{"id":79400,"date":"2024-09-24T18:12:07","date_gmt":"2024-09-24T18:12:07","guid":{"rendered":"https:\/\/ekamu.net\/?p=79400"},"modified":"2024-09-24T18:12:07","modified_gmt":"2024-09-24T18:12:07","slug":"cin-nvidiaya-yerli-bir-alternatif-ariyor-hangi-secenekler-var","status":"publish","type":"post","link":"https:\/\/ekamu.net\/index.php\/2024\/09\/24\/cin-nvidiaya-yerli-bir-alternatif-ariyor-hangi-secenekler-var\/","title":{"rendered":"\u00c7in, Nvidia’ya yerli bir alternatif ar\u0131yor: Hangi se\u00e7enekler var?"},"content":{"rendered":"

<\/span> ABD<\/strong>\u2019nin y\u0131llar boyunca \u00c7in<\/strong>\u2019in yar\u0131 iletken sekt\u00f6r\u00fcne uygulad\u0131\u011f\u0131 yapt\u0131r\u0131mlar, Pekin\u2019i yerli \u00e7ip<\/strong> \u00fcretimini h\u0131zland\u0131rmaya zorlad\u0131. Yapay zeka ve temel modellerin patlamas\u0131, \u00c7in’in \u00e7ip end\u00fcstrisinde lider bir rol oynama hedefini daha da h\u0131zland\u0131rd\u0131. \u015eimdiye kadar yapay zekalar\u0131n e\u011fitilmesini sa\u011flayan Nvidia <\/strong>man\u015fetlerde boy g\u00f6sterse de firma, sadece baz\u0131 \u00e7iplerini \u00c7in\u2019e g\u00f6nderebiliyor. Bu durum \u00c7in\u2019in kendi yerli \u00e7ip giri\u015fimlerine olan ilgiyi yeniden canland\u0131rd\u0131. Peki \u00c7in\u2019den Nvidia\u2019ya bir rakip \u00e7\u0131kabilir mi? Asl\u0131na bakarsan\u0131z baz\u0131 se\u00e7enekler var. <\/figure>\n

Huawei bir numaral\u0131 se\u00e7enek<\/b><\/p>\n

<\/span> Huawei<\/strong>, telekom\u00fcnikasyon altyap\u0131s\u0131ndan t\u00fcketici elektroni\u011fine kadar geni\u015f bir yelpazede faaliyet g\u00f6steren bir teknoloji devi olarak \u00c7in\u2019in yerli \u00e7ip \u00fcretiminde \u00f6ne \u00e7\u0131k\u0131yor. Huawei\u2019nin HiSilicon <\/strong>adl\u0131 \u00e7ip tasar\u0131m birimi zaten bir s\u00fcredir Ascend <\/strong>serisi veri merkezi i\u015flemcileri tasarl\u0131yor. Bu i\u015flemciler, veri merkezlerinde yapay zeka modellerini e\u011fitmek i\u00e7in kullan\u0131lan sunucular\u0131n bir par\u00e7as\u0131 olarak sat\u0131l\u0131yor. Atlas <\/strong>markas\u0131 alt\u0131nda sat\u0131lan bu yapay zeka sunucular\u0131 \u015firketin sekt\u00f6rdeki konumunu g\u00f6steriyor. <\/figure>\n

Huawei, mevcut Ascend 910B<\/strong> \u00e7ipini kullan\u0131yor ve yak\u0131n gelecekte piyasaya s\u00fcrmeyi planlad\u0131\u011f\u0131 Ascend 910C<\/strong> ile Nvidia\u2019n\u0131n H100 <\/strong>\u00fcr\u00fcn\u00fcyle rekabet etmeyi hedefliyor. Huawei\u2019nin bu alanda sadece donan\u0131m de\u011fil, yaz\u0131l\u0131m ekosistemi ve geli\u015ftiriciler i\u00e7in sundu\u011fu ara\u00e7larla da avantaj sa\u011flad\u0131\u011f\u0131 belirtiliyor.<\/p>\n

