HPC News Bytes 20250203: DeepSeek Lessons, Intel Reroutes GPU Roadmap, LANL and OpenAI for National Security, Nuclear Reactors for Google Data Centers

The HPC-AI world was upended last week by DeepSeek AI benchmark numbers, as the dust settles we offer commentary on what it may, at this stage, mean: Five lessons from DeepSeek, Intel GPU rack scale architecture ….

Power Hungry: Google in Data Center Agreement for Small Modular Nuclear Reactors

IT, like nature, hates a vacuum. Actually, in today’s world, IT is a virtual force of nature. IT’s enormous appetite for power is on a trajectory to strain grids everywhere, and in no IT sector is this more keenly felt than in HPC-AI, the home of generative AI, where insatiable power demand could hobble genAI […]

Nvidia, AMD, Intel and Google Debut Chips in MLPerf Inference Benchmark for GenAI

Today, MLCommons announced new results for its industry-standard MLPerf Inference v4.1 benchmark suite, which delivers machine learning (ML) system performance benchmarking in an architecture-neutral, representative, and ….

Open Compute Project Foundation and Hyperscalers to Trial Low-Carbon ‘Green Concrete’

AUSTIN, Texas, Aug. 20, 2024 — Today, the Open Compute Project Foundation (OCP) announces a collaboration to test development and deployment of low-embodied carbon concrete or “green concrete.” While numerous emerging technologies exist to achieve production of low carbon concrete, adoption has not yet scaled. This proactive and collaborative demonstration project is an important step towards […]

CoreWeave Adds Managers from Google, Oracle

ROSELAND, N.J., Aug. 8, 2024 — GPU cloud company CoreWeave today announced the appointment of Chen Goldberg as Senior Vice President of Engineering, as well as Sachin Jain as its new Chief Operating Officer. Goldberg has more than 25 years of experience leading global engineering teams, product R&D initiatives, and high-profile customer engagements with Fortune 500 enterprises. She joins […]

Ultra Accelerator Link Group for Data Center AI Connectivity Formed: AMD, Broadcom, Cisco, Google, HPE, Intel, Meta and Microsoft

BEAVERTON, Ore.– AMD, Broadcom, Cisco, Google, Hewlett Packard Enterprise (HPE), Intel, Meta and Microsoft today announced they have aligned to develop a new industry standard dedicated to advancing high-speed and low latency communication for scale-up AI systems linking in data centers. Called the Ultra Accelerator Link (UALink), this initial group will define and establish an […]

NVIDIA and Google DeepMind Collaborate on LLMs

Intended to make it easier for developers to create AI-powered applications with world-class performance, NVIDIA and Google today announced three new collaborations at Google I/O ’24. Using TensorRT-LLM, NVIDIA is working with Google to optimize two new models it introduced at the event: Gemma 2 and PaliGemma. These models are built from the same research and […]

MIPS Adds 3 Managers from NVIDIA, Google and SiFive

SAN JOSE – April 09, 2024 – MIPS, a developer of efficient and configurable IP compute cores, today announced the addition of three technology and semiconductor industry professionals dedicated to driving MIPS’ technical differentiation to support the company’s global expansion in the automotive, data center and embedded markets. Reporting directly to CEO Sameer Wasson, the new […]

Google in $1.67B AI Chip Patent Infringement Trial

Designing an AI specialty chip that gains traction is a sure way to riches. That’s why so much money – $1.67 billion – is at stake in a patent infringement lawsuit brought against Google. Reuters reported this week on a federal trial in Boston in which computer scientist Joseph Bates, founder of Singular Computing….

Google Launches AI Supercomputer Powered by Tens of Thousands of NVIDIA H100 GPUs

If a computer’s intelligence can be anthropomorphized, then an AI supercomputer that can scale to 26,000 GPUs (26 exaFLOPS AI throughput) is at the head of the class. That’s the case with Google’s new A3 GPU supercomputers for Google Cloud, introduced at the Google I/O 2023 conference. Google said A3 GPU VMs are designed to […]