Ukufika kwamaprosesa amasha e-Intel Core Ultra 200S Plus, aziwa ngokuthi i-Arrow Lake Refresh, kuletha okungaphezu nje kwe-GHz nama-cores. Kuza nesofthiwe ekhethekile kakhulu: i Ithuluzi Lokwenza Kahle Kakhulu le-Intel, noma i-BOT/IBOT, ubuchwepheshe obuklanyelwe ukuthuthukisa ukusebenza kwangempela, ikakhulukazi ku imidlalo enama-GPU aphansi kanye nemithwalo yemisebenzi enzima, ngaphandle kokuthi abathuthukisi bathinte umugqa owodwa wekhodi.
Le ndlela iyathakazelisa ngoba igxile endleleni ama-binary asetshenziswa ngayo ku-CPU, hhayi nje amandla aluhlaza e-chip. Esikhundleni sokuphinda uhlanganise izinhlelo noma imidlalo yokulungisa, i-Intel iphakamisa ungqimba lwe- ukwenza ngcono okunamandla kwekhodi ehlanganisiwe kakade Lobu buchwepheshe busebenza phakathi kwe-executable kanye ne-processor, buhlela kabusha imiyalelo futhi busebenzisa kangcono i-microarchitecture yangaphakathi. Konke lokhu kuvula impikiswano ethakazelisayo kakhulu: ingabe "kusese" "kulungile" ukuqhathanisa ama-processor uma omunye wabo esebenzisa ukwesekwa okunamandla kangaka?
Iyini ngempela i-Intel Binary Optimization Tool futhi ihluke kanjani kwezinye izindlela zokwenza ngcono?
Ithuluzi Lokwenza Kahle Kakhulu le-Intel, empeleni, ungqimba lokuhumusha okuhlakaniphile kanye nokwenza ngcono okusetshenziswe kuma-binary aqoqwe kakadeAyihlanganisi kabusha, ayihlanganisi, noma ayishintshi i-executable yokuqala yomdlalo noma uhlelo lokusebenza, kodwa ishintsha indlela i-binary ephakela ngayo i-CPU ukuze isebenze kahle kakhulu.
Lo mbono uvela enkingeni eyisisekelo ebikhona iminyaka eminingi: imidlalo nezinhlelo eziningi zakhiwa kucatshangelwa okulandelayo: izakhiwo ezindala, ama-consoles, noma ama-CPU ajwayelekileUmphumela uba ukuthi, uma uzisebenzisa kuhadiwe yesimanje njenge-Arrow Lake Refresh, i-CPU ayisetshenziswa ngokugcwele. Kukhona ukungasebenzi kahle, ukwehluleka kokubikezela kwegatsha, ukusetshenziswa kabi kwe-cache, noma ukuntuleka kwe-vectorization lapho kungenzeka khona.
Esikhundleni sokulungiswa kwe-compiler yendabuko noma ama-patches onjiniyela, i-BOT iphakamisa ukuthi i-Intel, kuma-laboratories ayo, Hlaziya leyo mithwalo yemisebenzi ezingeni elincane lokwakha futhi ukhiqize inguqulo elungiselelwe kahle yekhodi, kodwa ngaphandle kokuthinta ifayela le-.exe onalo kudiski.
Kuhlelo lwesofthiwe lomkhiqizo, i-BOT ijoyina amathuluzi afana ne-Intel Application Optimizer (APO) nezinye izingxenye ze-performance suite. Ngenkathi i-APO igxile kakhulu ku- ukwabiwa kwengqikithi kanye nentambo kanye nokusebenzisana nomhleli wesheduli Kusukela ohlelweni lokusebenza, i-BOT isebenza ngisho nangokwengeziwe, ekugelezeni kwemiyalelo esetshenziswa ku-CPU uqobo, okwenza zombili izinhlelo zihambisane esikhundleni sokudlulana.
Indlela i-Intel BOT esebenza ngayo ngaphakathi: i-HWPGO, i-microarchitecture, kanye namaphrofayili e-binary
Injini yobuchwepheshe ngemuva kwe-Intel BOT incike endleleni Ukuthuthukiswa Okuqondiswa Yiphrofayili Okusekelwe Kuhadiwe (HWPGO)Kalula nje, i-Intel ihlaziya indlela ama-binary aziphatha ngayo lapho esebenza ekwakhiweni kwayo, ithole izithiyo, futhi kusukela kulolo lwazi, ikhiqiza inguqulo elungiselelwe kahle yekhodi yomshini.
