{"id":31358,"date":"2019-10-31T21:40:51","date_gmt":"2019-10-31T18:40:51","guid":{"rendered":"https:\/\/prohoster.info\/blog\/tenzornye-i-rt-yadra-zanimayut-ne-tak-mnogo-mesta-na-graficheskih-protsessorah-nvidia-turing\/"},"modified":"2019-10-31T21:40:51","modified_gmt":"2019-10-31T18:40:51","slug":"tenzornye-i-rt-yadra-zanimayut-ne-tak-mnogo-mesta-na-graficheskih-protsessorah-nvidia-turing","status":"publish","type":"post","link":"https:\/\/prohoster.info\/da\/blog\/news\/tenzornye-i-rt-yadra-zanimayut-ne-tak-mnogo-mesta-na-graficheskih-protsessorah-nvidia-turing","title":{"rendered":"Tensor- og RT-kerner fylder ikke meget p\u00e5 NVIDIA Turing-grafikprocessorer","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"

Allerede under annonceringen af de f\u00f8rste GeForce RTX 20-serie grafikkort antog mange, at de massive dimensioner af Turing grafikkprocessorene skyldtes tilstedev\u00e6relsen af ekstra blokke: RT-kerner og tensor-kerner. Nu har en Reddit-bruger analyseret infrar\u00f8de billeder af Turing TU106 og TU116 grafikkprocessorene og konkluderet, at de nye beregningsblokke ikke optager s\u00e5 meget plads, som man tidligere har antaget.<\/p>\n

\"Tensor-<\/p>\n

Lad os f\u00f8rst minde om, at Turing TU106 grafikkprocessoren er den mindste og mest kompakte chip fra NVIDIA med specielle RT-kerner til ray tracing og tensor-kerner til acceleration af AI-funktioner. Den relaterede grafikkprocessor Turing TU116 mangler disse specielle beregningsblokke, og derfor blev det besluttet at sammenligne netop dem.<\/p>\n

\"Tensor-
\n\"Tensor-<\/p>\n

NVIDIA Turing-grafikprocessorer er opdelt i TPC-blokke, der inkluderer et par streaming multiprocessorer, som allerede indeholder alle beregningskernerne. Det viser sig, at Turing TU106-grafikprocessorens TPC-blokomr\u00e5de kun er 1,95 mm\u00b2 st\u00f8rre end Turing TU116, hvilket svarer til 22 %. Af dette omr\u00e5de udg\u00f8r 1,25 mm\u00b2 tensor-kerner, og kun 0,7 mm\u00b2 er RT-kerner.<\/p>\n

\"Tensor-
\n\"Tensor-<\/p>\n

Det viser sig, at uden de nye tensor- og RT-kerner ville den flagskibs-grafikprocessor Turing TU102, som ligger til grund for GeForce RTX 2080 Ti, ikke have en st\u00f8rrelse p\u00e5 754 mm\u00b2, men 684 mm\u00b2 (36 TPC). Turing TU104, som ligger til grund for GeForce RTX 2080, ville kunne have en st\u00f8rrelse p\u00e5 498 mm\u00b2 i stedet for 545 mm\u00b2 (24 TPC). Som det ses, ville de \u00e6ldre Turing-grafikprocessorer, selv uden tensor- og RT-kerner, stadig v\u00e6re ret store chips. Betydeligt st\u00f8rre end Pascal-grafikprocessorer.<\/p>\n

<\/center>
\n \"Tensor- <\/p>\n

Hvad kan s\u00e5 forklare disse betydelige dimensioner? For det f\u00f8rste er cache-hukommelsen for Turing grafikkprocessorerne blevet \u00f8get. St\u00f8rrelsen af shaderne er ogs\u00e5 blevet \u00f8get, og Turing chipperne har st\u00f8rre instruktionss\u00e6t og stor registerkapacitet. Alt dette har gjort det muligt ikke kun at \u00f8ge arealet, men ogs\u00e5 ydeevnen for Turing grafikkprocessorerne m\u00e6rkbart. For eksempel har den samme GeForce RTX 2060 baseret p\u00e5 TU106 n\u00e6sten det samme ydningsniveau som GeForce GTX 1080 baseret p\u00e5 GP104. Sidstn\u00e6vnte har forresten 25 % flere CUDA-kerner, selvom det optager et omr\u00e5de p\u00e5 314 mm\u00b2 mod 410 mm\u00b2 for den nye TU106. <\/p>\n

<\/p>\n

<\/center><\/p>\n

\t\t\t\t


\n\t\t\t\t\t\t\t\t<\/center>
\n
Kilde: 3dnews.ru<\/a><\/p>","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"

