{"id":102845,"date":"2022-01-03T15:36:38","date_gmt":"2022-01-03T13:36:38","guid":{"rendered":"https:\/\/prohoster.info\/blog\/novosti-interneta\/%d0%be%d0%bf%d1%83%d0%b1%d0%bb%d0%b8%d0%ba%d0%be%d0%b2%d0%b0%d0%bd-%d0%bd%d0%b0%d0%b1%d0%be%d1%80-%d0%bf%d0%b0%d1%82%d1%87%d0%b5%d0%b9-%d1%83%d1%81%d0%ba%d0%be%d1%80%d1%8f%d1%8e%d1%89%d0%b8%d1%85"},"modified":"2022-01-03T15:36:38","modified_gmt":"2022-01-03T13:36:38","slug":"%d0%be%d0%bf%d1%83%d0%b1%d0%bb%d0%b8%d0%ba%d0%be%d0%b2%d0%b0%d0%bd-%d0%bd%d0%b0%d0%b1%d0%be%d1%80-%d0%bf%d0%b0%d1%82%d1%87%d0%b5%d0%b9-%d1%83%d1%81%d0%ba%d0%be%d1%80%d1%8f%d1%8e%d1%89%d0%b8%d1%85","status":"publish","type":"post","link":"https:\/\/prohoster.info\/sq\/blog\/news\/%d0%be%d0%bf%d1%83%d0%b1%d0%bb%d0%b8%d0%ba%d0%be%d0%b2%d0%b0%d0%bd-%d0%bd%d0%b0%d0%b1%d0%be%d1%80-%d0%bf%d0%b0%d1%82%d1%87%d0%b5%d0%b9-%d1%83%d1%81%d0%ba%d0%be%d1%80%d1%8f%d1%8e%d1%89%d0%b8%d1%85","title":{"rendered":"Publikuar nj\u00eb grup patch-esh q\u00eb p\u00ebrshpejton nd\u00ebrtimin e b\u00ebrtham\u00ebs Linux nga 50-80%","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"
Ingo Molnar, nj\u00eb zhvillues i njohur i b\u00ebrtham\u00ebs Linux dhe autor i planifikuesit t\u00eb detyrave CFS (Completely Fair Scheduler), ka propozuar p\u00ebr diskutim n\u00eb list\u00ebn e shp\u00ebrndarjes s\u00eb zhvilluesve t\u00eb b\u00ebrtham\u00ebs Linux nj\u00eb seri patch-e, q\u00eb prekin m\u00eb shum\u00eb se gjysm\u00ebn e t\u00eb gjitha skedave n\u00eb tekstet e burimit t\u00eb b\u00ebrtham\u00ebs dhe q\u00eb ofrojn\u00eb nj\u00eb rritje t\u00eb shpejt\u00ebsis\u00eb s\u00eb rind\u00ebrtimit t\u00eb plot\u00eb t\u00eb b\u00ebrtham\u00ebs nga 50-80% n\u00eb var\u00ebsi t\u00eb konfigurimeve. Optimizimi i realizuar \u00ebsht\u00eb i ve\u00e7ant\u00eb sepse \u00ebsht\u00eb i shoq\u00ebruar me shtimin e grupit m\u00eb t\u00eb madh t\u00eb ndryshimeve n\u00eb histori t\u00eb zhvillimit t\u00eb b\u00ebrtham\u00ebs \u2014 p\u00ebrfshirja e 2297 patch-eve, q\u00eb ndryshojn\u00eb m\u00eb shum\u00eb se 25 mij\u00eb skeda (10 mij\u00eb skeda titujsh n\u00eb katalog\u00ebt 'include\/' dhe 'arch\/*\/include\/' dhe 15 mij\u00eb skeda me tekste burimi). <\/p>\n
Rritja e performanc\u00ebs arrihet p\u00ebrmes ndryshimit t\u00eb metod\u00ebs s\u00eb p\u00ebrpunimit t\u00eb skedave p\u00ebrfshir\u00ebse. Theksohet se, pas tridhjet\u00eb viteve t\u00eb zhvillimit t\u00eb b\u00ebrtham\u00ebs, gjendja e skedave p\u00ebrfshir\u00ebs ka kaluar n\u00eb nj\u00eb gjendje t\u00eb d\u00ebshp\u00ebruar p\u00ebr shkak t\u00eb numrit t\u00eb madh t\u00eb var\u00ebsive t\u00eb nd\u00ebrthurura midis skedave. Puna p\u00ebr rikonstruksionin e skedave p\u00ebrfshir\u00ebs ka zgjatur m\u00eb shum\u00eb se nj\u00eb vit dhe ka k\u00ebrkuar nj\u00eb ripunim t\u00eb r\u00ebnd\u00ebsish\u00ebm t\u00eb hierarkis\u00eb dhe var\u00ebsive. Gjat\u00eb rikonstruksionit \u00ebsht\u00eb realizuar pun\u00eb p\u00ebr ndarjen e definicioneve t\u00eb tipeve dhe API-ve p\u00ebr n\u00ebn-sistemet e ndryshme t\u00eb b\u00ebrtham\u00ebs. <\/p>\n
