{"id":181858,"date":"2026-06-02T14:48:21","date_gmt":"2026-06-02T12:48:22","guid":{"rendered":"https:\/\/prohoster.info\/blog\/novosti-interneta\/opublikovan-nbd-vram"},"modified":"2026-06-02T14:48:21","modified_gmt":"2026-06-02T12:48:22","slug":"opublikovan-nbd-vram","status":"publish","type":"post","link":"https:\/\/prohoster.info\/cs\/blog\/news\/opublikovan-nbd-vram","title":{"rendered":"Byl zve\u0159ejn\u011bn NBD-VRAM","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"<p>Byl zve\u0159ejn\u011bn otev\u0159en\u00fd projekt <strong>NBD-VRAM<\/strong>, kter\u00fd umo\u017e\u0148uje vyu\u017e\u00edt \u010d\u00e1st video pam\u011bti NVIDIA GPU jako swap prostor v Linuxu. Projekt je zam\u011b\u0159en p\u0159edev\u0161\u00edm na notebooky s integrovanou opera\u010dn\u00ed pam\u011bt\u00ed, kde nen\u00ed mo\u017en\u00e9 RAM roz\u0161\u00ed\u0159it, ale p\u0159itom je v syst\u00e9mu diskr\u00e9tn\u00ed grafick\u00e1 karta NVIDIA RTX\/GTX s nevyu\u017e\u00edvanou VRAM. K\u00f3d je naps\u00e1n v C a shellu, je \u0161\u00ed\u0159en pod licenc\u00ed <strong>MIT<\/strong>.<\/p>\n<p>My\u0161lenka NBD-VRAM je jednoduch\u00e1: pokud syst\u00e9m ji\u017e za\u010d\u00edn\u00e1 p\u0159ech\u00e1zet na swap na SSD, je mo\u017en\u00e9 um\u00edstit p\u0159ed SSD dal\u0161\u00ed meziprostor \u2014 video pam\u011b\u0165. Autor uv\u00e1d\u00ed p\u0159\u00edklad s notebookem na <strong>RTX 3070 Laptop<\/strong>: z 8 GB VRAM bylo vy\u010dlen\u011bno 7 GB na swap, a celkov\u011b s RAM, zram a SSD swap syst\u00e9m z\u00edskal p\u0159ibli\u017en\u011b 46 GB adresovateln\u00e9 pam\u011bti. P\u0159edpokl\u00e1dan\u00fd po\u0159\u00e1dek p\u0159ete\u010den\u00ed je n\u00e1sleduj\u00edc\u00ed: nejprve se pou\u017e\u00edv\u00e1 RAM, pot\u00e9 VRAM jako rychl\u00fd swap, n\u00e1sledn\u011b zram, a teprve pot\u00e9 SSD.<\/p>\n<p>Technicky NBD-VRAM nep\u0159id\u00e1v\u00e1 nov\u00fd ovlada\u010d j\u00e1dra. Mal\u00fd d\u00e9mon vy\u010dle\u0148uje pam\u011b\u0165 grafick\u00e9 karty prost\u0159ednictv\u00edm <strong>CUDA Driver API<\/strong>, pot\u00e9 ji p\u0159ed\u00e1v\u00e1 j\u00e1dru Linuxu jako blokov\u00e9 za\u0159\u00edzen\u00ed prost\u0159ednictv\u00edm <strong>NBD<\/strong> \u2014 Network Block Device \u2014 p\u0159es Unix socket. Po p\u0159ipojen\u00ed standardn\u00edm nbd-client se v syst\u00e9mu objev\u00ed \/dev\/nbdX, kter\u00e9 lze naform\u00e1tovat jako b\u011b\u017en\u00fd swap prost\u0159ednictv\u00edm mkswap a aktivovat p\u0159es swapon.