Nvidia Blackwell Ultra Is Official: the First Gpu with Double Grid Design, Nv-Hbi, and 208 Billion Transistors to Master Ai
4 Articles
4 Articles
From generation to generation, Nvidia pushes the limits of massively parallel computing further. After Hopper, who had marked a first break in the acceleration of the generative AI, the Blackwell architecture inaugurates a new stage: that of multi-die chips designed for large-scale "d
With the "Blackwell Ultra," NVIDIA is setting a new benchmark in AI acceleration. While the original Blackwell chip was already considered a radical departure from previous GPU designs, Blackwell Ultra goes even further, with twice the transistor density of Hopper, an aggressive dual-die architecture, and impressive performance scaling up to 15 petaflops in the new NVFP4 format. The chip is […] Source
The path that leads from Blackwell to Rubin includes a step, Blackwell Ultra. NVIDIA detailed the design of this chip in a blog article whose title transcribes the word: Inside NVIDIA Blackwell Ultra: The Chip Powering the AI Factory Era...
NVIDIA has moved sheet and officially presented its Blackwell Ultra GPU architecture as the core of a new era of computation for Artificial Intelligence. And not only that, since the design combines advanced integration of silicon with a software ecosystem in line with the current demands expected by its partners. This GPU is designed to power large-scale data centers (called by the company itself as IA's FAB) and offers a qualitative leap in pe…
Coverage Details
Bias Distribution
- There is no tracked Bias information for the sources covering this story.
Factuality
To view factuality data please Upgrade to Premium