AI accelerator
PulseAugur coverage of AI accelerator — every cluster mentioning AI accelerator across labs, papers, and developer communities, ranked by signal.
4 天有情绪数据
-
New techniques ZipCCL and FlashOverlap accelerate LLM training by optimizing communication
Researchers have developed ZipCCL, a lossless compression library designed to accelerate the distributed training of large language models by addressing communication bottlenecks. The library utilizes novel techniques l…
-
Google Cloud Next 发布新的 TPU 和 Gemini Enterprise Agent Platform
Google Cloud 宣布了新的 AI 创新，包括其第八代 Tensor Processing Units (TPUs)，专为推理和决策设计。该公司还推出了 Gemini Enterprise Agent Platform，旨在增强代理数据能力。这些更新是在 Google Cloud Next 活动中公布的。
-
AMD gains server CPU revenue share; Arm AGI CPUs secure $2B; Meta expands AWS Graviton use
AMD has achieved a significant milestone, capturing 46.2% of the x86 server CPU revenue share in Q1 2026, while Intel maintains a 70.4% share of the overall consumer PC market. In related news, Arm's new AGI CPUs have g…
-
Photonic processors offer energy-efficient alternative for deep learning computations
The future of deep learning may involve photonic processors that use light instead of electrons to perform calculations. This approach aims to reduce the significant energy demands of current neural networks, which rely…
-
Google launches new TPUs for AI training and inference
Google has unveiled its eighth-generation Tensor Processing Units (TPUs), featuring two specialized chips: TPU 8t for training and TPU 8i for inference. These new chips are designed to enhance the capabilities of AI mod…
-
CoreWeave enhances multi-cloud AI stack with Google Cloud interconnect and unified orchestration
CoreWeave has announced a suite of services aimed at simplifying multi-cloud AI infrastructure, including a direct interconnect with Google Cloud to reduce deployment times. The company also introduced SUNK Anywhere, a …
-
Microsoft and Google adapt data center planning for volatile AI demand
Microsoft and Google are adapting their data center planning strategies to accommodate the unpredictable nature of AI workloads. Both companies are moving away from fixed roadmaps towards continuous rebalancing, utilizi…
-
Anthropic 承诺向 Google Cloud 投入 2000 亿美元用于 AI 计算
Anthropic 已承诺在未来五年内斥资约 2000 亿美元购买 Google Cloud 服务和下一代 TPU，预计计算能力将于 2027 年开始上线。这笔交易显著深化了 Anthropic 与 Google 和 Broadcom 现有的合作关系，旨在为其 Claude 模型提供动力并满足不断增长的客户需求。该协议是 Google 的一项重大战略举措，巩固了其定制 TPU 战略和云积压订单，同时也使其能够更直接地与 NVIDIA …
-
Google DeepMind launches WeatherNext 2, an AI model for faster, more accurate weather forecasts
Google DeepMind has unveiled WeatherNext 2, an advanced AI model for weather forecasting that generates predictions 8x faster and with hourly resolution. This new model, built on a Functional Generative Network (FGN) ap…
-
Graphcore and Hugging Face partner to offer new IPU-ready AI models
Graphcore has partnered with Hugging Face to optimize its Intelligence Processing Unit (IPU) hardware for transformer models. This collaboration aims to improve the efficiency and performance of training and deploying l…