{"id":61834,"date":"2021-03-11T20:38:10","date_gmt":"2021-03-11T19:38:10","guid":{"rendered":"https:\/\/www.planet3dnow.de\/cms\/?p=61834"},"modified":"2021-03-11T20:38:10","modified_gmt":"2021-03-11T19:38:10","slug":"ampd-announces-machine-learning-cloud-initiative-built-around-amd-instinct-mi100-accelerators","status":"publish","type":"post","link":"https:\/\/www.planet3dnow.de\/cms\/61834-ampd-announces-machine-learning-cloud-initiative-built-around-amd-instinct-mi100-accelerators\/","title":{"rendered":"AMPD<\/span> Announces \u2018Machine Learning Cloud\u2019 Initiative Built Around AMD<\/span> Instinct\u2122 MI100<\/span> accelerators"},"content":{"rendered":"

- the World\u2019s Fas\u00adtest HPC<\/span> GPU<\/span> for Sci\u00aden\u00adti\u00adfic Research*<\/p>\n

Van\u00adcou\u00adver, Bri\u00adtish Colum\u00adbia \u2013 March 11, 2021 \u2013 AMPD<\/span> Ven\u00adtures Inc. (\u201cAMPD<\/span>\u201d or the \u201cCom\u00adpa\u00adny\u201d, CSE<\/span>: AMPD<\/span>; FRA<\/span>: 2Q0<\/span>) is plea\u00adsed to announ\u00adce a \u2018Machi\u00adne Lear\u00adning Cloud\u2019 initia\u00adti\u00adve desi\u00adgned to cater to the requi\u00adre\u00adments of aca\u00adde\u00admic insti\u00adtu\u00adti\u00adons and com\u00adpa\u00adnies in the arti\u00adfi\u00adci\u00adal intel\u00adli\u00adgence (\u201cAI<\/span>\u201d), machi\u00adne lear\u00adning and deep lear\u00adning sec\u00adtors. The plat\u00adform, fea\u00adturing AMD<\/span> Instinct\u2122 acce\u00adle\u00adra\u00adtors along with the AMD<\/span> ROCm\u2122 open soft\u00adware plat\u00adform, will initi\u00adal\u00adly be hos\u00adted at AMPD<\/span>\u2019s DC1<\/span> data cent\u00adre in Van\u00adcou\u00adver, Bri\u00adtish Colum\u00adbia, but is expec\u00adted to expand into other ter\u00adri\u00adto\u00adries over the coming months.<\/p>\n

<\/a><\/p>\n

AMD<\/span> has been focu\u00adsing on Data Cent\u00adre Deep Lear\u00adning solu\u00adti\u00adons, com\u00adbi\u00adning its AMD<\/span> ROCm\u2122 open soft\u00adware plat\u00adform with its powerful AMD<\/span> Instinct\u2122 acce\u00adle\u00adra\u00adtors. With an open-source tool\u00adset, AMD<\/span> ROCm\u2122 is desi\u00adgned to bring a rich foun\u00adda\u00adti\u00adon to advan\u00adced com\u00adpu\u00adting with the goal of sol\u00adving real-world pro\u00adblems. It is the first open-source exas\u00adca\u00adle-class plat\u00adform for acce\u00adle\u00adra\u00adted com\u00adpu\u00adting that\u2019s also pro\u00adgramming-lan\u00adguage inde\u00adpen\u00addent. It unlocks an array of dif\u00adfe\u00adrent open com\u00adpu\u00adte lan\u00adguages, com\u00adpi\u00adlers, libra\u00adri\u00ades, and tools desi\u00adgned from the ground up to meet the deman\u00adding needs of AI<\/span> sci\u00aden\u00adtists and rese\u00adar\u00adchers \u2014 hel\u00adping to acce\u00adle\u00adra\u00adte code deve\u00adlo\u00adp\u00adment and sol\u00adve the toug\u00adhest chal\u00adlenges in the world today.<\/p>\n