<\/p>\n<p>Autor zvl\u00e1\u0161\u0165 zd\u016fraz\u0148uje, \u017ee tento p\u0159\u00edstup byl zvolen kv\u016fli omezen\u00edm spot\u0159ebitelsk\u00fdch grafick\u00fdch karet NVIDIA. P\u0159\u00edm\u011bj\u0161\u00ed cesta prost\u0159ednictv\u00edm NVIDIA P2P API na GeForce, podle n\u011bj, narazila na EINVAL, proto\u017ee p\u0159\u00edslu\u0161n\u00e9 funkce jsou fakticky dostupn\u00e9 pouze pro profesion\u00e1ln\u00ed a serverov\u00e9 modely. Varianta s p\u0159\u00edm\u00fdm p\u0159\u00edstupem k BAR1 tak\u00e9 nefungovala: k dispozici je pouze mal\u00e9 mapovan\u00e9 pole, a \u010dten\u00ed z ostatn\u00ed \u010d\u00e1sti vrac\u00ed nuly. NBD p\u0159\u00edstup obch\u00e1z\u00ed toto omezen\u00ed, proto\u017ee pou\u017e\u00edv\u00e1 b\u011b\u017en\u00e9 CUDA kop\u00edrov\u00e1n\u00ed cuMemcpyHtoD a cuMemcpyDtoH.<\/p>\n<h2>Funkce<\/h2>\n<ul>\n<li>\n<p><strong>Vyu\u017eit\u00ed VRAM jako b\u011b\u017en\u00e9ho Linux swapu.<\/strong> Po spu\u0161t\u011bn\u00ed d\u00e9mon poskytuje video pam\u011b\u0165 jako \/dev\/nbd0 nebo jin\u00e9 NBD za\u0159\u00edzen\u00ed, kter\u00e9 pro j\u00e1dro vypad\u00e1 jako standardn\u00ed blokov\u00e9 za\u0159\u00edzen\u00ed.<\/p>\n<\/li>\n<li>\n<p><strong>Fungov\u00e1n\u00ed bez vlastn\u00edho modulu j\u00e1dra.<\/strong> Projekt nevy\u017eaduje psan\u00ed, sestavov\u00e1n\u00ed a spr\u00e1vu samostatn\u00e9ho modul j\u00e1dra, nevyu\u017e\u00edv\u00e1 intern\u00ed symboly NVIDIA ovlada\u010de a m\u011bl by p\u0159e\u017e\u00edt aktualizace j\u00e1dra a ovlada\u010de bez op\u011btovn\u00e9ho sestaven\u00ed.<\/p>\n<\/li>\n<li>\n<p><strong>Zam\u011b\u0159en\u00ed na spot\u0159ebitelsk\u00e9 NVIDIA GPU.<\/strong> V po\u017eadavc\u00edch jsou uvedeny NVIDIA GPU s podporou CUDA, v\u010detn\u011b spot\u0159ebitelsk\u00fdch karet RTX\/GTX, ofici\u00e1ln\u00edho NVIDIA ovlada\u010de s libcuda.so.1, modulu nbd v j\u00e1d\u0159e Linux, nbd-client, gcc a make. CUDA Toolkit v tomto p\u0159\u00edpad\u011b nen\u00ed pot\u0159eba.<\/p>\n<\/li>\n<li>\n<p><strong>Integrace systemd.<\/strong> Instalace pomoc\u00ed install.sh p\u0159id\u00e1v\u00e1 slu\u017ebu vram-swap-nbd, kterou lze spou\u0161t\u011bt p\u0159es systemctl; po instalaci se slu\u017eba povol\u00ed pro automatick\u00e9 spu\u0161t\u011bn\u00ed p\u0159i startu.<\/p>\n<\/li>\n<li>\n<p><strong>Nastaven\u00ed velikosti a priority swap.<\/strong> V systemd-unit lze nastavit VRAM_SETUP_SIZE_MB, tedy horn\u00ed limit alokovan\u00e9 VRAM, a VRAM_SWAP_PRIORITY, tedy prioritu swapov\u00e9ho za\u0159\u00edzen\u00ed. \u010c\u00edm vy\u0161\u0161\u00ed je priorita, t\u00edm d\u0159\u00edve Linux pou\u017eije tuto vrstvu swapu.<\/p>\n<\/li>\n<li>\n<p><strong>Automatick\u00e9 zmen\u0161en\u00ed po\u017eadovan\u00e9 velikosti.<\/strong> Pokud po\u017eadovan\u00fd objem VRAM nen\u00ed k dispozici, d\u00e9mon se sna\u017e\u00ed zmen\u0161it velikost po bloc\u00edch po 512 MiB, aby p\u0159esto alokoval dostupn\u00fd objem, nap\u0159\u00edklad pokud je \u010d\u00e1st pam\u011bti ji\u017e obsazena compositorem nebo grafickou relac\u00ed.<\/p>\n<\/li>\n<li>\n<p><strong>Kontroln\u00ed skripty.