Phakathi nalokhu kuhlaziywa, izinto ezifana nokwehluleka kokubikezela kwegatsha, amabhamuza epayipi, ukubambezeleka kwe-cache, kanye nokusetshenziswa kabi kwe-prefetcher kuyaqashelwa ngokuningiliziwe. Uma kutholakala amaphethini angasebenzi kahle, uhlelo lwakha i- iphrofayili yokulungisa ehlela kabusha imiyalelo ukunciphisa lezo zinkinga. Umgomo akusikho ukwenza umsebenzi omncane noma "ukweqa" imisebenzi, kodwa ukwenza inani elifanayo lomsebenzi ngendlela evumela i-CPU ukuthi igcine i-IPC esebenza kahle kakhulu.
Isici esibalulekile ukuthi yonke le nqubo yokwenza iphrofayili ayenzeki kwi-PC yakho, kodwa iyenzeka kumalebhu e-Intel. Isebenzisa amasu okwenza ngcono i-post-link, inkampani ikhiqiza amakhodi e-binary ahlelwe kabusha ngobuningi bemiyalelo obuthuthukisiweLawa maphrofayili aklanyelwe ngqo ukwakheka okuncane kwama-chip awo akamuva. Bese asakazwa njengengxenye yephakheji yokusebenza ukuze umsebenzisi ayisebenzise.
Emshinini wakho, uma usebenzisa i-Intel BOT, isevisi yemodi yomsebenzisi isebenza ngemuva. Le sevisi inesibopho sokwenza lokhu qapha lapho kukhishwa khona ama-binary ahambisanayo futhi uqondise kabusha ukusebenza kwayo ezindleleni ezilungiselelwe ezidalwe yi-Intel. Ifayela le-.exe ku-hard drive yakho alishintshiwe: yiziphi izinguquko eziyindlela imiyalelo eyilandelayo ngesikhathi sokusebenza, efana nendlela umshayeli wehluzo ashintsha ngayo ama-shaders ngezinguqulo ezilungiselelwe emidlalweni ethile.
Ngobuchwepheshe, ithuluzi lisebenza njengohlobo lwe-dynamic execution flow optimizer, lisebenzisa ulwazi lwalo lwangaphakathi lwe-Arrow Lake Refresh microarchitecture. Alisebenzi njengomshayeli wakudala noma i-patch yomdlalo; kuyinto ephakathi nendawo. hlela kabusha imenyu yemiyalelo ngaphandle kokushintsha "izingcezu", kuphela uhlelo lapho zikhonzwa khona kwi-CPU.
Ubudlelwano phakathi kwe-Intel BOT, i-APO, namanye amathuluzi okusebenza kwe-Intel
Esulweni lakamuva le-Intel, i-BOT ayizi yodwa: iyingxenye yephakheji ebanzi lapho kuhlangana khona obunye ubuchwepheshe. I-Intel Extreme Tuning Utility (XTU)I-Intel Application Optimizer (APO) kanye ne-Intel's Binary Optimization Tool (IBOT/BOT). Ingxenye ngayinye ihlanganisa isici esihlukile sokusebenza.
I-XTU igxile ohlangothini olujwayelekile kakhulu lwe Ukweqiwa kwe-CPU, ama-voltage kanye namapharamithaOkungukuthi, ukulungiswa kwehadiwe uqobo. I-APO isebenza kusofthiwe kanye nesendlalelo sesistimu yokusebenza, ilawula indlela imicu nemisebenzi esakazwa ngayo phakathi kwezinhlobo ezahlukene zama-core (ama-P-core, ama-E-core) futhi iqaphe ukwabiwa kwezinsiza ukuze izinhlelo zokusebenza eziyidingayo zizuze kakhulu kwihadiwe efanele.
I-BOT, yona, iya ezingeni eliphansi nakakhulu: ukugeleza kwemiyalelo okwenziwa yi-CPU kwe-binary ethileNakuba i-APO izama ukwenza umsebenzi uwele ku-kernel efanele ngesikhathi esifanele, i-BOT iyaqinisekisa ukuthi leyo miyalelo iyahlelwa futhi ihlelwe ngendlela efanelana kahle nokwakheka kwangaphakathi kwe-chip.
