Bigger isn’t at all times higher, however typically it’s. Cerebras Systems, an organization bent on accelerating machine studying techniques, constructed the world’s largest chip final yr. In the time since, it’s developed bespoke options to among the largest issues within the AI trade.
Founded in 2015, Cerebras is a kind of reunion tour for many of its C-suite executives. Prior to constructing chips the dimensions of dinner plates, the group was accountable for Sea Micro, an organization based in 2007 that finally bought to AMD for greater than $330 million in 2012.
Cerebras Systems is now valued at greater than $4 billion.
I interviewed Cerebras Systems co-founder and CEO Andrew Feldman to learn the way his firm was bellying as much as the {hardware} bar in 2022 — an period marked by huge tech’s domination of the AI trade.
Feldman appeared unaffected by the prospect of competitors from Silicon Valley’s elite.
According to him, Cerebras was constructed from the bottom as much as tackle all comers:
From day one we wished to be within the Computer History Museum. We wished to be within the corridor of fame.
That may sound like hubris to the outsider, however the dog-eat-dog world (or extra aptly, big-dog-buys-little-dog-and-absorbs-its-technology world) of laptop {hardware} isn’t any place for the meek.
In order for Cerebras to make its mark, Feldman and his group relied on their confirmed monitor document of success and ultra-high ambitions to safe entry into the market.
Per Feldman:
We’re system builders, which suggests we are unafraid of inauspicious issues… for those who’re 100-times higher than (huge tech’s finest efforts) then there’s nothing they will do.
So the large query is, are Cererbras’ humongous chips that a lot better than the competitors’s?
The reply: there actually isn’t any competitors. Its chips are greater than 50X bigger than the most important GPU available on the market (the sort of chip usually used to coach machine studying techniques).
According to Cerebras’ web site, the corporate’s CS-2 system, constructed on its WSE-2 “wafer” chips, can substitute a complete rack of processors usually used to coach AI fashions:
A single CS-2 usually delivers the wall-clock compute efficiency of many tens to a whole bunch of graphics processing models (GPU), or extra. In one system lower than one rack in measurement, the CS-2 delivers solutions in minutes or hours that may take days, weeks, or longer on massive multi-rack clusters of legacy, normal objective processors.
Based on the benchmarks the corporate’s revealed, the CS-2 not solely presents efficiency benefits, but it surely additionally has the potential to massively scale back the quantity of energy consumed throughout AI coaching periods. This is a win for companies making an attempt to unravel extraordinarily laborious issues and for the planet.
To study extra about Cerebras Systems, try the white paper for the CS-2 right here and the corporate web site right here.
https://thenextweb.com/news/cerebras-systems-dinner-plate-sized-chips-are-revolutionizing-field-of-ai