AMPD<\/span> expects to be the first com\u00adpa\u00adny in Cana\u00adda to offer the power of the AMD<\/span> Instinct MI100<\/span> acce\u00adle\u00adra\u00adtors on a ser\u00advice basis. The AMD<\/span> Instinct MI100<\/span> was announ\u00adced in Novem\u00adber 2020 as the world\u2019s fas\u00adtest High-Per\u00adfor\u00admance Com\u00adpu\u00adting (HPC<\/span>) GPU<\/span>, desi\u00adgned with AMD<\/span>\u2019s all-new AMD<\/span> CDNA<\/span>\u2122 archi\u00adtec\u00adtu\u00adre and beco\u00adming the industry\u2019s first data cen\u00adter GPU<\/span> to exceed 10 Tera\u00adflops (FP64<\/span>)*.<\/p>\n

AMPD<\/span> intends to offer sub\u00adsi\u00addi\u00adzed access to the\u00adse excep\u00adtio\u00adnal\u00adly powerful com\u00adpu\u00adte resour\u00adces to the aca\u00adde\u00admic arti\u00adfi\u00adci\u00adal intel\u00adli\u00adgence com\u00admu\u00adni\u00adty in Bri\u00adtish Colum\u00adbia, in addi\u00adti\u00adon to its pro\u00adgrams for com\u00admer\u00adcial cus\u00adto\u00admers who wish to assess their AI<\/span> com\u00adpu\u00adte opti\u00adons in a low-risk envi\u00adron\u00adment. AMPD<\/span> is working with aca\u00adde\u00admic insti\u00adtu\u00adti\u00adons across the pro\u00advin\u00adce, as well as the \u2018AI<\/span> in BC<\/span>\u2019 indus\u00adtry asso\u00adcia\u00adti\u00adon to iden\u00adti\u00adfy pro\u00adgrams that would bene\u00adfit from access to the plat\u00adform, inclu\u00adding a spe\u00adcial empha\u00adsis on pro\u00adgrams desi\u00adgned to increase access to machi\u00adne lear\u00adning tech\u00adno\u00adlo\u00adgy for under\u00adre\u00adpre\u00adsen\u00adted groups with the aim of hel\u00adping ensu\u00adre an inclu\u00adsi\u00adve and diver\u00adse arti\u00adfi\u00adci\u00adal intel\u00adli\u00adgence future.<\/p>\n

\u201c<\/span>It\u2019s clear that AMD<\/span> and AMPD<\/span> share the same com\u00admit\u00adment to the open-source com\u00admu\u00adni\u00adty and open-source tech\u00adno\u00adlo\u00adgies that are the dri\u00adving force behind ROCm\u2122 plat\u00adform inno\u00adva\u00adtions,\u201d said Brad McCre\u00addie, cor\u00adpo\u00adra\u00adte vice pre\u00adsi\u00addent, Data Cen\u00adter GPU<\/span> and Acce\u00adle\u00adra\u00adted Pro\u00adces\u00adsing, AMD<\/span>. \u201cWe are plea\u00adsed to pro\u00advi\u00adde AMPD<\/span> with the power of the\u00adse acce\u00adle\u00adra\u00adtors with the aim of faci\u00adli\u00adta\u00adting access for aca\u00adde\u00admic insti\u00adtu\u00adti\u00adons and rese\u00adar\u00adchers as cost effec\u00adtively as pos\u00adsi\u00adble. Our indus\u00adtry-dif\u00adfe\u00adren\u00adtia\u00adting approach to acce\u00adle\u00adra\u00adted com\u00adpu\u00adte and hete\u00adro\u00adge\u00adneous workload deve\u00adlo\u00adp\u00adment gives users unpre\u00adce\u00adden\u00adted fle\u00adxi\u00adbi\u00adli\u00adty, choice and plat\u00adform auto\u00adno\u00admy, and we are exci\u00adted that AMPD<\/span> plans to offer the\u00adse advan\u00adta\u00adges on a ser\u00advice basis.\u201d<\/p>\n

Addi\u00adtio\u00adnal\u00adly, both AMPD<\/span> and AMD<\/span> will sup\u00adport the com\u00admu\u00adni\u00adty via the ROCm\u2122 Git\u00adHub forums, inclu\u00adding gui\u00addance and sup\u00adport for HIP<\/span> (Hete\u00adro\u00adge\u00adneous-Com\u00adpu\u00adting Inter\u00adface for Por\u00adta\u00adbi\u00adli\u00adty). AMD<\/span> has deve\u00adlo\u00adped a suite of tools desi\u00adgned to ease con\u00adver\u00adsi\u00adon of CUDA<\/span> appli\u00adca\u00adti\u00adons to por\u00adta\u00adble C++ code by auto\u00adma\u00adting much of the con\u00adver\u00adsi\u00adon work via a source-to-source trans\u00adfor\u00adma\u00adti\u00adon from CUDA<\/span> to HIP<\/span>.<\/p>\n