<\/strong> V repozit\u00e1\u0159i je test-nbd.sh pro kou\u0159ov\u00fd test se z\u00e1pisem\/\u010dten\u00edm 1 MiB a test-fill.sh pro stresovou zkou\u0161ku cel\u00e9 VRAM odd\u00edlu.<\/p>\n<\/li>\n<li>\n<p><strong>Uveden\u00fd v\u00fdkon kolem 1,3 GB\/s.<\/strong> Na RTX 3070 Laptop autor zm\u011b\u0159il sekven\u010dn\u00ed z\u00e1pis 7 GB bloky po 4 MB p\u0159ibli\u017en\u011b na \u00farovni 1,3 GB\/s.<\/p>\n<\/li>\n<\/ul>\n<h2>Sc\u00e9n\u00e1\u0159e pou\u017eit\u00ed<\/h2>\n<p><strong>Notebooky s integrovanou RAM.<\/strong> Hlavn\u00ed sc\u00e9n\u00e1\u0159 \u2013 modern\u00ed notebooky, kde 16 nebo 32 GB opera\u010dn\u00ed pam\u011bti ji\u017e nesta\u010d\u00ed, ale nen\u00ed mo\u017en\u00e9 ji roz\u0161\u00ed\u0159it. Pokud m\u00e1 takov\u00fd stroj dedikovanou kartu RTX, \u010d\u00e1st VRAM lze pou\u017e\u00edt jako dal\u0161\u00ed swapovou vrstvu. To neprom\u011bn\u00ed VRAM na plnohodnotnou RAM, ale m\u016f\u017ee zachr\u00e1nit syst\u00e9m p\u0159ed n\u00e1hl\u00fdm p\u0159echodem na pomal\u00fd SSD swap nebo p\u0159ed OOM-killerem p\u0159i \u0161pi\u010dkov\u00fdch z\u00e1t\u011b\u017e\u00edch.<\/p>\n<p><strong>T\u011b\u017ek\u00e9 pracovn\u00ed prost\u0159ed\u00ed v\u00fdvoj\u00e1\u0159e.<\/strong> IDE, prohl\u00ed\u017ee\u010d s des\u00edtkami z\u00e1lo\u017eek, kontejnery Docker, m\u00edstn\u00ed datab\u00e1ze, sestaven\u00ed velk\u00fdch projekt\u016f a testovac\u00ed prost\u0159ed\u00ed snadno vytv\u00e1\u0159ej\u00ed kr\u00e1tkodob\u00e9 \u0161pi\u010dky v spot\u0159eb\u011b pam\u011bti. V takov\u00e9m sc\u00e9n\u00e1\u0159i m\u016f\u017ee NBD-VRAM fungovat jako buffer: neurychluje b\u011b\u017enou pr\u00e1ci, ale zm\u00edr\u0148uje okam\u017eiky, kdy RAM do\u0161la.<\/p>\n<p><strong>Sn\u00ed\u017een\u00ed z\u00e1t\u011b\u017ee na SSD swap.<\/strong> Pokud se na SSD \u010dasto pou\u017e\u00edv\u00e1 swap, nen\u00ed to jen pomalej\u0161\u00ed, ale tak\u00e9 to vytv\u00e1\u0159\u00ed zbyte\u010dn\u00e9 z\u00e1pisy na \u00falo\u017ei\u0161t\u011b. VRAM-swap lze nastavit s vy\u0161\u0161\u00ed prioritou, aby syst\u00e9m p\u0159i zapln\u011bn\u00ed RAM nejprve vyh\u00e1n\u011bl str\u00e1nky do grafick\u00e9 pam\u011bti a teprve pot\u00e9 se obracel na SSD. To je zvl\u00e1\u0161\u0165 aktu\u00e1ln\u00ed pro notebooky, kde je SSD \u010dasto neodn\u00edmateln\u00fd nebo drah\u00fd na v\u00fdm\u011bnu.<\/p>\n<p><strong>Kombinace se zram.<\/strong> Autor p\u0159\u00edmo popisuje sch\u00e9ma, kde VRAM swap dost\u00e1v\u00e1 vy\u0161\u0161\u00ed prioritu a jako prvn\u00ed \"nadzved\u00e1v\u00e1\" pam\u011b\u0165, zram se pou\u017e\u00edv\u00e1 na dal\u0161\u00ed \u00farovni a SSD z\u016fst\u00e1v\u00e1 posledn\u00ed lini\u00ed obrany. Takov\u00e9 sch\u00e9ma m\u016f\u017ee b\u00fdt u\u017eite\u010dn\u00e9 pro pracovn\u00ed stanice a notebooky, kde je d\u016fle\u017eit\u011bj\u0161\u00ed udr\u017eet reak\u010dnost syst\u00e9mu p\u0159i nedostatku pam\u011bti ne\u017e dos\u00e1hnout maxim\u00e1ln\u00ed p\u0159edv\u00eddatelnosti latenc\u00ed.