On the upcoming Sizzling Chips 2024 convention, NVIDIA engineers are set to current groundbreaking improvements geared toward enhancing information middle efficiency and vitality effectivity. The occasion, scheduled for August 25-27 at Stanford College, will function 4 talks by senior NVIDIA engineers, in line with NVIDIA Weblog.
NVIDIA Blackwell Platform
NVIDIA’s shows will spotlight the NVIDIA Blackwell platform, which integrates a number of chips, programs, and NVIDIA CUDA software program to drive the subsequent technology of AI purposes. The NVIDIA GB200 NVL72, a multi-node, liquid-cooled resolution, will even be showcased for its capability to attach 72 Blackwell GPUs and 36 Grace CPUs, setting a brand new benchmark for AI system design. Moreover, the NVLink interconnect know-how can be mentioned for its excessive throughput and low-latency capabilities, essential for generative AI.
Liquid Cooling Developments
A good portion of the convention will give attention to liquid cooling applied sciences. Ali Heydari, director of knowledge middle cooling and infrastructure at NVIDIA, will current designs for hybrid-cooled information facilities. These designs goal to retrofit present air-cooled information facilities with liquid-cooling items, providing a extra environment friendly and space-saving resolution. Heydari’s crew can also be collaborating with the U.S. Division of Vitality on the COOLERCHIPS program to develop superior cooling applied sciences. Utilizing the NVIDIA Omniverse platform, they’re creating digital twins to mannequin vitality consumption and cooling effectivity.
AI-Pushed Chip Design
NVIDIA can also be leveraging AI to reinforce semiconductor design. Mark Ren, director of design automation analysis at NVIDIA, will focus on how AI fashions enhance design high quality and productiveness by automating time-consuming duties. These fashions embody prediction and optimization instruments that assist engineers analyze and enhance designs quickly. AI brokers powered by massive language fashions (LLMs) are being developed to autonomously full duties, work together with designers, and be taught from a database of human and agent experiences. Ren will share examples of AI purposes in timing report evaluation, cell cluster optimization, and code technology.
These shows underscore NVIDIA’s dedication to pushing the boundaries of knowledge middle computing via modern applied sciences in AI and cooling options. The corporate’s efforts goal to ship unparalleled efficiency, effectivity, and optimization throughout varied sectors.
Picture supply: Shutterstock