- Ubuchwepheshe bamanje benqubo ye-14nm buzohlala busebenza kuze kube okungenani u-2021
- Izethulo ze-Intel zokushintshela kubuchwepheshe obusha zisho noma yimaphi amaphrosesa nemikhiqizo, kodwa hhayi eyedeskithophu.
- Ukukhiqizwa okuningi kwemikhiqizo ye-Intel kusetshenziswa ubuchwepheshe be-7nm kuzokwethulwa ngaphambi kuka-2022
- Zonke izinsiza zobunjiniyela zizodluliswa zisuka kubuchwepheshe benqubo ye-14 nm ziye ku-7 nm, kanti abanye ochwepheshe bazobandakanyeka kubuchwepheshe benqubo ye-10 nm.
Ukuvuza kumephu yomgwaqo kaDell thola umbono mayelana nezinhlelo ze-Intel zokukhulula amaphrosesa amasha, futhi imikhiqizo ye-14-nm kufanele ivele engxenyeni yedeskithophu isikhathi eside kakhulu, uma uthembele kulo mthombo wolwazi. Kodwa-ke, umcimbi we-Intel wabatshalizimali kuleli sonto ungasikhanyisela ngalesi simo ngokukhululwa kwemikhiqizo eyi-10-nm kanye ne-7-nm, futhi konke bekungaba kuhle uma kungenjalo ngokuthula okudangele kwabamele inkampani mayelana nesikhathi sokukhishwa kwedeskithophu entsha. amaprosesa.
Uhlelo lwangempela I-Intel bekumele yenze izinguquko ukuze ikwazi ukusebenzisa ubuchwepheshe be-10nm
Akuyona imfihlo ukuthi eminyakeni eyisithupha edlule i-Intel yayizethemba emandleni ayo okwenza kahle ukukhiqizwa kwe-serial kwama-processor angu-10nm ngo-2016. Njengoba abaphathi be-Intel, abakwazile ukushintsha ngalesi sikhathi, bachaze izikhathi ezingaphezu kwesisodwa, okuhlosiwe okunamandla kakhulu kwakhethwa ukukala kwejiyomethri kwama-transistors lapho kuhlelwa ukushintshela kubuchwepheshe benqubo ye-10-nm, futhi akuzange kwenzeke kahle ukukhiqiza. yemikhiqizo engu-10-nm phakathi nesikhathi esishiwo.
Ngonyaka odlule, ukulethwa kwamaphrosesa eselula e-10nm Cannon Lake kwaqala, kodwa ayefaneleka kuphela ukusetshenziswa kumadivayisi eselula amancane kakhulu, ayengenawo ama-cores angaphezu kwamabili, futhi uhlelo olungaphansi lwezithombe ze-on-chip kwadingeka luvalwe ngokuphelele. Empeleni, amavolumu okuphakelwa kweCannon Lake abengabalulekile, ngakho-ke i-Intel manje ikhombisa u-10 njengokuqala kwesikhathi sokuthuthukiswa kwenqubo ye-2019nm. Amaphrosesa e-10nm Ice Lake eselula azokwethulwa ngoJuni walo nyaka, ngaleso sikhathi ukulethwa kwawo kubakhiqizi bamalaptop azokuqala, futhi azokhipha amakhompyutha aphelile ngokusekelwe kuwo engxenyeni yesibili yonyaka.
Kuphela ngokwenguqulo esemthethweni, ubuchwepheshe be-Intel be-14-nm budlule ezizukulwaneni ezintathu ekuthuthukisweni kwayo kokuziphendukela kwemvelo, futhi kube nokuthuthuka okuncane nakakhulu. I-Intel iyaziqhenya ukusho ukuthi ukusebenza nge-watt ngayinye kuthuthuke ngo-14% kusukela esizukulwaneni sokuqala kuya kunqubo yesizukulwane sesithathu ye-20nm.
Ngaphezu kwalokho, uma ubheka izethulo zakamuva ze-Intel ezivela kumcimbi wotshalizimali kaMeyi, uzothola ukuthi umjikelezo wempilo wobuchwepheshe benqubo ye-14 nm unwetshiwe kuze kube ngu-2021 uhlanganisa. Ngaleso sikhathi, ukukhiqizwa kwe-serial kwemikhiqizo yokuqala ye-7nm kuzobe sekuqalile, futhi ubuchwepheshe benqubo ye-14nm buzoqhubeka bufanele uhla oluthile lwemikhiqizo ye-Intel.
Akuzange kukhulunywe ngokudlulisa amaphrosesa edeskithophu kubuchwepheshe be-7nm
Ngisho nokuvuza mayelana nezinhlelo ze-Intel esethulweni sikaDell akuzange kube nolwazi mayelana nesikhathi sokukhishwa kwamaphrosesa angu-10nm ukuze asetshenziswe kwideskithophu. Kulo mongo, amaphrosesa eselula asebenzisa amandla aphansi kakhulu, inani lawo lama-cores alidluli kwezine, avele kakhulu. Ngisho nakulokhu, ngeke zisetshenziswe kabanzi kuze kube ngu-2021. Ngaleso sikhathi, i-10nm Tiger Lake processors izobe isikhishiwe, ezohlinzeka ngokusekela i-PCI Express 4.0 futhi izokhiqizwa kusetshenziswa isizukulwane sesibili sobuchwepheshe be-10nm. Amaphrosesa e-Tiger Lake azophinde athole ihluzo emisha enama-cores okubulala angama-96, esebenzisa ukwakheka okujwayelekile okunemikhiqizo ehlukile eyamenyezelwa ngo-2020.