\u201c<\/span>BC<\/span> is home to a cri\u00adti\u00adcal mass of some 200 appli\u00aded AI<\/span> com\u00adpa\u00adnies as well as lea\u00adding aca\u00adde\u00admic and rese\u00adarch pro\u00adgrams,\u201d said James Hurst\u00adhouse, Chief Stra\u00adtegy Offi\u00adcer at AMPD<\/span>. \u201cWe expect that this cut\u00adting-edge GPU-based hos\u00adted envi\u00adron\u00adment pro\u00admo\u00adting open-source com\u00adpu\u00adting opti\u00adons will yield posi\u00adti\u00adve out\u00adco\u00admes for AI<\/span> com\u00adpa\u00adnies and aca\u00adde\u00admic rese\u00adar\u00adchers and we are incre\u00addi\u00adbly plea\u00adsed that a com\u00adpa\u00adny with the pedi\u00adgree of AMD<\/span> has reco\u00adgni\u00adzed AMPD<\/span> as a key go-to-mar\u00adket part\u00adner. We alre\u00ada\u00addy have initia\u00adti\u00adves under way when it comes to the appli\u00adca\u00adti\u00adon of arti\u00adfi\u00adci\u00adal intel\u00adli\u00adgence and machi\u00adne lear\u00adning in other AMPD<\/span> core sec\u00adtors, such as the visu\u00adal effects indus\u00adtry, as demons\u00adtra\u00adted by our joint indus\u00adtry aca\u00adde\u00admic pro\u00adject on arti\u00adfi\u00adci\u00adal intel\u00adli\u00adgence in moti\u00adon cap\u00adtu\u00adre in con\u00adjunc\u00adtion with AMPD<\/span> cus\u00adto\u00admer, Sho\u00adcap Enter\u00adtain\u00adment, and a team of AI<\/span> rese\u00adar\u00adchers at Simon Fraser University.\u201d<\/p>\n

Hurst\u00adhouse recent\u00adly mode\u00adra\u00adted two panel dis\u00adcus\u00adsions on the the\u00adme of arti\u00adfi\u00adci\u00adal intel\u00adli\u00adgence. On March 4th, 2021, he acted as Panel Chair for a ses\u00adsi\u00adon on \u2018How to Build an AI<\/span> Clus\u00adter\u2019 at the \u2018Glo\u00adbal AI<\/span>: How to Build Across Bor\u00adders\u2019 vir\u00adtu\u00adal con\u00adfe\u00adrence, hos\u00adted by CityA\u00adge, with the sup\u00adport of the AI<\/span> in BC<\/span> indus\u00adtry asso\u00adcia\u00adti\u00adon and the Cana\u00adda Korea Busi\u00adness Asso\u00adcia\u00adti\u00adon. Panelists included repre\u00adsen\u00adta\u00adti\u00adves from Mas\u00adter\u00adCard, Cis\u00adco AI<\/span> Lab, and AI<\/span> in BC<\/span>. Hurst\u00adhouse also mode\u00adra\u00adted a panel at the SPARK<\/span> FX<\/span> 2021 Vir\u00adtu\u00adal Con\u00adfe\u00adrence (\u201cSPARK<\/span>\u201d) with the title \u2018Crea\u00adti\u00adve Arti\u00adfi\u00adci\u00adal Intel\u00adli\u00adgence and Movie Making: The Artist in the Loop\u2019.<\/p>\n

The glo\u00adbal arti\u00adfi\u00adci\u00adal intel\u00adli\u00adgence (AI<\/span>) soft\u00adware mar\u00adket is fore\u00adcast to grow rapidly in the coming years, rea\u00adching around USD<\/span> $126 bil\u00adli\u00adon by 2025[1]. The over\u00adall AI<\/span> mar\u00adket includes a wide array of appli\u00adca\u00adti\u00adons such as natu\u00adral lan\u00adguage pro\u00adces\u00adsing, robo\u00adtic pro\u00adcess auto\u00adma\u00adti\u00adon, and machi\u00adne lear\u00adning. This growth is expec\u00adted to repre\u00adsent an oppor\u00adtu\u00adni\u00adty for a mul\u00adti-sec\u00adtor com\u00adpa\u00adny such as AMPD<\/span> that sup\u00adpli\u00ades GPU-based high-per\u00adfor\u00admance com\u00adpu\u00adting envi\u00adron\u00adments to host the\u00adse applications.<\/p>\n