Empeleni, lokhu kusho ukuthi uma umdlalo usohlwini lwezihloko ezisekelwayo futhi umsebenzisi esebenzisa imodi ehambisanayo, I-APO ne-BOT bangasebenzisanaI-APO iphatha ukwabiwa kahle kwemithwalo yemisebenzi, futhi i-BOT ikhipha ikhodi yomshini ephumayo. Kukulezi zimo lapho izinzuzo zokusebenza eziphawuleka kakhulu zibhalwe khona, ikakhulukazi emidlalweni engavumelani kahle nokwakhiwa kwe-Intel okuhlanganisiwe.
Kuyafaneleka ukuqaphela ukuthi, nakuba ifilosofi ye-BOT ikhumbuza eye-APO (zombili ziyizinhlelo zokuthuthukisa isofthiwe ezisetshenziswa "ngaphandle" kohlelo lokusebenza), Azikwazi ukushintshaniswa noma ukulinganaI-Intel iziveza njengamathuluzi ahambisanayo okulungisa ungqimba ngalunye lwesitaki sokusebenza: kusukela kuhadiwe kanye nokucushwa kwayo (XTU), ngokusebenzisa ukwabiwa kwenqubo (APO), kuya ekuhleleni kabusha ikhodi yomshini uqobo (BOT).
Izinzuzo zokusebenza kwemidlalo: kusukela ekuthuthukisweni okuncane kuya ekukhuphukeni okumangalisayo
I-Intel ithi, eqenjini layo lokuqala lezihloko ezisekelwayo, i-Binary Optimization Tool ifinyelela i- ukuthuthuka okumaphakathi cishe okungu-8% emidlalweningeziqongo eziphakeme kakhulu ezimweni ezithile. Okwamanje, sikhuluma ngohlu oluncishisiwe lwemidlalo ecishe ibe yi-12, kodwa imiphumela yokuqala iyamangaza.
Esinye sezibonelo ezicashunwe kakhulu yilezi Umthunzi weTomb RaiderLesi sihloko, ngenxa yokwenza ngcono kwaso kwasekuqaleni, asizange sisebenzise ngokugcwele ukwakheka okuhlanganisiwe kwamaphrosesa e-Intel anamuhla. Ngenhlanganisela ye-APO + BOT, ukwanda okungaba ngu-22% nangaphezulu kuye kwalinganiswa kwezinye izilinganiso, ngezimo lapho ukweqa kwe-FPS, isibonelo, kusuka kuzimele ezingu-298 kuya ku-375 ngomzuzwana, okumele ukweqa okucishe kube ngu-26%.
Kumdlalo wesimanje, olungiselelwe kangcono ama-CPU amanje, njenge Cyberpunk 2077Kodwa-ke, indaba ishintsha kakhulu. Kulokhu, ukuthuthuka kungamaphesenti ambalwa kuphela: kusukela cishe kuma-FPS angu-210 kuya ngaphezulu kwama-220 kuma-benchmark athile, noma kusukela cishe kuma-FPS angu-173 kuya cishe kuma-179 kwezinye izigcawu ezithile. Sikhuluma ngakho ukwanda cishe ngo-3-5%ezisamukelekile ngoba "zimahhala" kumsebenzisi, kodwa azishintshi kangako ulwazi.
Lokhu okuboniswa yilezi zibalo ukuthi i-BOT ayiyona imilingo noma umlingo omnyama, kodwa kunalokho Kuncike kakhulu ekutheni umdlalo wokuqala uhlelwe kahle kangakanani (noma kahle)Uma i-binary yokuqala iphambaniswe kabi kakhulu nokwakheka kwe-Intel, umthelela ungaba mkhulu kakhulu. Uma ilungiswe kahle, i-BOT ikhipha kuphela ama-FPS ambalwa engeziwe, angenza umehluko ezimweni ezibucayi noma ngamanani aphezulu okuvuselela.
Ngaphandle kwezwe lemidlalo, imiphumela ephawulekayo ibonwe nakwezinye izinhlelo zokusebenza ezisebenzisa izinsiza eziningi njenge Ukususwa Kwezinto noma ukucubungula i-HDRlapho ukuhlaziywa kwelabhorethri kubonisa ukwanda okungafinyelela ku-30% ngenxa yokwenziwa kwe-vector okunamandla kakhulu kwezingxenye zekhodi ye-scalar ekuqaleni.