<\/p>\n<p><strong>M\u00edstn\u00ed \u00falohy AI\/LLM kolem GPU, ale ne m\u00edsto VRAM pro model.<\/strong> NBD-VRAM nezvy\u0161uje grafickou pam\u011b\u0165 dostupnou pro CUDA aplikaci jako VRAM pro model. Je to obr\u00e1cen\u00fd sc\u00e9n\u00e1\u0159: nen\u00ed RAM pou\u017e\u00edv\u00e1na jako VRAM, ale VRAM je pou\u017e\u00edv\u00e1na jako swap pro b\u011b\u017enou pam\u011b\u0165 Linuxu. Proto projekt neumo\u017en\u00ed p\u0159\u00edmo nahr\u00e1t do GPU model v\u011bt\u0161\u00ed velikosti. M\u016f\u017ee b\u00fdt v\u0161ak u\u017eite\u010dn\u00fd na stroji, kde vedle LLM inference b\u011b\u017e\u00ed prohl\u00ed\u017ee\u010d, IDE, index\u00e1tory, Python prost\u0159ed\u00ed a kontejnery, a syst\u00e9mov\u00e1 RAM za\u010d\u00edn\u00e1 doch\u00e1zet.<\/p>\n<p><strong>Dom\u00e1c\u00ed a experiment\u00e1ln\u00ed pracovn\u00ed stanice.<\/strong> Projekt je zaj\u00edmav\u00fd pro u\u017eivatele, jejich\u017e grafick\u00e1 karta \u010dasto ne\u010dinn\u011b stoj\u00ed mimo hry, rendering nebo ML \u00falohy. Nap\u0159\u00edklad 8-12 GB VRAM na desktopov\u00e9 GeForce lze do\u010dasn\u011b p\u0159em\u011bnit na dodate\u010dnou vrstvu swap pro t\u011b\u017ek\u00e9 \u00falohy kompilace, zpracov\u00e1n\u00ed dat nebo spu\u0161t\u011bn\u00ed. <a class=\"wpil_keyword_link\" href=\"https:\/\/prohoster.info\/cs\/vps\/\" title=\"virtu\u00e1ln\u00edch stroj\u016f\" data-wpil-keyword-link=\"linked\">virtu\u00e1ln\u00edch stroj\u016f<\/a>.<\/p>\n<h2>Omezen\u00ed<\/h2>\n<p>NBD-VRAM nen\u00ed n\u00e1hradou za opera\u010dn\u00ed pam\u011b\u0165. P\u0159\u00edstup k takov\u00e9mu swapu prob\u00edh\u00e1 \u0159et\u011bzcem kernel swap \u2192 \/dev\/nbdX \u2192 nbd driver \u2192 Unix socket \u2192 daemon \u2192 CUDA copy \u2192 VRAM, tak\u017ee latence a chov\u00e1n\u00ed se budou li\u0161it od skute\u010dn\u00e9 RAM. Je to sp\u00ed\u0161e nouzov\u00e1 nebo p\u0159echodn\u00e1 vrstva mezi RAM a SSD ne\u017e zp\u016fsob \"p\u0159idat pam\u011b\u0165\" bez n\u00e1sledk\u016f.<\/p>\n<p>Projekt je tak\u00e9 v\u00e1z\u00e1n na ofici\u00e1ln\u00ed stack NVIDIA s CUDA. Nouveau\/Nova pro toto nevyhovuj\u00ed, proto\u017ee vy\u017eaduje libcuda.so.1. Phoronix <a rel=\"nofollow\" href=\"https:\/\/www.phoronix.com\/news\/NVIDIA-NBD-VRAM\">tak\u00e9 zaznamen\u00e1v\u00e1<\/a>, \u017ee NBD-VRAM byl vytvo\u0159en pr\u00e1v\u011b pro spot\u0159ebitelsk\u00e9 NVIDIA GPU, kde alternativn\u00ed p\u0159\u00edstupy pomoc\u00ed NVIDIA P2P API nefunguj\u00ed.