family as the target workload, the system addresses memory pressure and communication overhead through hierarchical parallelism and kernel execution optimizations.
A Reddit user asks why anyone would spend $50,000 to $60,000 on a dual RTX 6000 Pro AI server to run DeepSeek v4 Flash. The author notes that while they previously received feedback that owners do not regret buying the hardware, they are seeking clarification on the logic behind such a high price tag for this specific model.
Huawei has paused its international expansion of artificial intelligence chips because domestic demand within China is exceeding the available supply. This strategic shift means that competitors AMD and Nvidia no longer face immediate pressure from Huawei's global market entry.
A user has developed a custom implementation that allows the DeepSeek V4.1 model to run on NVIDIA A100 GPUs with performance exceeding the official API.
We use cookies to measure traffic and improve the site. You can accept or decline analytics cookies. Privacy policy