Models & Labs
NVIDIA Unveils 100% Liquid-Cooled AI Servers
NVIDIA Bloghigh confidence
Why it matters
- →Liquid cooling at higher temperatures reduces energy and water consumption in data centers.
- →The innovation allows for more efficient data center designs, eliminating the need for fans and chillers.
- →It sets a new standard for sustainable AI infrastructure, potentially influencing future data center construction.
NVIDIA has launched the Rubin generation of AI servers, which are the first to achieve 100% liquid cooling. This new infrastructure allows cooling liquids to operate at temperatures up to 45 degrees Celsius, reducing energy and water usage significantly. The design eliminates the need for fans and mechanical chillers, making data centers more efficient and environmentally friendly. This advancement could lead to substantial cost savings and a shift in how data centers are built and operated, particularly in regions with cooler climates.
Read originalMore from NVIDIA Blog
© NVIDIA BlogMarket & Regulationbusiness
NVIDIA Partners Transform Marketing with AI at Cannes Lions
NVIDIA's collaboration with industry leaders at Cannes Lions is reshaping the advertising and marketing landscape through AI. Companies like Alembic and Criteo are leveraging NVIDIA's advanced GPU technology to enhance decision intelligence and ad bidding processes, respectively. Alembic's use of NVIDIA DGX Vera Rubin systems allows for comprehensive causal modeling, providing businesses with clear insights into marketing effectiveness. Meanwhile, Criteo's partnership with NVIDIA has doubled the speed of AI model training, optimizing ad recommendations. This integration of AI into marketing workflows signifies a shift towards more autonomous and efficient operations, setting a new standard for the industry.
NVIDIA Blog
© NVIDIA BlogMarket & Regulationbusiness
France Boosts AI Infrastructure with NVIDIA Technologies
France is making significant strides in AI infrastructure, leveraging NVIDIA technologies to enhance its capabilities. With the construction of a new 44-megawatt data center by Mistral and the deployment of NVIDIA Blackwell instances by Scaleway, France is positioning itself as a key player in Europe's AI landscape. The collaboration with NVIDIA and other partners is fostering the development of open models tailored to local languages and cultural contexts, ensuring compliance with European regulations. This initiative marks a shift from pilot projects to full-scale AI production, promising to accelerate AI adoption across various industries in France.
NVIDIA Blog
More in Models & Labs
Models & Labsmodels
llama.cpp b9745 Release Enhances MTP Support
The latest b9745 release of llama.cpp introduces significant enhancements in multi-threaded processing (MTP) support, particularly with the addition of Step3.5/3.7 flash MTP3. This update includes new APIs like llama_set_mtp_layer_offset and llama_model_n_nextn_layer, which aim to improve the efficiency of multi-head processing. The release also addresses various platform-specific builds, including support for macOS, Linux, Windows, and openEuler, ensuring broader compatibility. While the update doesn't introduce new models, it refines the existing infrastructure, making llama.cpp more robust for developers working with diverse hardware configurations.
llama.cpp Releases
Models & Labsmodels
llama.cpp b9747 release enhances model tracking
The b9747 release of llama.cpp brings a notable improvement with real-time model load progress tracking, enhancing user interaction by offering immediate insights during loading. This update includes server-side improvements such as the addition of a mutex for notify_to_router, which ensures more reliable operations. While there are no new model architectures introduced, the release broadens its reach by supporting platforms like macOS, Linux, and Windows. This makes llama.cpp a more flexible tool for developers working in different environments, although some features like KleidiAI on Apple Silicon are not yet active. The inclusion of ROCm 7.2 and CUDA 12 and 13 DLLs further solidifies its utility across diverse hardware setups.
llama.cpp Releases
Models & Labscoding
Llama.cpp b9754 Release Enhances Grammar Parsing
The latest b9754 release of llama.cpp introduces a more stringent grammar generation with the implementation of an AC parser. This update focuses on refining the parsing capabilities, which could lead to more accurate and reliable code generation. The release also includes various platform-specific builds, such as support for macOS, Linux, Windows, and openEuler, although some features like KleidiAI on Apple Silicon remain disabled. While there are no groundbreaking new features, this release represents a steady improvement in the tool's parsing and platform compatibility.
llama.cpp Releases