Ekupheleni kuka-2019, amaphrosesa angu-10nm eLakefield anesakhiwo esiyinkimbinkimbi se-Foveros azokhululwa, okusho ukuhlanganiswa kwakho kokubili kokucabanga kwesistimu ne-RAM kuphakheji elilodwa. Ngisho neprosesa yehluzo yokuqala ye-Intel “okuthiwa ideskithophu” eminyakeni engamashumi amabili edlule izokhishwa ngo-2020 kusetshenziswa ubuchwepheshe be-10nm, kodwa amaphrosesa edeskithophu esimweni sokushintshela kubuchwepheshe be-10nm awazange ashiwo nhlobo emcimbini wabatshalizimali.
Kukhona nesiqiniseko esanele esigabeni seseva. Ngaphambi kokuthi kukhishwe ama-processor we-10nm Ice Lake-SP engxenyeni yokuqala yonyaka ozayo, kuzokhululwa ama-14nm Cooper Lake processors ahambisana nawo ngokwesakhiwo. Abamele i-Intel abacacisi ukuthi yibuphi ubuchwepheshe obuzosetshenziswa ukukhiqiza abazolandela i-Ice Lake-SP ngendlela yeSapphire Rapids, kodwa uNavin Shenoy uvumile ngesikhathi seseshini yemibuzo nezimpendulo nabahlaziyi ukuthi umkhiqizo wesibili ukhiqize kusetshenziswa ubuchwepheshe be-7nm ngemuva kwe-GPU yama-accelerator. ikhompuyutha kuzoba iyunithi yokucubungula emaphakathi yamaseva. Uma kucatshangelwa ukuthi izibulo le-7nm lizokhishwa ngo-2021, khona-ke zombili izikhathi zika-7 nezakamuva zifaneleka ngokulinganayo ukukhishwa kwephrosesa ephakathi kwe-2021nm server-class. I-Sapphire Rapids kulindeleke ukuthi iqale ukuvela ngo-2021, kanti umlandeli wayo uzofika ngo-2022.
Ngakho-ke, lapho ichaza izinhlelo zayo zamanje zokufuduka kubuchwepheshe benqubo ye-7nm, i-Intel ikhuluma ngokucacile ngama-GPU nama-CPU wezinhlelo zokusebenza zeseva, kodwa ishiya ideskithophu neselula ngaphandle kwesithombe.
Ukuhlaselwa kobuchwepheshe be-7nm: ithemba elikhohlisayo lemikhiqizo yedeskithophu
I-Intel CEO uRobert Swan wenze izitatimende ezimbalwa ezibalulekile mayelana nokuthuthukiswa kobuchwepheshe benqubo ye-7nm. Okokuqala, uthe ngemuva kuka-2021 lolu hlelo luzovumela inkampani ukuthi yehlise izindleko zokusebenza. Lokhu ukuzethemba kusekelwe eqinisweni lokuthi inkampani manje kufanele ithuthukise izinqubo ezintathu zobuchwepheshe ngokuhambisana: 14 nm, 10 nm no-7 nm. Ukuzama ukuhambisana nenqubo ye-10nm kwandisa izindleko, futhi uma inqubo ye-7nm isiqalile futhi isebenza, inkampani inethemba lokubuyisela izindleko ngaphansi kohlelo lwayo oluyinhloko iminyaka eminingana.
Okwesibili, uSwan uthe bonke abasebenzi bonjiniyela ababebambe iqhaza ekwakhiweni kwemikhiqizo ye-Intel ye-7nm bazosatshalaliswa ukuthuthukisa ubuchwepheshe be-14nm. Phakathi kwalokhu kokugcina, sazi amaphrosesa amaningi edeskithophu anenani elikhulu lama-cores kanye nezinga eliphezulu lokusebenza. Ingabe lokhu kusho ukuthi leli thimba lochwepheshe lizophumelela ekudaleni amaphrosesa we-desktop 7nm? Impendulo yalo mbuzo cishe kuyodingeka ifunwe ngale kweshumi leminyaka lamanje.
Okwesithathu, inhloko ye-Intel yachaza ukuthi ukukhiqizwa ngobuningi kwemikhiqizo ye-Intel esebenzisa ubuchwepheshe be-7-nm kuzokwethulwa ngo-2022 kuphela, ngemuva kokuvela kwephrosesa yokuqala yehluzo ekhishwe ngonyaka owedlule kusetshenziswa ubuchwepheshe be-7-nm kusetshenziswa i-ultra-hard ultraviolet lithography. . Ukuthi lezi kuzoba yideskithophu noma amaphrosesa eselula manje kunzima ukusho ngokuqiniseka, ngoba ngisho nasekulandeleni kokudlulisa imikhiqizo ezinqubweni ezintsha zobuchwepheshe, izinto ezibalulekile ze-Intel zishintshile.
Source: 3dnews.ru