<\/p>\n<p>Ve zjednodu\u0161en\u00e9 podob\u011b, NBD-VRAM je mal\u00fd, ale zaj\u00edmav\u00fd syst\u00e9mov\u00fd hack pro Linux: ned\u011bl\u00e1 z\u00e1zraky a nenahrazuje upgrade RAM, ale umo\u017e\u0148uje vyu\u017e\u00edt nevyu\u017eitou video pam\u011b\u0165 jako dal\u0161\u00ed vrstvu swap p\u0159ed SSD. Pro notebooky s p\u0159ip\u00e1jenou pam\u011bt\u00ed a dedikovanou kartou RTX m\u016f\u017ee b\u00fdt praktick\u00fdm zp\u016fsobem, jak zvl\u00e1dat \u0161pi\u010dkov\u00e9 zat\u00ed\u017een\u00ed bez okam\u017eit\u00e9ho p\u00e1du aplikac\u00ed nebo bolestiv\u00e9ho p\u0159echodu na pomal\u00e9 \u00falo\u017ei\u0161t\u011b.<\/p>\n<p>Zdroj: <a rel=\"nofollow\" href=\"https:\/\/www.linux.org.ru\/news\/hardware\/18308769\">linux.org.ru<\/a><\/p>","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"<p>\u041e\u043f\u0443\u0431\u043b\u0438\u043a\u043e\u0432\u0430\u043d \u043e\u0442\u043a\u0440\u044b\u0442\u044b\u0439 \u043f\u0440\u043e\u0435\u043a\u0442 NBD-VRAM, \u043f\u043e\u0437\u0432\u043e\u043b\u044f\u044e\u0449\u0438\u0439 \u0437\u0430\u0434\u0435\u0439\u0441\u0442\u0432\u043e\u0432\u0430\u0442\u044c \u0447\u0430\u0441\u0442\u044c \u0432\u0438\u0434\u0435\u043e\u043f\u0430\u043c\u044f\u0442\u0438 NVIDIA GPU \u043a\u0430\u043a swap-\u043f\u0440\u043e\u0441\u0442\u0440\u0430\u043d\u0441\u0442\u0432\u043e \u0432 Linux. \u041f\u0440\u043e\u0435\u043a\u0442 \u043e\u0440\u0438\u0435\u043d\u0442\u0438\u0440\u043e\u0432\u0430\u043d \u043f\u0440\u0435\u0436\u0434\u0435 \u0432\u0441\u0435\u0433\u043e \u043d\u0430 \u043d\u043e\u0443\u0442\u0431\u0443\u043a\u0438 \u0441 \u0440\u0430\u0441\u043f\u0430\u044f\u043d\u043d\u043e\u0439 \u043e\u043f\u0435\u0440\u0430\u0442\u0438\u0432\u043d\u043e\u0439 \u043f\u0430\u043c\u044f\u0442\u044c\u044e, \u0433\u0434\u0435 RAM \u043d\u0435\u043b\u044c\u0437\u044f \u0440\u0430\u0441\u0448\u0438\u0440\u0438\u0442\u044c, \u043d\u043e \u043f\u0440\u0438 \u044d\u0442\u043e\u043c \u0432 \u0441\u0438\u0441\u0442\u0435\u043c\u0435 \u0435\u0441\u0442\u044c \u0434\u0438\u0441\u043a\u0440\u0435\u0442\u043d\u0430\u044f \u0432\u0438\u0434\u0435\u043e\u043a\u0430\u0440\u0442\u0430 NVIDIA RTX\/GTX \u0441 \u043d\u0435\u0438\u0441\u043f\u043e\u043b\u044c\u0437\u0443\u0435\u043c\u043e\u0439 VRAM. \u041a\u043e\u0434 \u043d\u0430\u043f\u0438\u0441\u0430\u043d \u043d\u0430 C \u0438 shell, \u0440\u0430\u0441\u043f\u0440\u043e\u0441\u0442\u0440\u0430\u043d\u044f\u0435\u0442\u0441\u044f \u043f\u043e\u0434 \u043b\u0438\u0446\u0435\u043d\u0437\u0438\u0435\u0439 MIT. \u0418\u0434\u0435\u044f NBD-VRAM \u043f\u0440\u043e\u0441\u0442\u0430: \u0435\u0441\u043b\u0438 [&hellip;]<\/p>\n","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"author":8,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[702],"tags":[],"class_list":["post-181858","post","type-post","status-publish","format-standard","hentry","category-news"],"aioseo_notices":[],"aioseo_head":"\n\t\t<!