Graphcore, the UK-based AI chipmaker has unveiled new hardware and software innovations that push the boundaries of research and development in AI. The company has announced the second generation of its flagship Intelligence Processing Unit (IPU) chip, the GC200 or the Colossus MK2.
According to Graphcore, GC200 is the most complex processor ever made. The IPU chip is at the core of every IPU-Machine M2000, a plug-and-play Machine Intelligence compute blade that has been designed for easy deployment and supports systems that can grow to massive scale.
Karl Freund, Senior Analyst at Moor Insights stated, “These developments put Graphcore ‘first in line to challenge NVIDIA for datacenter AI’.”
Developed using TSMC’s latest 7nm process technology, each chip contains more than 59.4 billion transistors on a single 823sqmm die.
GC200 integrates 1,472 separate IPU-Cores and is capable of executing 8,832 separate parallel computing threads. Each IPU processor core gets a performance boost from a set of novel floating-point technologies developed by Graphcore, called AI-Float.
The chip is claimed to deliver some 250 Trillion Operations per Second (TOPS) across 1,472 cores and 900MB of In-Processor Memory interconnected across a 2.8Tb/s low-latency fabric.
According to a blog post, the Colossus IPUs are unique in having support for Stochastic Rounding on the arithmetic that is supported in hardware and runs at the full speed of the processor. This allows the Colossus Mk2 IPU to keep all arithmetic in 16bit formats, reducing memory requirements, saving on reading and writing energy and reducing energy in the arithmetic logic, while delivering full accuracy Machine Intelligence results.
The IPU-Machine M2000 is available standalone or can be purchased in data centre pods, which allow for the connection of up to 64,000 IPU chips to handle the very toughest workloads.
If you loved this story, do join our Telegram Community.
Also, you can write for us and be one of the 500+ experts who have contributed stories at AIM. Share your nominations here.
What's Your Reaction?
A Technical Journalist who loves writing about Machine Learning and Artificial Intelligence. A lover of music, writing and learning something out of the box. Contact: firstname.lastname@example.org