{"id":8714,"date":"2025-06-07T01:16:39","date_gmt":"2025-06-06T16:16:39","guid":{"rendered":"https:\/\/aireviewirush.com\/?p=8714"},"modified":"2025-06-07T01:16:39","modified_gmt":"2025-06-06T16:16:39","slug":"sandia-fires-up-a-mind-like-supercomputer-that-can-simulate-180-million-neurons","status":"publish","type":"post","link":"https:\/\/aireviewirush.com\/?p=8714","title":{"rendered":"Sandia Fires Up a Mind-Like Supercomputer That Can Simulate 180 Million Neurons"},"content":{"rendered":"


\n<\/p>\n

\n

Computer systems that function on the identical ideas because the mind may very well be key to slashing AI\u2019s large power payments<\/a>. Sandia Nationwide Laboratories has simply switched on a tool able to simulating between 150 and 180 million neurons.<\/p>\n

The race to construct ever-larger AI fashions<\/a> has yielded large leaps in functionality, nevertheless it\u2019s additionally massively elevated the sources AI requires for coaching and operation. Based on some estimates<\/a>, AI may now account for as a lot as 20 p.c of worldwide datacenter energy demand.<\/p>\n

The human mind<\/a> may present an answer to this rising drawback. The pc inside our heads solves issues past even the most important AI fashions, whereas drawing solely round 20 watts<\/a>. The sector of neuromorphic computing<\/a> is betting laptop {hardware} extra carefully mimicking the mind may assist us match each its energy and power effectivity.<\/p>\n

German startup SpiNNcloud has constructed a neuromorphic supercomputer referred to as SpiNNaker2, based mostly on expertise developed by Steve Furber, designer of ARM\u2019s groundbreaking chip structure. And at this time, Sandia introduced it had formally deployed the machine at its facility in New Mexico.<\/p>\n

\u201cThough GPU-based techniques can increase the effectivity of supercomputers by processing extremely parallel and math-intensive workloads a lot quicker than CPUs, brain-inspired techniques, just like the SpiNNaker2 system, provide an attractive various,\u201d Sandia analysis scientist Craig Winery stated in a press release<\/a>. \u201cThe brand new system delivers each spectacular efficiency and substantial effectivity beneficial properties.\u201d<\/p>\n

The neural networks powering trendy AI<\/a> are already loosely modeled on the mind, however solely at a really rudimentary stage. Neuromorphic computer systems dial up the organic realism with the hope that we will extra carefully replicate a few of the mind\u2019s most engaging qualities.<\/p>\n

In comparison with conventional machines, neuromorphic computer systems mimic the best way the mind communicates utilizing bursts of electrical energy. In typical neural networks, data strikes between neurons within the type of numbers whose worth can fluctuate. In distinction, neuromorphic computer systems use spiking neural networks the place data is contained within the timing of spikes between neurons.<\/p>\n

Within the typical strategy, every neuron prompts each time the community processes information even when the numbers it transmits don\u2019t contribute a lot to the end result. However in a spiking neural community, neurons are solely activated briefly once they have essential data to transmit, which suggests far fewer neurons draw energy at anyone time.<\/p>\n<\/div>\n

\n

You may run a spiking neural community on a traditional laptop, however to essentially see the advantages, you want chips specifically designed to help this novel strategy<\/a>. The SpiNNaker2 system options 1000’s of tiny Arm-based processing cores that function in parallel and talk utilizing very small messages.<\/p>\n

Crucially, the cores aren\u2019t all the time on, like they’d be in a traditional laptop. They\u2019re event-based, which suggests they solely get up and course of information once they obtain a message\u2014or spike\u2014earlier than going again into idle mode. Altogether, SpiNNcloud claims this makes their machine 18 occasions extra power environment friendly than techniques constructed with current graphics processing items (GPUs).<\/p>\n

“Our imaginative and prescient is to pioneer the way forward for synthetic intelligence,” stated Hector A. Gonzalez, cofounder and CEO of SpiNNcloud. \u201cWe\u2019re thrilled to associate with Sandia on this enterprise, and to see the system being delivered to life first-hand.\u201d<\/p>\n

The primary problem dealing with neuromorphic computing is that it operates in essentially alternative ways in comparison with current AI techniques. This makes it tough to translate between the 2 disciplines. A scarcity of software program instruments and supporting infrastructure additionally makes it exhausting to get began.<\/p>\n

However as AI\u2019s power payments mount, the promise of vastly improved power effectivity is a compelling one. This second often is the one neuromorphic computing has been ready for.<\/p>\n<\/div>\n\n","protected":false},"excerpt":{"rendered":"

Computer systems that function on the identical ideas because the mind may very well be key to slashing AI\u2019s large power payments. Sandia Nationwide Laboratories has simply switched on a tool able to simulating between 150 and 180 million neurons. The race to construct ever-larger AI fashions has yielded large leaps in functionality, nevertheless it\u2019s […]<\/p>\n","protected":false},"author":1,"featured_media":8716,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[21],"tags":[],"class_list":["post-8714","post","type-post","status-publish","format-standard","has-post-thumbnail","category-robotics"],"_links":{"self":[{"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/posts\/8714","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=8714"}],"version-history":[{"count":1,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/posts\/8714\/revisions"}],"predecessor-version":[{"id":8715,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/posts\/8714\/revisions\/8715"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/media\/8716"}],"wp:attachment":[{"href":"https:\/\/aireviewirush.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=8714"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=8714"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=8714"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}