Uhlaziyo lwabaqhubi be-NVIDIA 440.100 kunye ne-390.138 kunye nobuthathaka bususiwe

Inkampani yeNVIDIA thaca ukukhutshwa kwabaqhubi abanini I-NVIDIA 440.100 (LTS) kunye 390.138 ("ilifa" le-GPU GF1xx "Fermi"), eyaphelisa ingozi ubuthathaka, okunokukuvumela ukuba unyuse amalungelo akho kwinkqubo. Umqhubi uyafumaneka Linux (ARM, x86_64), iFreeBSD (x86_64) kunye neSolaris (x86_64).

• I-CVE-2020-5963 bubuthathaka kwiCUDA Driver Interprocess Communication API enokukhokelela ekukhanyeni kwenkonzo, ukubulawa kwekhowudi ephakamileyo, okanye ukuvuza kolwazi.
• I-CVE-2020-5967 bubuthathaka kumqhubi we-UVM obangelwa yimeko yogqatso enokukhokelela ekukhanyeni kwenkonzo.

Inguqulo ye-440.100 ikwabandakanya inkxaso yee-GPU ezintsha
IGeForce GTX 1650 Ti, iGeForce GTX 1650 Ti eneMax-Q, iGeForce RTX 2060 eneMax-Q kunye neQuadro T1000 eneMax-Q, isixhobo esingachazwanga igama lesixhobo se-“Connector-N” songezelelwe useto lwe-X11, olunokusetyenziswa kuyo. i ConnectedMonitor ukhetho lokulinganisa ukudibanisa imonitha ngaphandle kokwazi iindlela ezikhoyo zoqhagamshelwano.
Inguqulelo 390.132 yongeza ukuhambelana kwekernel Linux I-5.6 kunye Oracle Linux 7.7, nakwienkqubo ezine-kernel Linux 5.4 yongeze inkxaso kwi-PRIME Synchronization.

Ukongeza, waqala Ukuvavanya inguqulelo ye-beta yesebe elitsha le-450.x, apho oku kuphuculwa kulandelayo kunokuqatshelwa:

• Inkxaso eyongeziweyo ye-GPU A100-PCIE-40GB, A100-PG509-200,
  A100-SXM4-40GB, GeForce GTX 1650 Ti, GeForce RTX 2060 kunye Max-Q kunye
  I-Quadro T1000 ene-Max-Q;
• IVulkan API ngoku ixhasa umboniso othe ngqo kwiziboniso eziqhagamshelwe ngeDisplayPort Multi-Stream Transport (DP-MST);
• Inkxaso eyongeziweyo yolwandiso lwe-OpenGL glNamedBufferPageCommitmentARB;
• Ilayibrari eyongeziweyo ye-libnvidia-ngx.so ngokuphunyezwa kwenkxaso yetekhnoloji NVIDIA NGX;
• Ukubonwa okuphuculweyo kwezixhobo ezenziwe nguVulkan kwiinkqubo ezinomncedisi we-X.Org;
• Ithala leencwadi le-libnvidia-fatbinaryloader.so lisusiwe ekusasazweni, ukusebenza kwalo kusasazwa phakathi kwamanye amathala eencwadi;
• Izixhobo zolawulo lwamandla anamandla ziye zandiswa ngokukwazi ukucima amandla ememori yevidiyo;
• I-VDPAU yongeza inkxaso yemigangatho yevidiyo ye-16-bit kunye nokukwazi ukukhawulezisa ukuchazwa kwe-HEVC 10/12-bit streams;
• Inkxaso eyongeziweyo yemowudi yokuLolwa koMfanekiso ye-OpenGL kunye nezicelo zeVulkan;
• Kususwe i IgnoreDisplayDevices X uqwalaselo lomncedisi ukhetho;
• Inkxaso eyongeziweyo PRIME Ungqamaniso yokunikezela ngenye i-GPU kwinkqubo usebenzisa umqhubi we-x86-video-amdgpu. Kuyenzeka ukusebenzisa izikrini eziqhagamshelwe kwi-NVIDIA GPU kwindima ethi "Reverse PRIME" ukubonisa iziphumo zenye i-GPU kwiinkqubo ezine-GPU ezininzi;
• Ngokungagqibekanga, ulungiselelo olunemisonto emininzi lwe-OpenGL luvaliwe ngenxa yokuhlehla kwezinye iimeko.

umthombo: opennet.ru