\u0415\u0449\u0451 \u0432\u043e \u0432\u0440\u0435\u043c\u044f \u0430\u043d\u043e\u043d\u0441\u0430 \u043f\u0435\u0440\u0432\u044b\u0445 \u0432\u0438\u0434\u0435\u043e\u043a\u0430\u0440\u0442 GeForce RTX 20-\u0439 \u0441\u0435\u0440\u0438\u0438 \u043c\u043d\u043e\u0433\u0438\u0435 \u043f\u043e\u0441\u0447\u0438\u0442\u0430\u043b\u0438, \u0447\u0442\u043e \u0441\u0432\u043e\u0438\u043c\u0438 \u0441\u043e\u0432\u0441\u0435\u043c \u043d\u0435 \u043c\u0430\u043b\u0435\u043d\u044c\u043a\u0438\u043c \u0433\u0430\u0431\u0430\u0440\u0438\u0442\u0430\u043c\u0438 \u0433\u0440\u0430\u0444\u0438\u0447\u0435\u0441\u043a\u0438\u0435 \u043f\u0440\u043e\u0446\u0435\u0441\u0441\u043e\u0440\u044b Turing \u043e\u0431\u044f\u0437\u0430\u043d\u044b \u043d\u0430\u043b\u0438\u0447\u0438\u044e \u0434\u043e\u043f\u043e\u043b\u043d\u0438\u0442\u0435\u043b\u044c\u043d\u044b\u0445 \u0431\u043b\u043e\u043a\u043e\u0432: RT-\u044f\u0434\u0435\u0440 \u0438 \u0442\u0435\u043d\u0437\u043e\u0440\u043d\u044b\u0445 \u044f\u0434\u0435\u0440. \u0422\u0435\u043f\u0435\u0440\u044c \u0436\u0435 \u043e\u0434\u0438\u043d \u0438\u0437 \u043f\u043e\u043b\u044c\u0437\u043e\u0432\u0430\u0442\u0435\u043b\u0435\u0439 Reddit \u043f\u0440\u043e\u0430\u043d\u0430\u043b\u0438\u0437\u0438\u0440\u043e\u0432\u0430\u043b \u0438\u043d\u0444\u0440\u0430\u043a\u0440\u0430\u0441\u043d\u044b\u0435 \u0441\u043d\u0438\u043c\u043a\u0438 \u0433\u0440\u0430\u0444\u0438\u0447\u0435\u0441\u043a\u0438\u0445 \u043f\u0440\u043e\u0446\u0435\u0441\u0441\u043e\u0440\u043e\u0432 Turing TU106 \u0438 TU116 \u0438 \u0437\u0430\u043a\u043b\u044e\u0447\u0438\u043b, \u0447\u0442\u043e \u043d\u043e\u0432\u044b\u0435 \u0432\u044b\u0447\u0438\u0441\u043b\u0438\u0442\u0435\u043b\u044c\u043d\u044b\u0435 \u0431\u043b\u043e\u043a\u0438 \u0437\u0430\u043d\u0438\u043c\u0430\u044e\u0442 \u043d\u0435 \u0442\u0430\u043a \u043c\u043d\u043e\u0433\u043e \u043c\u0435\u0441\u0442\u0430, […]<\/p>\n","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"author":1,"featured_media":23324,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[702],"tags":[],"class_list":["post-31358","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-news"],"aioseo_notices":[],"aioseo_head":"\n\t\t\n\t\n\t\n\t\n\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\n","aioseo_head_json":{"title":"\ud83e\udd47Tensor- og RT-kernerne fylder ikke meget p\u00e5 NVIDIA Turing grafiske processorer | ProHoster","description":"Allerede under annonceringen af de f\u00f8rste GeForce RTX 20-serie grafikkort mente mange, at de store dimensioner af Turing-grafikprocessorerne skyldes tilstedev\u00e6relsen af \u200b\u200byderligere kerner: RT-kerner og tensor-kerner.","canonical_url":"https:\/\/prohoster.info\/da\/blog\/news\/tenzornye-i-rt-yadra-zanimayut-ne-tak-mnogo-mesta-na-graficheskih-protsessorah-nvidia-turing","robots":"max-image-preview:large","keywords":"","webmasterTools":{"miscellaneous":""},"schema":null,"og:locale":"da_DK","og:site_name":"ProHoster | \u041a\u0443\u043f\u0438\u0442\u044c \u043d\u0430\u0434\u0435\u0436\u043d\u044b\u0439 \u0445\u043e\u0441\u0442\u0438\u043d\u0433 \u0434\u043b\u044f \u0441\u0430\u0439\u0442\u043e\u0432 \u0441 \u0437\u0430\u0449\u0438\u0442\u043e\u0439 \u043e\u0442 DDoS, VPS VDS \u0441\u0435\u0440\u0432\u0435\u0440\u044b","og:type":"article","og:title":"\ud83e\udd47\u0422\u0435\u043d\u0437\u043e\u0440\u043d\u044b\u0435 \u0438 RT-\u044f\u0434\u0440\u0430 \u0437\u0430\u043d\u0438\u043c\u0430\u044e\u0442 \u043d\u0435 \u0442\u0430\u043a \u043c\u043d\u043e\u0433\u043e \u043c\u0435\u0441\u0442\u0430 \u043d\u0430 \u0433\u0440\u0430\u0444\u0438\u0447\u0435\u0441\u043a\u0438\u0445 \u043f\u0440\u043e\u0446\u0435\u0441\u0441\u043e\u0440\u0430\u0445 NVIDIA Turing | ProHoster","og:description":"\u0415\u0449\u0451 \u0432\u043e \u0432\u0440\u0435\u043c\u044f \u0430\u043d\u043e\u043d\u0441\u0430 \u043f\u0435\u0440\u0432\u044b\u0445 \u0432\u0438\u0434\u0435\u043e\u043a\u0430\u0440\u0442 GeForce RTX 20-\u0439 \u0441\u0435\u0440\u0438\u0438 \u043c\u043d\u043e\u0433\u0438\u0435 \u043f\u043e\u0441\u0447\u0438\u0442\u0430\u043b\u0438, \u0447\u0442\u043e \u0441\u0432\u043e\u0438\u043c\u0438 \u0441\u043e\u0432\u0441\u0435\u043c \u043d\u0435 \u043c\u0430\u043b\u0435\u043d\u044c\u043a\u0438\u043c \u0433\u0430\u0431\u0430\u0440\u0438\u0442\u0430\u043c\u0438 \u0433\u0440\u0430\u0444\u0438\u0447\u0435\u0441\u043a\u0438\u0435 \u043f\u0440\u043e\u0446\u0435\u0441\u0441\u043e\u0440\u044b Turing \u043e\u0431\u044f\u0437\u0430\u043d\u044b \u043d\u0430\u043b\u0438\u0447\u0438\u044e \u0434\u043e\u043f\u043e\u043b\u043d\u0438\u0442\u0435\u043b\u044c\u043d\u044b\u0445 \u0431\u043b\u043e\u043a\u043e\u0432: RT-\u044f\u0434\u0435\u0440 \u0438 \u0442\u0435\u043d\u0437\u043e\u0440\u043d\u044b\u0445 \u044f\u0434\u0435\u0440.","og:url":"https:\/\/prohoster.info\/da\/blog\/news\/tenzornye-i-rt-yadra-zanimayut-ne-tak-mnogo-mesta-na-graficheskih-protsessorah-nvidia-turing","og:image":"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg","og:image:secure_url":"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg","og:image:width":350,"og:image:height":350,"article:published_time":"2019-10-31T18:40:51+00:00","article:modified_time":"2019-10-31T18:40:51+00:00","article:publisher":"https:\/\/www.facebook.com\/prohoster","article:author":"https:\/\/www.facebook.com\/prohoster"},"aioseo_meta_data":{"post_id":"31358","title":null,"description":null,"keywords":null,"keyphrases":null,"primary_term":null,"canonical_url":null,"og_title":null,"og_description":null,"og_object_type":"default","og_image_type":"default","og_image_url":null,"og_image_width":null,"og_image_height":null,"og_image_custom_url":null,"og_image_custom_fields":null,"og_video":null,"og_custom_url":null,"og_article_section":null,"og_article_tags":null,"twitter_use_og":false,"twitter_card":"default","twitter_image_type":"default","twitter_image_url":null,"twitter_image_custom_url":null,"twitter_image_custom_fields":null,"twitter_title":null,"twitter_description":null,"schema":{"blockGraphs":[],"customGraphs":[],"default":{"data":{"Article":[],"Course":[],"Dataset":[],"FAQPage":[],"Movie":[],"Person":[],"Product":[],"ProductReview":[],"Car":[],"Recipe":[],"Service":[],"SoftwareApplication":[],"WebPage":[]},"graphName":"","isEnabled":true},"graphs":[]},"schema_type":null,"schema_type_options":null,"pillar_content":false,"robots_default":true,"robots_noindex":false,"robots_noarchive":false,"robots_nosnippet":false,"robots_nofollow":false,"robots_noimageindex":false,"robots_noodp":false,"robots_notranslate":false,"robots_max_snippet":null,"robots_max_videopreview":null,"robots_max_imagepreview":"large","priority":null,"frequency":null,"local_seo":null,"seo_analyzer_scan_date":"2026-01-21 05:47:29","breadcrumb_settings":null,"limit_modified_date":false,"reviewed_by":null,"ai":null,"created":"2021-03-01 03:18:39","updated":"2026-01-21 05:47:29","focus_keyword":null,"additional_keywords":null,"truseo_locale":null},"gt_translate_keys":[{"key":"link","format":"url"}],"_links":{"self":[{"href":"https:\/\/prohoster.info\/da\/wp-json\/wp\/v2\/posts\/31358","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/prohoster.info\/da\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/prohoster.info\/da\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/prohoster.info\/da\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/prohoster.info\/da\/wp-json\/wp\/v2\/comments?post=31358"}],"version-history":[{"count":0,"href":"https:\/\/prohoster.info\/da\/wp-json\/wp\/v2\/posts\/31358\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/prohoster.info\/da\/wp-json\/wp\/v2\/media\/23324"}],"wp:attachment":[{"href":"https:\/\/prohoster.info\/da\/wp-json\/wp\/v2\/media?parent=31358"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/prohoster.info\/da\/wp-json\/wp\/v2\/categories?post=31358"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/prohoster.info\/da\/wp-json\/wp\/v2\/tags?post=31358"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}