NVIDIA’s Marc Hamilton on Building Cambridge-1 Supercomputer During Pandemic


Considering that NVIDIA declared construction of the U.K.’s most powerful AI supercomputer — Cambridge-one — Marc Hamilton, vice president of options architecture and engineering, has been (remotely) overseeing its constructing across the pond.

The process, which will be offered for U.K. health care scientists to perform on pressing challenges, is currently being developed on NVIDIA DGX SuperPOD architecture for a whopping 400 petaflops of AI effectiveness.

Located at Kao Data, a info middle using 100 p.c renewable strength, Cambridge-1 would rank among the world’s major a few most vitality-efficient supercomputers on the newest Environmentally friendly500 checklist.

Hamilton factors to the focus of major healthcare businesses in the U.K. as a most important rationale for NVIDIA’s final decision to build Cambridge-one.

AstraZeneca, GSK, Guy’s and St Thomas’ NHS Basis Belief, King’s College or university London, and Oxford Nanopore have previously introduced their intent to harness the supercomputer for study in the coming months.

Building has been progressing at NVIDIA’s regular pace-of-mild rate, with just final installations and original assessments remaining.

Hamilton promises to provide the most recent updates on Cambridge-1 at GTC 2021.

Crucial Factors From This Episode:

  • Hamilton gives listeners an explainer on Cambridge-1’s scalable models, or constructing blocks — NVIDIA DGX A100 techniques — and how just 20 of them can deliver the equivalent of hundreds of CPUs.
  • NVIDIA intends for Cambridge-one to speed up corporate exploration in addition to that of universities. Among them are King’s School London, which has already declared that it’ll be working with the method.


“With only 20 [DGX A100] servers, you can develop just one of the prime 500 supercomputers in the world” — Marc Hamilton [9:14]

“This is the initially time we’re getting an NVIDIA supercomputer by our engineers and opening it up to our associates, to our customers, to use” — Marc Hamilton [10:17]

You May possibly Also Like:

How AI Can Boost the Analysis and Therapy of Diseases

Drugs — specially radiology and pathology — have become more facts-pushed. The Massachusetts Standard Clinic Center for Clinical Info Science — led by Mark Michalski — promises to speed up that, employing AI systems to spot styles that can boost the detection, diagnosis and treatment method of health conditions.

NVIDIA Main Scientist Monthly bill Dally on Wherever AI Goes Subsequent

This podcast is comprehensive of words from the clever. 1 of the pillars of the pc science environment, NVIDIA’s Invoice Dally joins to share his viewpoint on the world of deep discovering and AI in standard.

The Buck Begins Right here: NVIDIA’s Ian Buck on What is Up coming for AI

Ian Buck, general supervisor of accelerated computing at NVIDIA, shares his insights on how fairly unsophisticated end users can harness AI via the proper software package. Buck helped lay the foundation for GPU computing as a Stanford doctoral prospect, and sent the keynote handle at GTC DC 2019.

Leave a comment

Your email address will not be published.