Mes ndryshimeve t\u00eb b\u00ebra: ndarja e skedave titujsh t\u00eb nivelit t\u00eb lart\u00eb nga nj\u00ebri tjetri, p\u00ebrjashtimi i lidhjeve t\u00eb skedave titujsh p\u00ebr funksionet inline, ndarja e skedave titujsh p\u00ebr llojet dhe API-t\u00eb, sigurimi i ndarjes s\u00eb nd\u00ebrtimit t\u00eb skedave titujsh (m\u00eb shum\u00eb se 80 skeda kishin var\u00ebsi t\u00eb drejtp\u00ebrdrejta q\u00eb pengonin nd\u00ebrtimin, q\u00eb ishin paraqitur p\u00ebrmes skedave t\u00eb tjera), shtimi automatik i var\u00ebsive p\u00ebr skedat '.h' dhe '.c', optimizimi hap pas hapi i skedave titujsh, p\u00ebrdorimi i modit 'CONFIG_KALLSYMS_FAST=y', konsolidimi selektiv i skedave C n\u00eb nj\u00ebsi nd\u00ebrtimi p\u00ebr t\u00eb reduktuar numrin e skedave objekti. <\/p>\n
Si rezultat, puna e kryer ka lejuar t\u00eb reduktohet p\u00ebr 1-2 rend n\u00eb madh\u00ebsin\u00eb e skedave titujsh q\u00eb p\u00ebrpunohen n\u00eb faz\u00ebn e post-p\u00ebrpunimit. P\u00ebr shembull, para optimizimit, p\u00ebrdorimi i sked\u00ebs titujsh 'linux\/gfp.h' \u00e7onte n\u00eb shtimin e 13543 linjave t\u00eb kodit dhe lidhjen e 303 skedave titujsh t\u00eb varur, nd\u00ebrsa pas optimizimit madh\u00ebsia u zvog\u00eblua n\u00eb 181 linja dhe 26 skeda t\u00eb varura. Ose nj\u00eb shembull tjet\u00ebr: gjat\u00eb p\u00ebrpunimit t\u00eb sked\u00ebs 'kernel\/pid.c' pa patch, lidhen 94 mij\u00eb linja kodi, shumica e t\u00eb cilave nuk p\u00ebrdoren n\u00eb pid.c. Ndara e skedave titujsh leht\u00ebsoi reduktimin e volumit t\u00eb kodit t\u00eb p\u00ebrpunuar me nj\u00eb t\u00eb tret\u00ebn, duke zvog\u00ebluar numrin e linjave t\u00eb p\u00ebrpunuara n\u00eb 36 mij\u00eb. <\/p>\n
Me nj\u00eb p\u00ebrmbledhje t\u00eb plot\u00eb t\u00eb b\u00ebrjes s\u00eb b\u00ebrtham\u00ebs me komand\u00ebn \u00abmake -j96 vmlinux\u00bb, aplikimi i patch-eve n\u00eb sistemin testues tregoi nj\u00eb reduktim t\u00eb koh\u00ebs s\u00eb nd\u00ebrtimit t\u00eb deg\u00ebs v5.16-rc7 nga 231.34 n\u00eb 129.97 sekonda (nga 15.5 n\u00eb 27.7 nd\u00ebrtime n\u00eb or\u00eb), si dhe rriti efikasitetin e p\u00ebrdorimit t\u00eb b\u00ebrthamave t\u00eb CPU gjat\u00eb nd\u00ebrtimit. Me nd\u00ebrtimin inkremental, efekti i optimizimit \u00ebsht\u00eb edhe m\u00eb i duksh\u00ebm \u2014 koha e rikonstruktimeve pas ndryshimeve n\u00eb skedar\u00ebt e titujve u reduktua ndjesh\u00ebm (nga 112% n\u00eb 173% n\u00eb var\u00ebsi t\u00eb skedarit t\u00eb titujve q\u00eb ndryshohet). Optimizimet aktualisht jan\u00eb t\u00eb disponueshme vet\u00ebm p\u00ebr arkitektur\u00ebn ARM64, MIPS, Sparc dhe x86 (32 dhe 64-bit). \n Burimi: opennet.ru<\/a> <\/p>","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"