{"id":684,"date":"2023-11-17T08:55:28","date_gmt":"2023-11-17T08:55:28","guid":{"rendered":"https:\/\/tastycounter.net\/index.php\/2023\/11\/17\/microsoft-cuoi-cung-cung-nhay-vao-tu-san-xuat-chip-ai-canh-tranh-nvidia\/"},"modified":"2023-11-17T08:55:28","modified_gmt":"2023-11-17T08:55:28","slug":"microsoft-cuoi-cung-cung-nhay-vao-tu-san-xuat-chip-ai-canh-tranh-nvidia","status":"publish","type":"post","link":"https:\/\/tastycounter.net\/index.php\/2023\/11\/17\/microsoft-cuoi-cung-cung-nhay-vao-tu-san-xuat-chip-ai-canh-tranh-nvidia\/","title":{"rendered":"Microsoft cu\u1ed1i c\u00f9ng c\u0169ng nh\u1ea3y v\u00e0o t\u1ef1 s\u1ea3n xu\u1ea5t chip AI c\u1ea1nh tranh Nvidia"},"content":{"rendered":"<\/p>\n
\n

Ng\u00e0y 15\/11, Microsoft \u0111\u00e3 gi\u1edbi thi\u1ec7u Azure Maia 100 l\u00e0 “m\u1ed9t trong nh\u1eefng m\u1eabu chip ti\u00ean ti\u1ebfn nh\u1ea5t tr\u00ean ti\u1ebfn tr\u00ecnh 5 nanomet” v\u1edbi 105 t\u1ef7 b\u00f3ng b\u00e1n d\u1eabn, t\u1ea1i h\u1ed9i ngh\u1ecb d\u00e0nh cho nh\u00e0 ph\u00e1t tri\u1ec3n Ignite \u1edf Seattle.<\/p>\n

Rani Borkar, Ph\u00f3 ch\u1ee7 t\u1ecbch Microsoft cho bi\u1ebft, Maia 100 \u0111ang \u0111\u01b0\u1ee3c th\u1eed nghi\u1ec7m v\u1ec1 kh\u1ea3 n\u0103ng hu\u1ea5n luy\u1ec7n chatbot AI tr\u00ean c\u00f4ng c\u1ee5 t\u00ecm ki\u1ebfm Copilot – t\u00ean m\u1edbi c\u1ee7a Bing Chat, GPT-3.5-Turbo, m\u1ed9t m\u00f4 h\u00ecnh ng\u00f4n ng\u1eef l\u1edbn t\u1eeb OpenAI v\u00e0 tr\u1ee3 l\u00fd m\u00e3 h\u00f3a GitHub Copilot.<\/p>\n

\"H\u1ec7
H\u1ec7 th\u1ed1ng chip Maia 100 v\u00e0 gi\u00e1 \u0111\u1ee1 Sidekicks. \u1ea2nh: Microsoft.<\/figcaption><\/figure>\n

T\u1ea1i s\u1ef1 ki\u1ec7n, Microsoft c\u0169ng c\u00f4ng b\u1ed1 Cobalt 100, d\u1ef1a tr\u00ean ki\u1ebfn tr\u00fac ARM, 64-bit, ch\u1ee9a 128 l\u00f5i \u0111i\u1ec7n to\u00e1n. Chip h\u01b0\u1edbng \u0111\u1ebfn h\u1ec7 th\u1ed1ng th\u1ef1c hi\u1ec7n t\u00e1c v\u1ee5 t\u00ednh to\u00e1n th\u00f4ng th\u01b0\u1eddng. N\u00f3 c\u00f3 hi\u1ec7u su\u1ea5t ho\u1ea1t \u0111\u1ed9ng cao c\u00f9ng kh\u1ea3 n\u0103ng gi\u1ea3m 40% m\u1ee9c ti\u00eau th\u1ee5 \u0111i\u1ec7n n\u0103ng so v\u1edbi c\u00e1c chip ARM kh\u00e1c m\u00e0 c\u00e1c h\u1ec7 th\u1ed1ng \u0111\u00e1m m\u00e2y Azure \u0111ang s\u1eed d\u1ee5ng.<\/p>\n

<\/ins><\/div>\n

Cobalt hi\u1ec7n \u0111ang \u0111\u01b0\u1ee3c Microsoft th\u1eed nghi\u1ec7m c\u00f4ng vi\u1ec7c x\u1eed l\u00fd c\u00e1c h\u1ec7 th\u1ed1ng ph\u1ea7n m\u1ec1m \u0111\u00e1m m\u00e2y g\u1ed3m Microsoft Teams v\u00e0 Azure SQL.<\/p>\n