[1] https:\/\/www.statista.com\/statistics\/607716\/worldwide-artificial-intelligence-market-revenues
\nAMD<\/span>, the AMD<\/span> logo, AMD<\/span> Instinct, ROCm, AMD<\/span> CDNA<\/span>, and com\u00adbi\u00adna\u00adti\u00adons the\u00adreof are trade\u00admarks of Advan\u00adced Micro Devices, Inc.
\n*https:\/\/www.amd.com\/en\/press-releases\/2020\u201311-16-amd-announces-world-s-fastest-hpc-accelerator-for-scientific-research<\/p>\n","protected":false},"excerpt":{"rendered":"

- the World\u2019s Fas\u00adtest HPC<\/span> GPU<\/span> for Sci\u00aden\u00adti\u00adfic Research*<\/p>\n

Van\u00adcou\u00adver, Bri\u00adtish Colum\u00adbia \u2013 March 11, 2021 \u2013 AMPD<\/span> Ven\u00adtures Inc. (\u201cAMPD<\/span>\u201d or the \u201cCom\u00adpa\u00adny\u201d, CSE<\/span>: AMPD<\/span>; FRA<\/span>: 2Q0<\/span>) is plea\u00adsed to announ\u00adce a \u2018Machi\u00adne Lear\u00adning Cloud\u2019 initia\u00adti\u00adve desi\u00adgned to cater to the requi\u00adre\u00adments of aca\u00adde\u00admic insti\u00adtu\u00adti\u00adons and com\u00adpa\u00adnies in the arti\u00adfi\u00adci\u00adal intel\u00adli\u00adgence (\u201cAI<\/span>\u201d), machi\u00adne lear\u00adning and deep lear\u00adning sec\u00adtors. The plat\u00adform, fea\u00adturing AMD<\/span> Instinct\u2122 acce\u00adle\u00adra\u00adtors along with the AMD<\/span> ROCm\u2122 open soft\u00adware plat\u00adform, will initi\u00adal\u00adly be hos\u00adted at AMPD<\/span>\u2019s DC1<\/span> data cent\u00adre in Van\u00adcou\u00adver, Bri\u00adtish Colum\u00adbia, but is expec\u00adted to expand into other ter\u00adri\u00adto\u00adries over the coming months.<\/p>\n

<\/a>
\n (\u2026) Wei\u00adter\u00adle\u00adsen \u00bb<\/a><\/p>\n","protected":false},"author":593,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"wp_typography_post_enhancements_disabled":false,"ngg_post_thumbnail":0,"footnotes":""},"categories":[22],"tags":[966,2534,2464],"class_list":["post-61834","post","type-post","status-publish","format-standard","hentry","category-pressemitteilungen","tag-amd","tag-ampd","tag-instinct-mi100","entry"],"share_on_mastodon":{"url":"","error":""},"_links":{"self":[{"href":"https:\/\/www.planet3dnow.de\/cms\/wp-json\/wp\/v2\/posts\/61834","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.planet3dnow.de\/cms\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.planet3dnow.de\/cms\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.planet3dnow.de\/cms\/wp-json\/wp\/v2\/users\/593"}],"replies":[{"embeddable":true,"href":"https:\/\/www.planet3dnow.de\/cms\/wp-json\/wp\/v2\/comments?post=61834"}],"version-history":[{"count":1,"href":"https:\/\/www.planet3dnow.de\/cms\/wp-json\/wp\/v2\/posts\/61834\/revisions"}],"predecessor-version":[{"id":61836,"href":"https:\/\/www.planet3dnow.de\/cms\/wp-json\/wp\/v2\/posts\/61834\/revisions\/61836"}],"wp:attachment":[{"href":"https:\/\/www.planet3dnow.de\/cms\/wp-json\/wp\/v2\/media?parent=61834"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.planet3dnow.de\/cms\/wp-json\/wp\/v2\/categories?post=61834"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.planet3dnow.de\/cms\/wp-json\/wp\/v2\/tags?post=61834"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}