Latest Updates NVIDIA Kicks Off the Next Generation of AI With Rubin
NVIDIA today kickstarted the next generation of AI with the launch of the NVIDIA Rubin platform, comprising six new chips designed to deliver one
Latest Updates Understanding Supercomputing: OxRSE Training
Understanding Supercomputing - Performance In supercomputing, we are normally interested in numerical computations: what is the answer to 0.234 + 3.456, or
Latest Updates The evolution of AI interconnects
For more than eight years, Marvell has delivered silicon photonics technology for successive generations of data center optical modules used in
Latest Updates Intel® Rack Scale Design Architecture White Paper
For example, a user may want to compose a node with a 2-socket Intel® Xeon compute module, a 500 Gb Intel® OptaneTM module, an Intel® FPGA module and a 4 TB HDD storage module to meet the
Latest Updates How many optical modules are required for NVIDIA chips?
Hyperscale clusters with hundreds or thousands of GPUs, like AI supercomputing setups, may need thousands of optical modules, especially for 800G or 1.6T links.
Latest Updates How to Build a Supercomputer from Scratch Like a Pro？
Discover how to build a powerful supercomputer from scratch with essential tools, components, and expert tips in this detailed guide.
Latest Updates What Is Supercomputing and How Does It Work?
Supercomputing is a high-performance computing platform that accelerates scientific, AI, and space research, offering benefits in speed, efficiency, and innovation.
Latest Updates Science 101: Supercomputing
Argonne National Laboratory Science 101: Supercomputing There are computers. And then there are super computers. Most personal and work computers are
Latest Updates Supercomputer architecture
A SGI Altix supercomputer with 23,000 processors at the CINES facility in France Approaches to supercomputer architecture have taken dramatic turns since the
Latest Updates Networking chips and modules for AI data centers:
A 25,000-processor cluster needs 75,000 interconnects. A million processors will need 10 million interconnects. More of those connections will be
Latest Updates How to Build a Quantum Supercomputer: Scaling Challenges and
Scaling up to many more tiles would clearly require new designs for a large cryostat. Optical connects Mirhosseini et al.(2023)could provide an alternative solution allowing easier scaling with modular
Latest Updates Understanding Optical Module Demand in Evolving Data
With the continuous evolution of network architectures, the number of optical modules required per server rack has increased significantly. So, how
Latest Updates How Many Optical Modules Does One GPU Need?
Explore the factors influencing the number of optical modules required for GPUs in various networking architectures. Learn about different network card and switch
Latest Updates Benner_FDIP.101024 NAL
For this program, we needed a new generation of optical components – Denser, faster, more configurable, equally reliable, & much more tightly-integrateable into the system
Latest Updates NVIDIA Unveils Next-Gen Rubin and Feynman
Two modules fit on a 1 rack-unit-high, water-cooled motherboard, and 18 compute boards fit in a 72 GPU NVL72 rack, along with switchboards to
Latest Updates Chapter 1
For our installation, we installed only 18 racks|16 compute racks and two I/O racks|with two blanking panels (to control air circulation) at the end of each cold aisle.
Latest Updates directory-list-2.4.txt/directory-list-2.4.txt at main
Customer stories Events & webinars Ebooks & reports Business insights GitHub Skills
Latest Updates AI Data Center Upgrades 2025: Best 400G & 800G
🚀 1. Why AI Data Center Upgrades in 2025 Are All About Optical Speed The explosion in AI and machine learning model sizes, the proliferation of
Latest Updates Nvidia Weaves Silicon Photonics Into InfiniBand And
But it would take twice as much space.) In any event, the scenario that Nvidia chose would have 2.4 million optical transceivers, the pluggable
Latest Updates Why future supercomputing requires optics
Could optical technology offer a solution to the heat generation and bandwidth limitations that the computing industry is starting to face?
Latest Updates Why do AI Data Centers Need 800G Optical Modules?
To meet the growing bandwidth demand, 800G optical modules are becoming a trend. The growth of LPO technology In the 800G optical module era,
Latest Updates Scaling and networking a modular photonic quantum
This combines all the primitive components as discrete, scalable rack-deployed modules networked over fibre-optic interconnects, including 84
Latest Updates AI Data Centers: Scaling Up and Scaling Out
Executive Summary This document provides in-depth commentary on the technical building blocks enabling scaling up and scaling out in modern AI data centers. By highlighting key industry
Latest Updates Get on the Optical Bus
IBM''s light-powered links overcome the greatest speed bump in supercomputing: interconnect bandwidth
Latest Updates Understanding Optical Module Demand in Evolving Data
So, how many optical modules does a data center typically need? In this post, we will explore the usage of optical modules in traditional three-tier,
Latest Updates NVIDIA GTC 2025 Unveils Revolutionary Chips,
A fully populated rack is expected to deliver up to 15 exaflops of FP4 inference performance, marking a 14-fold increase over earlier systems. The new
Latest Updates Data center design requirements for AI workloads. A
Traditional computing racks operate at power densities between 5kW and 15kW, but AI workloads, driven by the use of GPUs, push these requirements
Latest Updates NVIDIA Unveils Next-Gen Rubin and Feynman
Blackwell GB200 modules have around four times the silicon area dedicated to GPU, and the same area for CPU, compared to GH200 modules.
Latest Updates What is Supercomputing?
Supercomputing resources give researchers the processing power needed to investigate trillions of data points at once. Modern supercomputers help