-- All in One SEO 5.0.3 - aioseo.com -->\n\t<meta name=\"description\" content=\"\u041e\u043f\u0443\u0431\u043b\u0438\u043a\u043e\u0432\u0430\u043d \u043e\u0442\u043a\u0440\u044b\u0442\u044b\u0439 \u043f\u0440\u043e\u0435\u043a\u0442 NBD-VRAM, \u043f\u043e\u0437\u0432\u043e\u043b\u044f\u044e\u0449\u0438\u0439 \u0437\u0430\u0434\u0435\u0439\u0441\u0442\u0432\u043e\u0432\u0430\u0442\u044c \u0447\u0430\u0441\u0442\u044c \u0432\u0438\u0434\u0435\u043e\u043f\u0430\u043c\u044f\u0442\u0438 NVIDIA GPU \u043a\u0430\u043a swap-\u043f\u0440\u043e\u0441\u0442\u0440\u0430\u043d\u0441\u0442\u0432\u043e \u0432 Linux.\" \/>\n\t<meta name=\"robots\" content=\"max-image-preview:large\" \/>\n\t<meta name=\"author\" content=\"Erik Peterson\"\/>\n\t<link rel=\"canonical\" href=\"https:\/\/prohoster.info\/cs\/blog\/news\/opublikovan-nbd-vram\" \/>\n\t\t<meta name=\"generator\" content=\"All in One SEO (AIOSEO) 5.0.3\" \/>\n\t\t<meta property=\"og:locale\" content=\"cs_CZ\" \/>\n\t\t<meta property=\"og:site_name\" content=\"ProHoster | \u041a\u0443\u043f\u0438\u0442\u044c \u043d\u0430\u0434\u0435\u0436\u043d\u044b\u0439 \u0445\u043e\u0441\u0442\u0438\u043d\u0433 \u0434\u043b\u044f \u0441\u0430\u0439\u0442\u043e\u0432 \u0441 \u0437\u0430\u0449\u0438\u0442\u043e\u0439 \u043e\u0442 DDoS, VPS VDS \u0441\u0435\u0440\u0432\u0435\u0440\u044b\" \/>\n\t\t<meta property=\"og:type\" content=\"article\" \/>\n\t\t<meta property=\"og:title\" content=\"\ud83e\udd47\u041e\u043f\u0443\u0431\u043b\u0438\u043a\u043e\u0432\u0430\u043d NBD-VRAM | ProHoster\" \/>\n\t\t<meta property=\"og:description\" content=\"\u041e\u043f\u0443\u0431\u043b\u0438\u043a\u043e\u0432\u0430\u043d \u043e\u0442\u043a\u0440\u044b\u0442\u044b\u0439 \u043f\u0440\u043e\u0435\u043a\u0442 NBD-VRAM, \u043f\u043e\u0437\u0432\u043e\u043b\u044f\u044e\u0449\u0438\u0439 \u0437\u0430\u0434\u0435\u0439\u0441\u0442\u0432\u043e\u0432\u0430\u0442\u044c \u0447\u0430\u0441\u0442\u044c \u0432\u0438\u0434\u0435\u043e\u043f\u0430\u043c\u044f\u0442\u0438 NVIDIA GPU \u043a\u0430\u043a swap-\u043f\u0440\u043e\u0441\u0442\u0440\u0430\u043d\u0441\u0442\u0432\u043e \u0432 Linux.\" \/>\n\t\t<meta property=\"og:url\" content=\"https:\/\/prohoster.info\/cs\/blog\/news\/opublikovan-nbd-vram\" \/>\n\t\t<meta property=\"og:image\" content=\"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg\" \/>\n\t\t<meta property=\"og:image:secure_url\" content=\"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg\" \/>\n\t\t<meta property=\"og:image:width\" content=\"350\" \/>\n\t\t<meta property=\"og:image:height\" content=\"350\" \/>\n\t\t<meta property=\"article:published_time\" content=\"2026-06-02T12:48:22+00:00\" \/>\n\t\t<meta property=\"article:modified_time\" content=\"2026-06-02T12:48:22+00:00\" \/>\n\t\t<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/prohoster\" \/>\n\t\t<meta property=\"article:author\" content=\"https:\/\/www.facebook.com\/prohoster\" \/>\n\t\t<!