Ukuhlaziywa kwe-vectorization ejulile kanye nemiyalelo: icala le-Geekbench
Enye yezindlela ezicacile zokuqonda lokho okwenziwa yi-BOT ukubheka lokho i-Primate Labs, ebhekene nakho GeekbenchBahlole ngokucophelela ukuthi ukusebenza kwabo kokulinganisa kwashintsha kanjani lapho ithuluzi le-Intel lisebenza. Ukuze benze lokhu, basebenzise i-Intel Software Development Emulator (SDE), elinganisa inani nohlobo lwemiyalelo esetshenzisiwe.
Ekugijimeni okujwayelekile kwe-Geekbench 6 ngaphandle kwe-bot, ukuhlolwa kuthathe cishe Imiyalelo eyi-1,26 trillion ukuqedela. Njengoba i-BOT ivuliwe, leso sibalo sehle saya cishe ku-1,08 trillion, okusho ukuthi cishe kuzoncishiswa ngo-14% enanini lemiyalelo iyonke. Ngamanye amazwi, umsebenzi wenziwa ngobuncane nangendlela ephumelelayo, ngaphandle kokusika izici noma ukusebenzisa izinqamuleli.
Uma ihlukaniswa ngohlobo lwemiyalelo, ifilosofi yaleli thuluzi iba sobala nakakhulu. Inani lemiyalelo ye-scalar liyehla kusukela nxazonke 220.000 billion kuya cishe ku-84.600 billion, kuyilapho imiyalelo ye-vector (SSE2, AVX2, njll.) ikhuphuka isuka ku-1.250 billion yaya cishe ku-18.300 billion, okungukuthi, ukwanda cishe izikhathi ezingu-13,7 kulolu hlobo lwemiyalelo.
Lokhu kucacisa kahle ukuthi i-BOT izinikele kakhulu ku- Guqula izingxenye zekhodi ye-scalar ezingasebenzi kahle zibe ikhodi ye-vectorized okusebenzisa kangcono amayunithi e-SIMD kuma-processor e-Intel. Lapho kwenziwa khona imisebenzi eminingi elula nephindaphindayo ngaphambilini, manje iqoqwe ngemisebenzi ye-vector ecubungula amaphuzu amaningi edatha ngesikhathi esifanayo, into ehambisana kahle nomklamo wangaphakathi wezakhiwo ezincane zakamuva zenkampani.
Lokhu kuhlela okukhulu akwenziwa ngokungaboni. Kusekelwe ekuhleleni i-hardware (HWPGO) kanye nokwenza ngcono kwe-binary-level post-optimization okwenziwa yi-Intel kuma-lab ayo, okuchaza ukuthi kungani, ngaphandle, i-BOT ibhekwa njenge- ibhokisi elimnyama eliyinkimbinkimbi kakhuluUmsebenzisi ubona kuphela ukuthi i-benchmark noma umdlalo usebenza ngokushesha, kodwa akanalo ulwazi oluqondile lokuthi yiziphi izinguquko ezisetshenziswe endleleni yokusebenza.
Iyahambisana nemidlalo embalwa, i-Arrow Lake Refresh kuphela, futhi idinga ukwenziwa isebenze ngesandla.
Naphezu kwamandla ayo, ithuluzi liza nemikhawulo eminingana ephawulekayo. Okokuqala ukuthi, okungenani kulesi sigaba sokuqala, Ukuhambisana kwe-Intel BOT kukhawulelwe ohlwini olufushane lwemidlalocishe izihloko ezikhethiwe eziyishumi nambili. I-Intel ibonise ukuthi izokwandisa ikhathalogi, kodwa okwamanje ububanzi bulinganiselwe kakhulu.
Umkhawulo wesibili obalulekile ukuthi isici esikhethekile samaprosesa e-Arrow Lake Refresh (i-Core Ultra 200S Plus), ngokusekelwa kwe-Linux kuncike ezixazululweni ezifana I-Proton 11Akuyona into ongayisebenzisa ezizukulwaneni ezidlule, okwenza i-BOT ibe yisici esihlukanisayo salo mndeni we-CPU uma kuqhathaniswa namamodeli adlule, futhi ngokwengxenye, uma kuqhathaniswa nomncintiswano oqondile.
