{"id":578205,"date":"2020-05-14T12:45:33","date_gmt":"2020-05-14T10:45:33","guid":{"rendered":"https:\/\/www.tomshardware.fr\/?p=578205"},"modified":"2023-06-22T17:47:10","modified_gmt":"2023-06-22T15:47:10","slug":"le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors","status":"publish","type":"post","link":"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/","title":{"rendered":"L’Ampere A100 de NVIDIA comptabilise 54 milliards de transistors\u00a0!"},"content":{"rendered":"\n

La carte Ampere A100 de NVIDIA, arm\u00e9e d\u2019un GPU Ampere GA100 grav\u00e9 en 7 nm, pourrait bien \u00e9tablir un nouveau record en ce qui concerne le nombre de transistors. En effet, selon de r\u00e9centes informations, elle comporte 54 milliards de transistors, sans compter ceux de la m\u00e9moire HBM2. M\u00eame l\u2019\u00e9norme GPU Xe d\u2019Intel pr\u00e9sent\u00e9 par Raja Koduri il y a quelques jours et modestement baptis\u00e9 “le P\u00e8re de Tous”<\/em> ne rivaliserait avec le monstre de NVIDIA.<\/p>\n\n\n\n

\"Image<\/figure><\/div>\n\n\n\n

Au-del\u00e0 des chiffres, NVIDIA clame des performances en mati\u00e8re d\u2019IA et de calcul simple pr\u00e9cision (FP32) vingt fois sup\u00e9rieures \u00e0 celles d\u2019une carte Tesla V100 bas\u00e9e sur un GPU Volta. D\u2019ailleurs, pour illustrer l\u2019am\u00e9lioration dans le domaine de l\u2019IA, l\u2019entreprise a imagin\u00e9 une nouvelle r\u00e9f\u00e9rence de calcul intitul\u00e9e TF32 (Tensor Float 32). Elle utilise la mantisse de 10 bits du FP16 et l\u2019exposant 8 bits du FP32. En mati\u00e8re de calcul double pr\u00e9cision (FP64), le gain par rapport \u00e0 Volta est de 2,5.<\/p>\n\n\n\n

Une RTX 3080 Ti avec une bande passante m\u00e9moire de 863 Go\/s et une puissance de 21 TLFOPS ?<\/strong><\/a><\/p>\n\n\n\n

Huit cartes dans une solution DGX-A100, pour du 5 PFLOPS<\/h2>\n\n\n\n

Ces cartes Ampere A100 int\u00e8grent notamment la solution DGX-A100. Un syst\u00e8me associant huit cartes A100 qu\u2019on a r\u00e9cemment pu voir dans le four de Jensen Huang, PDG de NVIDIA<\/a>. Une solution DGX-A100 d\u00e9livre une puissance de calcul de 5 PFLOPS. En revanche, \u00e0 moins que vous ne soyez tr\u00e8s riche, n\u2019esp\u00e9rez pas en installer une dans votre ordinateur, puisque chaque pi\u00e8ce se n\u00e9gocie 199 000 dollars. En outre, notez que contrairement \u00e0 ce qu\u2019on sugg\u00e9rait hier, les cartes Ampere A100 ne sont pas compatibles avec les socket GPU Volta V100, comme en atteste la photo fournie par VideoCardz. Le bond technologique est en tout cas bien r\u00e9el, puisque selon les dires de NVIDIA rapport\u00e9s par EETimes, un seul rack de DGX-A100 remplace 25 racks avec CPU de centres de donn\u00e9es, pour un prix dix fois moindre et une consommation d\u2019\u00e9nergie vingt fois inf\u00e9rieure.<\/p>\n\n\n\n

Nos confr\u00e8res de Tom’s Hardware US<\/a> ont rassembl\u00e9 toutes les caract\u00e9ristiques dans le tableau ci-dessous :<\/p>\n\n\n\n

Carte<\/td>NVIDIA Tesla P100<\/td>NVIDIA Tesla V100<\/td>NVIDIA A100<\/td><\/tr>
GPU<\/td>GP100<\/td>GV100<\/td>GA100<\/td><\/tr>
Process gravure TSMC<\/td>16 nm FinFET+<\/td>12 nm FFN<\/td>7 nm N7<\/td><\/tr>