-- All in One SEO -->\n\n","aioseo_head_json":{"title":"\ud83e\udd47Byl zve\u0159ejn\u011bn NBD-VRAM | ProHoster","description":"Byl zve\u0159ejn\u011bn otev\u0159en\u00fd projekt NBD-VRAM, kter\u00fd umo\u017e\u0148uje vyu\u017e\u00edt \u010d\u00e1st video pam\u011bti GPU NVIDIA jako swapov\u00e9 m\u00edsto v Linuxu.","canonical_url":"https:\/\/prohoster.info\/cs\/blog\/news\/opublikovan-nbd-vram","robots":"max-image-preview:large","keywords":"","webmasterTools":{"miscellaneous":""},"schema":null,"og:locale":"cs_CZ","og:site_name":"ProHoster | \u041a\u0443\u043f\u0438\u0442\u044c \u043d\u0430\u0434\u0435\u0436\u043d\u044b\u0439 \u0445\u043e\u0441\u0442\u0438\u043d\u0433 \u0434\u043b\u044f \u0441\u0430\u0439\u0442\u043e\u0432 \u0441 \u0437\u0430\u0449\u0438\u0442\u043e\u0439 \u043e\u0442 DDoS, VPS VDS \u0441\u0435\u0440\u0432\u0435\u0440\u044b","og:type":"article","og:title":"\ud83e\udd47\u041e\u043f\u0443\u0431\u043b\u0438\u043a\u043e\u0432\u0430\u043d NBD-VRAM | ProHoster","og:description":"\u041e\u043f\u0443\u0431\u043b\u0438\u043a\u043e\u0432\u0430\u043d \u043e\u0442\u043a\u0440\u044b\u0442\u044b\u0439 \u043f\u0440\u043e\u0435\u043a\u0442 NBD-VRAM, \u043f\u043e\u0437\u0432\u043e\u043b\u044f\u044e\u0449\u0438\u0439 \u0437\u0430\u0434\u0435\u0439\u0441\u0442\u0432\u043e\u0432\u0430\u0442\u044c \u0447\u0430\u0441\u0442\u044c \u0432\u0438\u0434\u0435\u043e\u043f\u0430\u043c\u044f\u0442\u0438 NVIDIA GPU \u043a\u0430\u043a swap-\u043f\u0440\u043e\u0441\u0442\u0440\u0430\u043d\u0441\u0442\u0432\u043e \u0432 Linux.","og:url":"https:\/\/prohoster.info\/cs\/blog\/news\/opublikovan-nbd-vram","og:image":"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg","og:image:secure_url":"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg","og:image:width":350,"og:image:height":350,"article:published_time":"2026-06-02T12:48:22+00:00","article:modified_time":"2026-06-02T12:48:22+00:00","article:publisher":"https:\/\/www.facebook.com\/prohoster","article:author":"https:\/\/www.facebook.com\/prohoster"},"aioseo_meta_data":[],"gt_translate_keys":[{"key":"link","format":"url"}],"_links":{"self":[{"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/posts\/181858","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/users\/8"}],"replies":[{"embeddable":true,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/comments?post=181858"}],"version-history":[{"count":1,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/posts\/181858\/revisions"}],"predecessor-version":[{"id":182088,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/posts\/181858\/revisions\/182088"}],"wp:attachment":[{"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/media?parent=181858"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/categories?post=181858"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/tags?post=181858"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}