{"id":48287,"date":"2026-04-07T18:09:14","date_gmt":"2026-04-07T18:09:14","guid":{"rendered":"https:\/\/www.europesays.com\/cz\/48287\/"},"modified":"2026-04-07T18:09:14","modified_gmt":"2026-04-07T18:09:14","slug":"jensen-huang-dekuje-geforce-za-rozsireni-nvidia-cuda-byl-to-riskantni-krok","status":"publish","type":"post","link":"https:\/\/www.europesays.com\/cz\/48287\/","title":{"rendered":"Jensen Huang d\u011bkuje GeForce za roz\u0161\u00ed\u0159en\u00ed Nvidia CUDA. Byl to riskantn\u00ed krok"},"content":{"rendered":"

\u0160\u00e9f Nvidie, Jensen Huang, v ned\u00e1vn\u00e9m rozhovoru zavzpom\u00ednal na riskantn\u00ed krok s uveden\u00edm rozhran\u00ed CUDA do grafick\u00fdch karet GeForce. Tehdy to firmu skoro polo\u017eilo. <\/p>\n

N\u011bkdy riskantn\u00ed kroky vych\u00e1z\u00ed a n\u011bkdy zase ne. Ten od Nvidie, kdy firma p\u0159ed dv\u011bma desetilet\u00edmi p\u0159edstavila rozhran\u00ed CUDA do prvn\u00edch grafick\u00fdch karet GeForce 8800, mezi ty \u0161\u0165astn\u00e9 pat\u0159\u00ed, ne-li mezi ty nej\u0161\u0165astn\u011bj\u0161\u00ed. Firma se toti\u017e sna\u017eila neb\u00fdt jen spole\u010dnost\u00ed pro hern\u00ed GPU, ale firmou pro v\u00fdpo\u010detn\u00ed \u00falohy pomoc\u00ed GPU. Kdy\u017e toti\u017e p\u0159i\u0161ly do hry programovateln\u00e9 pixel shadery, zjistilo se, \u017ee na GPU velmi dob\u0159e b\u011b\u017e\u00ed i n\u011bkter\u00e9 v\u00fdpo\u010detn\u00ed \u00falohy (zpravidla v\u011bdeck\u00e9), kter\u00e9 nemus\u00ed m\u00edt spojitost s 3D grafikou (obecn\u011b se takov\u00e9 nasazen\u00ed ozna\u010dovalo GPGPU).\u00a0<\/p>\n

Ostatn\u011b pamatuju, jak jsem u\u017e v roce 2008 (pouh\u00fd rok po ofici\u00e1ln\u00edm uveden\u00ed technologie CUDA) kupoval grafiku GeForce<\/a> pro v\u00fdvoj akcelerovan\u00fdch algoritm\u016f pro Nvidia CUDA jako velmi slibn\u00e9 technologii do budoucna a v\u00fdvoj pro toto rozhran\u00ed jsem na univerzit\u011b b\u011bhem sv\u00e9 v\u011bdeck\u00e9 kari\u00e9ry je\u0161t\u011b n\u011bkolikr\u00e1t zopakoval.\u00a0<\/p>\n

Jensen Huang podotknul, \u017ee v\u00fdznamn\u00fdm technologick\u00fdm p\u0159\u00ednosem tehdej\u0161\u00ed novinky byla podpora FP32, kter\u00e1 mno\u017estv\u00ed \u00faloh, pro kter\u00e9 se dala akcelerace pomoc\u00ed GPU pou\u017e\u00edt, d\u00e1le v\u00fdznamn\u011b roz\u0161\u00ed\u0159ila (do t\u00e9 doby pou\u017e\u00edvan\u00e9 technologie GPGPU v\u00fdpo\u010dt\u016f si musely poradit bez t\u00e9to podpory, co\u017e mno\u017estv\u00ed akcelerovateln\u00fdch \u00faloh limitovalo).\u00a0<\/p>\n

Podpora Nvidia CUDA v GeForce byl podle Jensena velmi riskantn\u00ed krok. Z hrub\u00e9 mar\u017ee 35 % si t\u00edmto v Nvidii nav\u00fd\u0161ili n\u00e1klady o 50 % a firma to m\u00e1lem nep\u0159e\u017eila. Hodnota firmy dle n\u011bj poklesla a\u017e na pouh\u00fdch 1,5 miliardy USD (pro zaj\u00edmavost, dnes je 4,3 bilion\u016f USD, tedy zhruba 3000 kr\u00e1t v\u00edce). Jensen tak dnes d\u011bkuje kart\u00e1m GeForce, \u017ee pomohly roz\u0161\u00ed\u0159it technologii CUDA za hranice hern\u00edch GPU.\u00a0<\/p>\n

I kdy\u017e s\u00e1zka se finan\u010dn\u011b za\u010dala zhmot\u0148ovat a\u017e o desetilet\u00ed pozd\u011bji, plody pr\u00e1ce jsou jasn\u011b viditeln\u00e9 a Nvidia je u\u017e tak\u0159ka nedosti\u017en\u00e1. Jej\u00ed ekosyst\u00e9m je za ty roky tak roz\u0161\u00ed\u0159en\u00fd, pou\u017e\u00edvan\u00fd a v mnoha kruz\u00edch naprosto standardn\u00ed, \u017ee konkurence u\u017e ji jen lep\u0161\u00edm hardwarem nem\u016f\u017ee jen taj dohnat (pokud to je v\u016fbec mo\u017en\u00e9 ud\u011blat jen lep\u0161\u00edm HW), a je jen ot\u00e1zkou, zda by to s takovou historicky roz\u0161\u00ed\u0159enou SW podporou dok\u00e1zala i v kombinaci lep\u0161\u00edho softwaru.<\/p>\n","protected":false},"excerpt":{"rendered":"\u0160\u00e9f Nvidie, Jensen Huang, v ned\u00e1vn\u00e9m rozhovoru zavzpom\u00ednal na riskantn\u00ed krok s uveden\u00edm rozhran\u00ed CUDA do grafick\u00fdch karet…\n","protected":false},"author":2,"featured_media":48288,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[13],"tags":[18163,7702,1902,74,73,78,75,77,76],"class_list":{"0":"post-48287","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-veda-a-technika","8":"tag-cuda","9":"tag-geforce-rtx","10":"tag-novinka","11":"tag-science","12":"tag-science-technology","13":"tag-technika","14":"tag-technology","15":"tag-veda","16":"tag-veda-a-technika"},"share_on_mastodon":{"url":"https:\/\/pubeurope.com\/@cz\/116364768897836681","error":""},"_links":{"self":[{"href":"https:\/\/www.europesays.com\/cz\/wp-json\/wp\/v2\/posts\/48287","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/cz\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/cz\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/cz\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/cz\/wp-json\/wp\/v2\/comments?post=48287"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/cz\/wp-json\/wp\/v2\/posts\/48287\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/cz\/wp-json\/wp\/v2\/media\/48288"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/cz\/wp-json\/wp\/v2\/media?parent=48287"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/cz\/wp-json\/wp\/v2\/categories?post=48287"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/cz\/wp-json\/wp\/v2\/tags?post=48287"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}