Tag: high-performance

  • The Register: Nvidia GPU roadmap confirms it: Moore’s Law is dead and buried

    Source URL: https://www.theregister.com/2025/03/29/nvidia_moores_law/ Source: The Register Title: Nvidia GPU roadmap confirms it: Moore’s Law is dead and buried Feedly Summary: More silicon, more power, more pain for datacenter operators Comment As Jensen Huang is fond of saying, Moore’s Law is dead – and at Nvidia GTC this month, the GPU-slinger’s chief exec let slip just…

  • Hacker News: Every Flop Counts: Scaling a 300B LLM Without Premium GPUs

    Source URL: https://arxiv.org/abs/2503.05139 Source: Hacker News Title: Every Flop Counts: Scaling a 300B LLM Without Premium GPUs Feedly Summary: Comments AI Summary and Description: Yes Summary: This technical report presents advancements in training large-scale Mixture-of-Experts (MoE) language models, namely Ling-Lite and Ling-Plus, highlighting their efficiency and comparable performance to industry benchmarks while significantly reducing training…

  • Simon Willison’s Weblog: deepseek-ai/DeepSeek-V3-0324

    Source URL: https://simonwillison.net/2025/Mar/24/deepseek/ Source: Simon Willison’s Weblog Title: deepseek-ai/DeepSeek-V3-0324 Feedly Summary: deepseek-ai/DeepSeek-V3-0324 Chinese AI lab DeepSeek just released the latest version of their enormous DeepSeek v3 model, baking the release date into the name DeepSeek-V3-0324. The license is MIT, the README is empty and the release adds up a to a total of 641 GB…

  • Slashdot: US Expands Export Blacklist To Keep Computing Tech Out of China

    Source URL: https://hardware.slashdot.org/story/25/03/26/2053233/us-expands-export-blacklist-to-keep-computing-tech-out-of-china?utm_source=rss1.0mainlinkanon&utm_medium=feed Source: Slashdot Title: US Expands Export Blacklist To Keep Computing Tech Out of China Feedly Summary: AI Summary and Description: Yes Summary: The U.S. government has expanded its export blacklist by adding 80 entities, primarily from China, aiming to prevent the acquisition of advanced American technology for military use, including AI and…

  • The Register: Microsoft walking away from datacenter leases (probably) isn’t a sign the AI bubble is bursting

    Source URL: https://www.theregister.com/2025/03/26/microsoft_ai_apocalypse/ Source: The Register Title: Microsoft walking away from datacenter leases (probably) isn’t a sign the AI bubble is bursting Feedly Summary: Why lease space that can’t power or cool 120kW racks – or the next-gen 600kW monsters? Comment Microsoft has walked away from negotiations to lease two gigawatts worth of datacenter capacity…

  • The Register: Schneider Electric pumps $700M into US ops as AI datacenter demand surges

    Source URL: https://www.theregister.com/2025/03/26/schneider_electric_ai_investment/ Source: The Register Title: Schneider Electric pumps $700M into US ops as AI datacenter demand surges Feedly Summary: Meanwhile, Apple is lining up ‘$1B’ of Nvidia Blackwell Ultra kit Schneider Electric plans to spend $700 million through 2027 to expand its US operations and bolster the supply of its power equipment necessary…

  • Cloud Blog: Speed up checkpoint loading time at scale using Orbax on JAX

    Source URL: https://cloud.google.com/blog/products/compute/unlock-faster-workload-start-time-using-orbax-on-jax/ Source: Cloud Blog Title: Speed up checkpoint loading time at scale using Orbax on JAX Feedly Summary: Imagine training a new AI / ML model like Gemma 3 or Llama 3.3 across hundreds of powerful accelerators like TPUs or GPUs to achieve a scientific breakthrough. You might have a team of powerful…

  • Simon Willison’s Weblog: deepseek-ai/DeepSeek-V3-0324

    Source URL: https://simonwillison.net/2025/Mar/24/deepseek/ Source: Simon Willison’s Weblog Title: deepseek-ai/DeepSeek-V3-0324 Feedly Summary: deepseek-ai/DeepSeek-V3-0324 Chinese AI lab DeepSeek just released the latest version of their enormous DeepSeek v3 model, baking the release date into the name DeepSeek-V3-0324. The license is MIT, the README is empty and the release adds up a to a total of 641 GB…

  • Hacker News: Aiter: AI Tensor Engine for ROCm

    Source URL: https://rocm.blogs.amd.com/software-tools-optimization/aiter:-ai-tensor-engine-for-rocm™/README.html Source: Hacker News Title: Aiter: AI Tensor Engine for ROCm Feedly Summary: Comments AI Summary and Description: Yes Summary: The text discusses AMD’s AI Tensor Engine for ROCm (AITER), emphasizing its capabilities in enhancing performance across various AI workloads. It highlights the ease of integration with existing frameworks and the significant performance…