Ngaphezu kwalokho, ukuze ajabulele lokhu kuthuthukiswa, umsebenzisi kufanele enze izinyathelo ezithile: okwamanje, i-BOT iyasebenza nge- "Imodi Ethuthukisiwe" ngaphakathi kwephakheji yokusebenza kwe-Intel futhi idinga ukuqala kabusha uhlelo ukuze amaphrofayili asetshenziswe kahle. Akuyona inqubo eyinkimbinkimbi, kodwa ayibonakali nhlobo.
I-Intel igcizelele ukuthi ifuna ukuthi ulwazi lokugcina lube ngokuzenzakalelayo ngangokunokwenzeka, futhi empeleni iningi lomlingo lenzeka buthule uma usulilungisile, kodwa okwamanje kukhona ingxenye ethile ye ubunzima kanye nokukhethekile okunciphisa umthelela wayo omkhulu phakathi kwabasebenzisi abangenantshiseko noma labo abangazimisele kangako ukuthinta izilungiselelo ezithuthukisiwe.
Elinye iphuzu okufanele licatshangelwe ukuthi, ngokusebenza ezingeni eliphansi kangaka endleleni yokusebenza kwe-binary, i-BOT okwamanje iyasebenza. kuvinjelwe ezindaweni eziningi ezibucayiikakhulukazi emidlalweni eku-inthanethi enezinhlelo eziqinile zokulwa nokukhohlisa.
Izinkinga ngezinhlelo zokulwa nokukhohlisa kanye nokungabaza emhlabeni wokulinganisa
Enye yezindlela ezibucayi kakhulu ze-Intel BOT ukusebenzisana kwayo ne- izinhlelo zokulwa nokukhohlisa emidlalweni yabadlali abaningiNgenxa yokuthi ithuluzi lishintsha indlela i-binary esetshenziswa ngayo ngesikhathi sokusebenza, ezinye izinhlelo zokulwa nokukhohlisa, njenge-Ricochet noma i-Vanguard, zingase zichaze lokhu njengomzamo wokulawula umdlalo futhi ziwuphawule njengokuziphatha okusolisayo.
Lokhu kusho ukuthi, okwamanje, I-BOT ingase ingafaneleki ezinkundleni zokuxhumana ezincintisanayolapho ubuqotho bamakhasimende buyinto eza kuqala. Kuze kube nokuqondana okucacile phakathi kwe-Intel nabathengisi abalwa nokukhohlisa, noma izindlela ezithile zokuqinisekisa ukuthi lokhu kuthuthukiswa akulethi izinzuzo ezingafanele, lesi sici cishe sizohlala sinqunyelwe emidlalweni yomdlali oyedwa noma okuhlangenwe nakho ngaphandle kwezinhlelo ezinjalo zokulwa nokukhohlisa ezinolaka.
Elinye iphuzu elibalulekile lempikiswano livele emkhakheni wezilinganiso zokwenziwa. I-Primate Labs, inkampani ephethe i-Geekbench, ithe ukusetshenziswa kwama-bot kungenza ukuphazamisa ukusebenza kwemiphumelangoba kushintsha kakhulu ukuziphatha okulindelekile kwe-executable. Endaweni yokulinganisa, lapho "ukusebenza okumsulwa" kohlelo kufanele kulinganiswe khona, ungqimba lwangaphandle lokwenza ngcono olufana nalolu lushintsha umdlalo.
Ukuze kugcinwe ukucaca, i-Geekbench izomaka ngokusobala ukugijima lapho ithola khona ukungenelela kwe-Intel BOT. Inguqulo 6.7 yesilinganiso izokwengeza i- ifulegi elithile lokuhlonza imiphumela "ethuthukisiwe yi-BOT"ukuze zihlukaniswe kalula ezilinganisweni ezivamile futhi zingahlanganiswa zibe amazinga ngaphandle kokucaciswa okufanele.
Lesi simo sivula ingxoxo ethokozisayo mayelana nendlela okufanele sichaze ngayo izilinganiso esikhathini lapho isofthiwe yokwenza ngcono ingakwazi ukuhlela kabusha ngokuyisisekelo umsebenzi owenziwa yi-CPUUmugqa phakathi kwamandla "angempela" ehadiwe kanye nosizo olunikezwa yisofthiwe uyafiphala, futhi kusiphoqa ukuthi sicabange kabusha ngalokho esikulinganisayo lapho siqhathanisa amaprosesa amabili anezilungiselelo ezihlukile kakhulu.