Alibaba ve Baidu<\/b><\/p>\n

<\/span> \u00c7in\u2019in en b\u00fcy\u00fck internet \u015firketlerinden Baidu<\/strong>, otonom ara\u00e7lar ve sunucular i\u00e7in Kunlun <\/strong>adl\u0131 kendi \u00e7iplerini tasarl\u0131yor. Alibaba <\/strong>ise Hanguang 800<\/strong> adl\u0131 bir yapay zeka \u00e7\u0131kar\u0131m \u00e7ipi geli\u015ftirdi. Ancak her iki \u015firket de Nvidia\u2019n\u0131n \u00e7iplerini sat\u0131n almay\u0131 s\u00fcrd\u00fcr\u00fcyor. Alibaba\u2019n\u0131n bu \u00e7ipi, e-ticaret platformunda \u00f6neri sistemlerini h\u0131zland\u0131rmak i\u00e7in kullan\u0131l\u0131yor. Baidu ise Kunlun \u00e7ipini veri merkezlerinde ve otonom s\u00fcr\u00fc\u015f alan\u0131nda entegre etmi\u015f durumda. <\/figure>\n

Biren Technology<\/b><\/p>\n

<\/span> Biren Technology<\/strong>, Nvidia\u2019ya benzer \u015fekilde genel ama\u00e7l\u0131 GPU\u2019lar tasarl\u0131yor ve bu donan\u0131mlar\u0131n \u00fczerine uygulamalar geli\u015ftirmek i\u00e7in bir yaz\u0131l\u0131m platformu sunuyor. \u015eirketin Bili serisi <\/strong>\u00e7ipleri, veri merkezlerinde yapay zeka e\u011fitimi i\u00e7in kullan\u0131l\u0131yor. Ancak Biren, ABD\u2019nin kara listesine eklenmi\u015f durumda ve dolay\u0131s\u0131yla Amerikan teknolojisine eri\u015fimi k\u0131s\u0131tlanm\u0131\u015f vaziyette. <\/figure>\n

Cambricon Technologies, Moore Threads ve Enflame Technology<\/b><\/p>\n

Cambricon Technologies<\/strong>, yapay zeka modellerini e\u011fitmek ve bu modelleri cihazlar \u00fczerinde \u00e7al\u0131\u015ft\u0131rmak i\u00e7in farkl\u0131 t\u00fcrlerde yar\u0131 iletkenler tasarl\u0131yor. \u015eirketin \u00fcr\u00fcnleri, veri merkezlerinden cihazlara kadar geni\u015f bir alanda kullan\u0131l\u0131yor. Ancak, Cambricon son y\u0131llarda finansal zorluklar ya\u015fam\u0131\u015f ve 2022 y\u0131l\u0131nda i\u015ften \u00e7\u0131karmalar yapm\u0131\u015ft\u0131. ABD\u2019nin kara listesinde yer ald\u0131\u011f\u0131 i\u00e7in Amerikan teknolojisine eri\u015fim konusunda s\u0131n\u0131rlamalarla kar\u015f\u0131 kar\u015f\u0131ya.<\/p>\n

<\/span> Moore Threads<\/strong>, 2020 y\u0131l\u0131nda kurulmu\u015f gen\u00e7 bir \u015firket olmas\u0131na ra\u011fmen, b\u00fcy\u00fck yapay zeka modellerini e\u011fitmek i\u00e7in kullan\u0131lan GPU\u2019laer geli\u015ftiriyor. \u015eirketin veri merkezi \u00fcr\u00fcn\u00fc olan MTT KUAE<\/strong>, b\u00fcy\u00fck \u00f6l\u00e7ekli yapay zeka e\u011fitimlerinde kullan\u0131lmak \u00fczere tasarland\u0131. Moore Threads, k\u00fcresel bir GPU lideri olmay\u0131 hedefliyor ve bu ama\u00e7 do\u011frultusunda TikTok\u2019un sahibi ByteDance gibi b\u00fcy\u00fck yat\u0131r\u0131mc\u0131lardan destek al\u0131yor. Ayr\u0131ca Sequoia ve GGV Capital gibi risk sermayesi firmalar\u0131 da \u015firketin yat\u0131r\u0131mc\u0131lar\u0131 aras\u0131nda yer al\u0131yor. <\/figure>\n