Architecture GPU<\/td>NVIDIA Pascal<\/td>NVIDIA Volta<\/td>NVIDIA Ampere<\/td><\/tr>
Surface GPU<\/td>610 mm\u00b2<\/td>815 mm\u00b2<\/td>826 mm\u00b2<\/td><\/tr>
Nombre de transistors<\/td>15,3 milliards<\/td>21,1 milliards<\/td>54,2 milliards<\/td><\/tr>
Facteur de forme GPU<\/td>SXM<\/td>SXM2<\/td>SXM4<\/td><\/tr>
SMs<\/td>56<\/td>80<\/td>108<\/td><\/tr>
TPCs<\/td>28<\/td>40<\/td>54<\/td><\/tr>
C\u0153urs FP32 \/ SM<\/td>64<\/td>64<\/td>64<\/td><\/tr>
C\u0153urs FP32 \/ GPU<\/td>3584<\/td>5120<\/td>6912<\/td><\/tr>
C\u0153urs FP64 \/ SM<\/td>32<\/td>32<\/td>32<\/td><\/tr>
C\u0153urs FP64 \/ GPU<\/td>1792<\/td>2560<\/td>3456<\/td><\/tr>
C\u0153urs INT32 \/ SM<\/td>NA<\/td>64<\/td>64<\/td><\/tr>
C\u0153urs INT32 \/ GPU<\/td>NA<\/td>5120<\/td>6912<\/td><\/tr>
C\u0153urs Tensor \/ SM<\/td>NA<\/td>8<\/td>4<\/td><\/tr>
C\u0153urs Tensor \/ GPU<\/td>NA<\/td>640<\/td>432<\/td><\/tr>
Fr\u00e9quence Boost GPU<\/td>1480 MHz<\/td>1530 MHz<\/td>1410 MHz<\/td><\/tr>
Pointe FP16 Tensor TFLOPS with FP16 Accumulate<\/td>NA<\/td>125<\/td>312\/624<\/td><\/tr>
Pointe FP16 Tensor TFLOPS avec FP32 accumul\u00e9<\/td>NA<\/td>125<\/td>312\/624<\/td><\/tr>
Pointe BF16 Tensor TFLOPS avec FP32 accumul\u00e9<\/td>NA<\/td>NA<\/td>312\/624<\/td><\/tr>
Pointe TF32 Tensor TFLOPS<\/td>NA<\/td>NA<\/td>156\/312<\/td><\/tr>
Pointe FP64 Tensor TFLOPS<\/td>NA<\/td>NA<\/td>19,5<\/td><\/tr>
Pointe INT8 Tensor TOPS<\/td>NA<\/td>NA<\/td>624\/1248<\/td><\/tr>
Pointe INT4 Tensor TOPS<\/td>NA<\/td>NA<\/td>1248\/2496<\/td><\/tr>
Pointe FP16 TFLOPS<\/td>21,2<\/td>31,4<\/td>78<\/td><\/tr>
Pointe BF16 TFLOPS<\/td>NA<\/td>NA<\/td>39<\/td><\/tr>
Pointe FP32 TFLOPS<\/td>10,6<\/td>15,7<\/td>19,5<\/td><\/tr>
Pointe FP64 TFLOPS<\/td>5,3<\/td>7,8<\/td>9,7<\/td><\/tr>
Pointe INT32 TOPS<\/td>NA<\/td>15,7<\/td>19,5<\/td><\/tr>
Unit\u00e9s de texture<\/td>224<\/td>320<\/td>432<\/td><\/tr>
Taille interface m\u00e9moire<\/td>4096-bit HBM2<\/td>4096-bit HBM2<\/td>5120-bit HBM2<\/td><\/tr>
Quantit\u00e9 m\u00e9moire<\/td>16 GB<\/td>32 GB \/ 16 GB<\/td>40 GB<\/td><\/tr>
Taux de donn\u00e9es m\u00e9moire<\/td>703 MHz DDR<\/td>877,5 MHz DDR<\/td>1215 MHz DDR<\/td><\/tr>
Bande passante m\u00e9moire<\/td>720 Go\/sec<\/td>900 Go\/sec<\/td>1,6 To\/sec<\/td><\/tr>
Taille cache L2<\/td>4 096 Ko<\/td>6 144 Ko<\/td>40 960 Ko<\/td><\/tr>
M\u00e9moire partag\u00e9e \/ SM<\/td>64 Ko<\/td>Configurable jusqu’\u00e0 96 Ko<\/td>Configurable jusqu’\u00e0 164  Ko<\/td><\/tr>
Taille enregistrement fichier \/ SM<\/td>256 Ko<\/td>256 Ko<\/td>256 Ko<\/td><\/tr>
Taille enregistrement fichier \/ GPU<\/td>14 336 Ko<\/td>20 480 Ko<\/td>27 648 Ko<\/td><\/tr>
TDP<\/td>300 Watts<\/td>300 Watts<\/td>400 Watts<\/td><\/tr><\/tbody><\/table><\/div><\/figure>\n\n\n\n
\"Image<\/figure><\/div>\n\n\n\n
\"Image<\/figure><\/div>\n","protected":false},"excerpt":{"rendered":"

