Get featured on INDIAai

Contribute your expertise or opinions and become part of the ecosystem!

The race for creating processing chips that can train Artificial Intelligence (AI) systems has largely been dominated by the America chip-giant, Nvidia. Earlier this year, the company unveiled Nvidia A100, which held the title of the world’s largest processor. 

But it looks like it faces tough competition as merely a few months later, the British chip designer Graphcore has unveiled the Colossus MK2 or GC200 IPU, the world’s most complex chip. 

The Colossus MK2 has 59.4 billion transistors and is a significant upgrade from its’ predecessor the Colossus MK1 as it has improved its performance by eight times. In comparison, Nvidia’s A100 only 54 billion transistors. 

One GC200 chip contains 1,472 independent processor cores and 8,832 separate parallel threads, backed by the support of 900MB in-processor RAM. The 1UM2000 comes at a cost of $ 32,450. Four GC200 chips put into a 1U M2000 creates a one petaflop capacity of total AI compute.

Graphcore announced that this pizza-box sized IPU is “completely plug-and-play” and that customers will be able to connect up to 64,000 IPUs together for a total of 16 exaflops of computing power. 

Citadel, a Chicago-based hedge fund, commissioned a detailed independent analysis into the Mk1 and M1000 machine, which was earlier published in the pre-print server arxiv.org. "The IPU architecture and its compute paradigm were co-designed from the ground up specifically to tackle machine intelligence workloads," the researchers state. "For that reason, they incarnate certain design choices that depart radically from more common architectures like CPUs and GPUs, and might be less familiar to the reader."

For some sorts of workloads, such as pseudo-random number generation, the IPU has an advantage over Nvidia’s GPU but state that researchers said it provided a lower quality of randomness. "The IPU’s performance advantage over the GPU doubles in a per-board comparison. We are not qualified to judge which platform the performance-quality trade-off favours."

Graphcore’s capacity to challenge the chip-giant Nvidia comes from the impressive list of backers like Microsoft, BMW's i Ventures, Dell Technologies, Samsung Electronics, and Demis Hassabis, the co-founder of Google's DeepMind. Earlier this year, the company raised $ 150 million, bringing its final valuation at $ 1.95 billion! 

J.P. Morgan is currently evaluating Graphcore’s chips for possible adoption. 

Want to publish your content?

Publish an article and share your insights to the world.

Get Published Icon
ALSO EXPLORE

DISCLAIMER

The information provided on this page has been procured through secondary sources. In case you would like to suggest any update, please write to us at support.ai@mail.nasscom.in