\"Cobalt<\/p>\n

Tr\u01b0\u1edbc Microsoft, Google v\u00e0 Amazon c\u0169ng \u0111\u00e3 t\u1ef1 s\u1ea3n xu\u1ea5t chip x\u1eed l\u00fd. Google ti\u00ean phong v\u1edbi m\u1eabu Tensor Processing Unit (TPU) v\u00e0o n\u0103m 2016. Theo sau l\u00e0 Amazon v\u1edbi lo\u1ea1t chip g\u1ed3m Graviton, Trainium v\u00e0 Inferentia.<\/p>\n

Microsoft cho ti\u1ebft, d\u00f9 t\u1ef1 t\u1ea1o chip m\u1edbi nh\u01b0ng h\u00e3ng v\u1eabn s\u1ebd h\u1ee3p t\u00e1c v\u1edbi Nvidia v\u00e0 AMD trong vi\u1ec7c trang b\u1ecb chip cho Azure. C\u00f4ng ty d\u1ef1 ki\u1ebfn s\u1ebd trang b\u1ecb cho h\u1ec7 th\u1ed1ng c\u1ee7a m\u00ecnh GPU H200 m\u1edbi nh\u1ea5t c\u1ee7a Nvidia v\u00e0 MI300 c\u1ee7a AMD v\u00e0o n\u0103m t\u1edbi.<\/p>\n

Nvidia c\u00f9ng v\u1eeba m\u1edbi ra m\u1eaft m\u1eabu chip H200 – chip AI m\u1ea1nh nh\u1ea5t th\u1ebf gi\u1edbi. H200 hi\u1ec7u su\u1ea5t g\u1ea7n g\u1ea5p \u0111\u00f4i so v\u1edbi H100 trong th\u1eed nghi\u1ec7m v\u1edbi m\u00f4 h\u00ecnh ng\u00f4n ng\u1eef l\u1edbn Llama 2 c\u1ee7a Meta v\u1edbi 70 t\u1ef7 tham s\u1ed1. Khi ra m\u1eaft v\u00e0o n\u0103m sau, H200 s\u1ebd \u0111\u01b0\u1ee3c trang b\u1ecb tr\u00ean Amazon Web Services, Google Cloud, Microsoft Azure v\u00e0 Oracle.<\/p>\n<\/div>\n","protected":false},"excerpt":{"rendered":"

Ng\u00e0y 15\/11, Microsoft \u0111\u00e3 gi\u1edbi thi\u1ec7u Azure Maia 100 l\u00e0 “m\u1ed9t trong nh\u1eefng m\u1eabu chip ti\u00ean ti\u1ebfn nh\u1ea5t tr\u00ean ti\u1ebfn tr\u00ecnh 5 nanomet” v\u1edbi 105 t\u1ef7 b\u00f3ng b\u00e1n d\u1eabn, t\u1ea1i h\u1ed9i ngh\u1ecb d\u00e0nh cho nh\u00e0 ph\u00e1t tri\u1ec3n Ignite \u1edf Seattle. Rani Borkar, Ph\u00f3 ch\u1ee7 t\u1ecbch Microsoft cho bi\u1ebft, Maia 100 \u0111ang \u0111\u01b0\u1ee3c th\u1eed nghi\u1ec7m v\u1ec1 […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-684","post","type-post","status-publish","format-standard","hentry","category-khong-phan-loai"],"_links":{"self":[{"href":"https:\/\/tastycounter.net\/index.php\/wp-json\/wp\/v2\/posts\/684","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/tastycounter.net\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/tastycounter.net\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/tastycounter.net\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/tastycounter.net\/index.php\/wp-json\/wp\/v2\/comments?post=684"}],"version-history":[{"count":0,"href":"https:\/\/tastycounter.net\/index.php\/wp-json\/wp\/v2\/posts\/684\/revisions"}],"wp:attachment":[{"href":"https:\/\/tastycounter.net\/index.php\/wp-json\/wp\/v2\/media?parent=684"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/tastycounter.net\/index.php\/wp-json\/wp\/v2\/categories?post=684"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/tastycounter.net\/index.php\/wp-json\/wp\/v2\/tags?post=684"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}