Pour des cartes Ampere A100 vingt fois plus performantes que les Tesla V100 en calcul FP32.<\/p>","protected":false},"author":87,"featured_media":578210,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[487,4072],"tags":[538],"hubs":[],"class_list":["post-578205","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-actualite","category-cartes-graphiques","tag-nvidia"],"acf":{"display_mode":"medium-img","post_show_excerpt":true,"post_source":"","hide_sidebar":false,"hide_advertising":false},"yoast_head":"\nL'Ampere A100 de NVIDIA comptabilise 54 milliards de transistors\u00a0!<\/title>\n<meta name=\"description\" content=\"Pour des cartes Ampere A100 vingt fois plus performantes que les Tesla V100 en calcul FP32.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/\" \/>\n<meta property=\"og:locale\" content=\"fr_FR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"L'Ampere A100 de NVIDIA comptabilise 54 milliards de transistors\u00a0!\" \/>\n<meta property=\"og:description\" content=\"Pour des cartes Ampere A100 vingt fois plus performantes que les Tesla V100 en calcul FP32.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/\" \/>\n<meta property=\"og:site_name\" content=\"Tom\u2019s Hardware\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/TomsHardwareFrance\/\" \/>\n<meta property=\"article:published_time\" content=\"2020-05-14T10:45:33+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2023-06-22T15:47:10+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/www.tomshardware.fr\/content\/uploads\/sites\/3\/2020\/05\/nvidia-tesla-a100-1.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"2000\" \/>\n\t<meta property=\"og:image:height\" content=\"1700\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"author\" content=\"R\u00e9mi Bouvet\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@tomshardware_fr\" \/>\n<meta name=\"twitter:site\" content=\"@tomshardware_fr\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/#article\",\"isPartOf\":{\"@id\":\"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/\"},\"author\":{\"name\":\"R\u00e9mi Bouvet\",\"@id\":\"https:\/\/www.tomshardware.fr\/#\/schema\/person\/5ccbd8cbd7e2cc86da6a517f62b4d005\"},\"headline\":\"L’Ampere A100 de NVIDIA comptabilise 54 milliards de transistors\u00a0!\",\"datePublished\":\"2020-05-14T10:45:33+00:00\",\"dateModified\":\"2023-06-22T15:47:10+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/\"},\"wordCount\":624,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\/\/www.tomshardware.fr\/#organization\"},\"image\":{\"@id\":\"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/#primaryimage\"},\"thumbnailUrl\":\"https:\/\/www.tomshardware.fr\/content\/uploads\/sites\/3\/2020\/05\/nvidia-tesla-a100-1.jpg\",\"keywords\":[\"NVIDIA\"],\"articleSection\":[\"Actualit\u00e9\",\"Cartes graphiques\"],\"inLanguage\":\"fr-FR\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/\",\"url\":\"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/\",\"name\":\"L'Ampere A100 de NVIDIA comptabilise 54 milliards de transistors\u00a0!\",\"isPartOf\":{\"@id\":\"https:\/\/www.tomshardware.fr\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/#primaryimage\"},\"image\":{\"@id\":\"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/#primaryimage\"},\"thumbnailUrl\":\"https:\/\/www.tomshardware.fr\/content\/uploads\/sites\/3\/2020\/05\/nvidia-tesla-a100-1.jpg\",\"datePublished\":\"2020-05-14T10:45:33+00:00\",\"dateModified\":\"2023-06-22T15:47:10+00:00\",\"description\":\"Pour des cartes Ampere A100 vingt fois plus performantes que les Tesla V100 en calcul FP32.