Kuyilo lwamaqonga e-x86, kuye kwavela iindlela ezimbini ezincedisanayo. Ngolunye uhlobo, kufuneka siqhubele phambili ekudibaniseni izixhobo zekhompyuter kunye nolawulo kwi-chip enye. Indlela yesibini ikhuthaza ukuhanjiswa koxanduva: iprosesa ixhotyiswe ngebhasi esebenza kakhulu eyenza i-peripheral scalable ecosystem. Yenza isiseko se-Intel C620 yenkqubo ye-logic topology kumaqonga aphezulu.
Umahluko osisiseko kwi-chipset ye-Intel C610 yangaphambili kukwandiswa kwejelo lonxibelelwano phakathi kweprosesa kunye neeperipherals ezibandakanyiweyo kwi-PCH chip ngokusebenzisa amakhonkco e-PCIe kunye nebhasi yeDMI yemveli.
Makhe sijonge ngakumbi kutshintsho lwebhulorho esemazantsi ye-Intel Lewisburg: zeziphi iindlela zendaleko kunye nenguquko ezithe zandisa amandla ayo ekunxibelelaneni nabalungiseleli?
Utshintsho lwe-Evolution kunxibelelwano lwe-CPU-PCH
Njengenxalenye yendlela yokuziphendukela kwemvelo, umjelo oyintloko wonxibelelwano phakathi kwe-CPU kunye nebhulorho esezantsi, eyi-DMI (Direct Media Interface) ibhasi, ifumene inkxaso ye-PCIe x4 Gen3 mode kunye nokusebenza kwe-8.0 GT / S. Ngaphambili, kwi-Intel C610 PCH, unxibelelwano phakathi kweprosesa kunye nengqiqo yenkqubo yenziwa kwimodi ye-PCIe x4 Gen 2 kwi-5.0 GT / S bandwidth.
Ukuthelekiswa kwenkqubo yokusebenza kwengqondo ye-Intel C610 kunye neC620
Qaphela ukuba le subsystem igcinwe kakhulu kuneechwephesha zePCIe ezakhelwe ngaphakathi zeprosesa, edla ngokusetyenziselwa ukuqhagamshela ii-GPUs kunye ne-NVMe drives, apho i-PCIe 3.0 isetyenziswe ixesha elide kwaye ukutshintshela kwi-PCI Express Gen4 kucwangcisiwe.
Utshintsho olutshintsho kunxibelelwano lwe-CPU-PCH
Utshintsho oluguqukayo lubandakanya ukongezwa kweendlela ezintsha zonxibelelwano zePCIe CPU-PCH, ezibizwa ngokuba zii-Uplinks ezongezelelweyo. Ngokwasemzimbeni, la ngamazibuko amabini e-PCI Express asebenza kwiimowudi ze-PCIe x8 Gen3 kunye ne-PCIe x16 Gen3, zombini i-8.0 GT/S.
Ukusebenzisana phakathi kwe-CPU kunye ne-Intel C620 PCH, iibhasi ezi-3 zisetyenziswa: i-DMI kunye nezibuko ezimbini ze-PCI Express
Kutheni bekuyimfuneko ukuhlaziya i-topology ekhoyo yonxibelelwano kunye ne-Intel C620? Okokuqala, ukuya kuthi ga kwi-4x 10GbE abalawuli benethiwekhi kunye nokusebenza kwe-RDMA banokudibaniswa kwi-PCH. Okwesibini, isizukulwana esitsha kunye nesikhawulezayo se-Intel QuickAssist Technology (QAT) coprocessors, ebonelela ngenkxaso ye-hardware yoxinzelelo kunye ne-encryption, inoxanduva lokubethela i-traffic network kunye notshintshiselwano kunye ne-subsystem yokugcina. Kwaye ekugqibeleni, "injini entsha" - , eya kufumaneka kuphela kwii-OEMs.
Scalability kunye bhetyebhetye
Ipropati ebalulekileyo yikhono lokukhetha ukukhetha kuphela i-topology yoxhulumaniso lwe-PCH, kodwa kunye nezinto eziphambili zemithombo yangaphakathi ye-chip ekufikeleleni kwiitshaneli zonxibelelwano olukhawulezayo kunye neprosesa ephakathi (abaqhubeki). Ukongezelela, kwi-EPO ekhethekileyo (i-EndPoint Mode kuphela), uxhulumaniso lwe-PCH lwenziwa kwimeko yesixhobo esiqhelekileyo se-PCI Express esiqulethe izixhobo ze-10 GbE kunye ne-Intel QAT. Ngexesha elifanayo, i-interface ye-DMI yakudala, kunye nenani le-subsystems yeLifa, eboniswe kumnyama kumzobo, zikhutshaziwe.
Uyilo lwangaphakathi lwe-Intel C620 PCH chip
Kwithiyori, oku kwenza kube nokwenzeka ukusebenzisa ngaphezu kwe-Intel C620 PCH chip kwinkqubo, ukulinganisa i-10 GbE kunye ne-Intel QAT yokusebenza ukuhlangabezana neemfuno zokusebenza. Kwangaxeshanye, imisebenzi yeLifa efunekayo kuphela kwikopi enye inokuvulwa kuphela kwenye yeetshiphusi zePCH ezifakiweyo.
Ke, intetho yokugqibela kuyilo iya kuba yeyomphuhlisi weqonga, esebenza ngesiseko sezinto zombini zobuchwepheshe kunye nokuthengisa ngokuhambelana nokumiswa kwemveliso nganye ethile.
umthombo: www.habr.com