Nokho, ngokombono womsebenzisi wokugcina, ingxoxo ayiyona ifilosofi kangako: uma umdlalo noma uhlelo lokusebenza lusebenza ngokushesha okukhulu kukhompyutha yabo ngoba i-Intel ikwazile ukuthuthukisa indlela yokusebenza ngaphandle kokudela ikhwalithi noma izici, umuzwa uzoba nje ukuthi "Iprosesa yayo isebenza kangcono"yize okunye kwalokho kunconywa kungenxa yesendlalelo sokwenza ngcono okubili hhayi i-silicon kuphela.
Izinzuzo ezisebenzayo kanye nekusasa le-Intel BOT uma kuqhathaniswa nomncintiswano
Uma sibheka isithombe esikhulu, i-BOT yethulwa njengohlobo "lwesikhali esiyimfihlo" se-Intel vala ingxenye yegebe lokusebenza bebhekene nezinketho ezinamandla kakhulu ezivela kumncintiswano, ikakhulukazi emidlalweni, ngaphambi kokufika kwezakhiwo zesikhathi esizayo njengeNova Lake kanye ne-BLLC yayo ethenjisiwe.
Inzuzo enkulu yaleli thuluzi ukuthi, lapho i-executable yokuqala ingalungiselelwanga kahle ukwakheka kwe-Intel, Inzuzo ingaba nkulu kakhulu. ngaphandle kokuthi umsebenzisi kudingeke ashintshe noma yini ngaphakathi komdlalo ngokwawo, noma ukuthi unjiniyela kudingeke akhiphe i-patch ethile. Kuyindlela yokubuyisela ukusebenza okulahlekile ezihlokweni ezazenzelwe amanye amapulatifomu noma ama-console ekuqaleni.
Ngaphezu kwalokho, i-BOT ayinciphisi ikhwalithi yokubuka noma izigcawu ezisikiwe noma i-physics: umsebenzi wohlelo uhlala ufana. Okushintshwayo ukuthi imiyalelo ihlelwe kanjani ukuze i-CPU... umatasa kakhulu wenza umsebenzi owusizo futhi ungalindi kakhulu. ngokusebenzisa idatha, ama-cache, noma izibikezeli. Ngokombono wesipiliyoni sokudlala, lokhu kuhunyushwa ku-FPS ezinzile kakhulu, futhi kwezinye izimo, ukuhlukahluka okuncane kokusebenza ezigcawini eziyinkimbinkimbi.
Inselele enkulu yesikhathi esizayo izoba yiyo ukukhuliswa nokwamukelwa emboniniUkuze i-Intel BOT ibe nomthelela omkhulu ngempela, uhlu lwemidlalo nezinhlelo zokusebenza ezisekelwayo kuzodingeka lukhule kakhulu, futhi abadlali ababalulekile ohlelweni lwe-ecosystem (isofthiwe ye-anticheat, abathuthukisi be-benchmark, izitudiyo zokuthuthukisa, njll.) kuzodingeka bazivumelanise nokuba khona kwalolu hlaka olusha lwe-middleware.
Kuzoba mnandi futhi ukubona ukuthi umncintiswano uphendula kanjani. Uma leli qhinga lokwenza ngcono i-binary lisebenza kahle futhi lamukelwa kahle yimakethe, bekungeke kumangaze uma ezinye izinkampani zingahlola amasu afanayo. izixazululo ezifanayo zokwenza ikhodi ngemuva kokusebenza kahle noma ujule kakhulu ekusebenziseni abashayeli kanye nezendlalelo eziphakathi ukuze uthole okuningi ngehadiwe.
Sekukonke, i-Intel Binary Optimization Tool ikhombisa amandla amakhulu kulabo abafuna ukucindezela yonke i-FPS yokugcina noma iphuzu elengeziwe emisebenzini ethile, kodwa inezindawo ezibalulekile ezingacacile maqondana nokuhambisana, ukucaca, kanye nobubanzi. Uma i-Intel ikwazi ukwandisa ukwesekwa kwayo, ithuthukise ukuhlanganiswa kwayo nesofthiwe yokulwa nokukhohlisa, futhi ilungise ukuba khona kwayo kuma-benchmarks, ingaba yi- ingxenye ebalulekile yesu labo lokusebenza eminyakeni ezayo, ngale kokuthuthukiswa kwehadiwe kuphela.