\",\"inLanguage\":\"fr-FR\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/#primaryimage\",\"url\":\"https:\/\/www.tomshardware.fr\/content\/uploads\/sites\/3\/2020\/05\/nvidia-tesla-a100-1.jpg\",\"contentUrl\":\"https:\/\/www.tomshardware.fr\/content\/uploads\/sites\/3\/2020\/05\/nvidia-tesla-a100-1.jpg\",\"width\":2000,\"height\":1700},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/www.tomshardware.fr\/#website\",\"url\":\"https:\/\/www.tomshardware.fr\/\",\"name\":\"Tom\u2019s Hardware\",\"description\":\"Toute l'info hardware et gaming !\",\"publisher\":{\"@id\":\"https:\/\/www.tomshardware.fr\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/www.tomshardware.fr\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"fr-FR\"},{\"@type\":\"Organization\",\"@id\":\"https:\/\/www.tomshardware.fr\/#organization\",\"name\":\"Tom\u2019s Hardware\",\"url\":\"https:\/\/www.tomshardware.fr\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\/\/www.tomshardware.fr\/#\/schema\/logo\/image\/\",\"url\":\"https:\/\/www.tomshardware.fr\/content\/uploads\/sites\/3\/2023\/06\/th.png\",\"contentUrl\":\"https:\/\/www.tomshardware.fr\/content\/uploads\/sites\/3\/2023\/06\/th.png\",\"width\":1000,\"height\":1000,\"caption\":\"Tom\u2019s Hardware\"},\"image\":{\"@id\":\"https:\/\/www.tomshardware.fr\/#\/schema\/logo\/image\/\"},\"sameAs\":[\"https:\/\/www.facebook.com\/TomsHardwareFrance\/\",\"https:\/\/x.com\/tomshardware_fr\"]},{\"@type\":\"Person\",\"@id\":\"https:\/\/www.tomshardware.fr\/#\/schema\/person\/5ccbd8cbd7e2cc86da6a517f62b4d005\",\"name\":\"R\u00e9mi Bouvet\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\/\/www.tomshardware.fr\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/secure.gravatar.com\/avatar\/a08bb358c2a9fc36265e8f3225c5de08?s=64&d=mm&r=g\",\"contentUrl\":\"https:\/\/secure.gravatar.com\/avatar\/a08bb358c2a9fc36265e8f3225c5de08?s=64&d=mm&r=g\",\"caption\":\"R\u00e9mi Bouvet\"},\"description\":\"Pas du tout convaincu que \\\"c'\u00e9tait mieux avant\\\", je pr\u00e9f\u00e8re m'int\u00e9resser \u00e0 demain plut\u00f4t que fantasmer hier. Alors devant une carte Voodoo, j'aurais certainement la m\u00eame attitude qu'un singe devant une \u0153uvre d'art. Pour finir, je le confesse, je pr\u00e9f\u00e8re Fallout 4 \u00e0 Fallout, Doom 2016 \u00e0 Doom, Civilization VI \u00e0 Civilization, eFootball \u00e0 PE... non, l\u00e0, faut tout de m\u00eame pas d\u00e9conner.\",\"sameAs\":[\"www.linkedin.com\/in\/r\u00e9mi-bouvet-96049a151\"],\"honorificPrefix\":\"Sieur\",\"gender\":\"Homme\",\"knowsLanguage\":[\"Fran\u00e7ais\",\"Anglais\",\"Burgonde\"],\"jobTitle\":\"R\u00e9dacteur\",\"worksFor\":\"Galaxie Media\",\"url\":\"https:\/\/www.tomshardware.fr\/author\/remi\/\"}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"L'Ampere A100 de NVIDIA comptabilise 54 milliards de transistors\u00a0!","description":"Pour des cartes Ampere A100 vingt fois plus performantes que les Tesla V100 en calcul FP32.