\u00a0<\/p>\n

<\/span> Enflame Technology<\/strong> ise \u00c7in’de kendini Nvidia’ya kar\u015f\u0131 yerli bir alternatif olarak konumland\u0131rmaya \u00e7al\u0131\u015fan bir ba\u015fka start-up. \u015eirket, yapay zeka e\u011fitimi ve s\u00fcre\u00e7lerine odaklanan veri merkezleri i\u00e7in \u00e7ipler tasarl\u0131yor. \u00c7in’in en b\u00fcy\u00fck teknoloji \u015firketlerinden biri olan Tencent, Enflame’in yat\u0131r\u0131mc\u0131lar\u0131ndan biri. <\/figure>\n

Sonu\u00e7 olarak \u00c7in\u2019in elinde belki Nvidia ile ba\u015fa ba\u015f rekabet edebilecek \u015fimdilik bir \u015firket yok ancak \u00fclkenin elinde \u00e7ok zengin se\u00e7enekler ve yetenekler bulunuyor. Huawei\u2019nin \u00fcr\u00fcnleri Nvidia\u2019n\u0131n \u00fcr\u00fcnlerinden zay\u0131f olsa da rahatl\u0131kla \u00c7in\u2019de al\u0131c\u0131 bulabiliyor. Zira dedi\u011fimiz gibi; Nvidia, k\u0131s\u0131tlamalar nedeniyle \u00c7in\u2019e \u00fcr\u00fcnlerinin hepsini satam\u0131yor ve genellikle k\u0131rp\u0131lm\u0131\u015f \u00fcr\u00fcnleri piyasaya sokmas\u0131na izin veriliyor. Bu durum \u00c7inli firmalar\u0131 yerli \u00fcreticilere y\u00f6nlendiriyor. Yapt\u0131r\u0131mlar s\u0131k\u0131la\u015fm\u0131\u015f olsa da \u00c7in, devlet deste\u011fini firmalar\u0131n\u0131n \u00fczerinden esirgemiyor. Bu da firmalar\u0131n Ar-Ge harcamalar\u0131n\u0131 s\u00fcrd\u00fcrmesini sa\u011fl\u0131yor. Dolay\u0131s\u0131yla \u00c7in\u2019in yerli Nvidia\u2019s\u0131 i\u00e7in zemin haz\u0131r ve sadece zaman gerekiyor.<\/p>\n\n

<\/span><\/p>\n

Kaynak :\u00a0https:\/\/www.donanimhaber.com\/cin-nvidia-ya-yerli-bir-alternatif-ariyor-hangi-secenekler-var–182045<\/span><\/p>\n","protected":false},"excerpt":{"rendered":"

ABD\u2019nin y\u0131llar boyunca \u00c7in\u2019in yar\u0131 iletken sekt\u00f6r\u00fcne uygulad\u0131\u011f\u0131 yapt\u0131r\u0131mlar, Pekin\u2019i yerli \u00e7ip \u00fcretimini h\u0131zland\u0131rmaya zorlad\u0131. Yapay zeka ve temel modellerin patlamas\u0131, \u00c7in’in \u00e7ip end\u00fcstrisinde lider bir rol oynama hedefini daha da h\u0131zland\u0131rd\u0131 …<\/p>\n","protected":false},"author":1,"featured_media":79401,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[8],"tags":[498,267,284,259,421],"class_list":["post-79400","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-teknoloji","tag-cip","tag-huawei","tag-nvidia","tag-sirket","tag-yapay-zeka"],"_links":{"self":[{"href":"https:\/\/ekamu.net\/index.php\/wp-json\/wp\/v2\/posts\/79400","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/ekamu.net\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/ekamu.net\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/ekamu.net\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/ekamu.net\/index.php\/wp-json\/wp\/v2\/comments?post=79400"}],"version-history":[{"count":1,"href":"https:\/\/ekamu.net\/index.php\/wp-json\/wp\/v2\/posts\/79400\/revisions"}],"predecessor-version":[{"id":79408,"href":"https:\/\/ekamu.net\/index.php\/wp-json\/wp\/v2\/posts\/79400\/revisions\/79408"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/ekamu.net\/index.php\/wp-json\/wp\/v2\/media\/79401"}],"wp:attachment":[{"href":"https:\/\/ekamu.net\/index.php\/wp-json\/wp\/v2\/media?parent=79400"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/ekamu.net\/index.php\/wp-json\/wp\/v2\/categories?post=79400"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/ekamu.net\/index.php\/wp-json\/wp\/v2\/tags?post=79400"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}