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/","og_locale":"fr_FR","og_type":"article","og_title":"L'Ampere A100 de NVIDIA comptabilise 54 milliards de transistors\u00a0!","og_description":"Pour des cartes Ampere A100 vingt fois plus performantes que les Tesla V100 en calcul FP32.","og_url":"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/","og_site_name":"Tom\u2019s Hardware","article_publisher":"https:\/\/www.facebook.com\/TomsHardwareFrance\/","article_published_time":"2020-05-14T10:45:33+00:00","article_modified_time":"2023-06-22T15:47:10+00:00","og_image":[{"width":2000,"height":1700,"url":"https:\/\/www.tomshardware.fr\/content\/uploads\/sites\/3\/2020\/05\/nvidia-tesla-a100-1.jpg","type":"image\/jpeg"}],"author":"R\u00e9mi Bouvet","twitter_card":"summary_large_image","twitter_creator":"@tomshardware_fr","twitter_site":"@tomshardware_fr","schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/#article","isPartOf":{"@id":"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/"},"author":{"name":"R\u00e9mi Bouvet","@id":"https:\/\/www.tomshardware.fr\/#\/schema\/person\/5ccbd8cbd7e2cc86da6a517f62b4d005"},"headline":"L’Ampere A100 de NVIDIA comptabilise 54 milliards de transistors\u00a0!","datePublished":"2020-05-14T10:45:33+00:00","dateModified":"2023-06-22T15:47:10+00:00","mainEntityOfPage":{"@id":"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/"},"wordCount":624,"commentCount":0,"publisher":{"@id":"https:\/\/www.tomshardware.fr\/#organization"},"image":{"@id":"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/#primaryimage"},"thumbnailUrl":"https:\/\/www.tomshardware.fr\/content\/uploads\/sites\/3\/2020\/05\/nvidia-tesla-a100-1.jpg","keywords":["NVIDIA"],"articleSection":["Actualit\u00e9","Cartes graphiques"],"inLanguage":"fr-FR","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/#respond"]}]},{"@type":"WebPage","@id":"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/","url":"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/","name":"L'Ampere A100 de NVIDIA comptabilise 54 milliards de transistors\u00a0!","isPartOf":{"@id":"https:\/\/www.tomshardware.fr\/#website"},"primaryImageOfPage":{"@id":"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/#primaryimage"},"image":{"@id":"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/#primaryimage"},"thumbnailUrl":"https:\/\/www.tomshardware.fr\/content\/uploads\/sites\/3\/2020\/05\/nvidia-tesla-a100-1.jpg","datePublished":"2020-05-14T10:45:33+00:00","dateModified":"2023-06-22T15:47:10+00:00","description":"Pour des cartes Ampere A100 vingt fois plus performantes que les Tesla V100 en calcul FP32.","inLanguage":"fr-FR","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/"]}]},{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/www.tomshardware.fr\/le-gpu-ampere-ga100-de-nvidia-comptabilise-54-milliards-de-transistors\/#primaryimage","url":"https:\/\/www.tomshardware.fr\/content\/uploads\/sites\/3\/2020\/05\/nvidia-tesla-a100-1.jpg","contentUrl":"https:\/\/www.tomshardware.fr\/content\/uploads\/sites\/3\/2020\/05\/nvidia-tesla-a100-1.jpg","width":2000,"height":1700},{"@type":"WebSite","@id":"https:\/\/www.tomshardware.fr\/#website","url":"https:\/\/www.tomshardware.fr\/","name":"Tom\u2019s Hardware","description":"Toute l'info hardware et gaming !","publisher":{"@id":"https:\/\/www.tomshardware.fr\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.tomshardware.fr\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"fr-FR"},{"@type":"Organization","@id":"https:\/\/www.tomshardware.fr\/#organization","name":"Tom\u2019s Hardware","url":"https:\/\/www.tomshardware.fr\/","logo":{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/www.tomshardware.fr\/#\/schema\/logo\/image\/","url":"https:\/\/www.tomshardware.fr\/content\/uploads\/sites\/3\/2023\/06\/th.png","contentUrl":"https:\/\/www.tomshardware.fr\/content\/uploads\/sites\/3\/2023\/06\/th.png","width":1000,"height":1000,"caption":"Tom\u2019s Hardware"},"image":{"@id":"https:\/\/www.tomshardware.fr\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/www.facebook.com\/TomsHardwareFrance\/","https:\/\/x.com\/tomshardware_fr"]},{"@type":"Person","@id":"https:\/\/www.tomshardware.fr\/#\/schema\/person\/5ccbd8cbd7e2cc86da6a517f62b4d005","name":"R\u00e9mi Bouvet","image":{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/www.tomshardware.fr\/#\/schema\/person\/image\/","url":"https:\/\/secure.gravatar.com\/avatar\/a08bb358c2a9fc36265e8f3225c5de08?s=64&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/a08bb358c2a9fc36265e8f3225c5de08?s=64&d=mm&r=g","caption":"R\u00e9mi Bouvet"},"description":"Pas du tout convaincu que \"c'\u00e9tait mieux avant\", je pr\u00e9f\u00e8re m'int\u00e9resser \u00e0 demain plut\u00f4t que fantasmer hier. Alors devant une carte Voodoo, j'aurais certainement la m\u00eame attitude qu'un singe devant une \u0153uvre d'art. Pour finir, je le confesse, je pr\u00e9f\u00e8re Fallout 4 \u00e0 Fallout, Doom 2016 \u00e0 Doom, Civilization VI \u00e0 Civilization, eFootball \u00e0 PE... non, l\u00e0, faut tout de m\u00eame pas d\u00e9conner.","sameAs":["www.linkedin.com\/in\/r\u00e9mi-bouvet-96049a151"],"honorificPrefix":"Sieur","gender":"Homme","knowsLanguage":["Fran\u00e7ais","Anglais","Burgonde"],"jobTitle":"R\u00e9dacteur","worksFor":"Galaxie Media","url":"https:\/\/www.tomshardware.fr\/author\/remi\/"}]}},"_links":{"self":[{"href":"https:\/\/www.tomshardware.fr\/wp-json\/wp\/v2\/posts\/578205"}],"collection":[{"href":"https:\/\/www.tomshardware.fr\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.tomshardware.fr\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.tomshardware.fr\/wp-json\/wp\/v2\/users\/87"}],"replies":[{"embeddable":true,"href":"https:\/\/www.tomshardware.fr\/wp-json\/wp\/v2\/comments?post=578205"}],"version-history":[{"count":0,"href":"https:\/\/www.tomshardware.fr\/wp-json\/wp\/v2\/posts\/578205\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.tomshardware.fr\/wp-json\/wp\/v2\/media\/578210"}],"wp:attachment":[{"href":"https:\/\/www.tomshardware.fr\/wp-json\/wp\/v2\/media?parent=578205"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.tomshardware.fr\/wp-json\/wp\/v2\/categories?post=578205"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.tomshardware.fr\/wp-json\/wp\/v2\/tags?post=578205"},{"taxonomy":"hubs","embeddable":true,"href":"https:\/\/www.tomshardware.fr\/wp-json\/wp\/v2\/hubs?post=578205"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}