Discussion in 'Videocards - NVIDIA GeForce Drivers Section' started by valorex, Oct 12, 2021.
Alright good to know,thanks !
REAL good drivers, FC6 was at 62 min 68 avg @Ultra, FX CAS ON, 4k (1440p x1.5) DXR reflection + shadow ON + hd textures + hdr ON. Now im getting 65 min 74 avg.....same settings. unbelieveable...
You truly believe that NVidia is compiling two entirely different sets of binaries for the same driver?
this driver introduces randam crashes on cod black ops cold war, reverting back to previous driver resolves the issue.
CPU - 10980XE
Yours is the only complaint I saw about it.
I play a bunch of competitive games except Valorant and don't feel any difference between the two drivers.
Using these latest drivers and loading my latest Far Cry 6 save gives me 45FPS instead of 125FPS, 3080TI, Windows 11
The people who have a clue are always called crazy by the folks that never got one.
There is absolutely nothing done differently with the kernel driver in DCH packages vs Standard packages, your experience is not related to the way the driver is packaged, all the same services and libraries are installed (minus anselconfig exe) just run from different locations.
Only for the compiled script and nvi/nvu files. everything else is just a batch file difference during the package build.
every instance of this i've followed up with on reddit or otherwise has turned out to be system instability that the driver is exposing, give the cpu some more juice or return memory to jdec profile.
Yeah this driver crashes COD Black ops for me too.Back to 472.12.
Its driver related as alot of ppl are experience it, the error is dx12
there is no problem whith my system as this happens on another system whith diferent specs (10920x rtx 3070)
never had a problem before on this game and Roling Back the driver fixes the problem.
Drivers are not perfect and this one has hits and misses, iam just reporting here like others are doing in cod forums reddit etc
There is absolutely zero difference between them. This is all in your imagination.
What i am seeing with black ops and this driver is a lot of reports of stuttering - which is same old same old following a driver update.
Its system related, because on stock systems the issue doesn't present.
and nvidia doesn't provide support for the operation of the driver (or lack of) when overclocks (or Undervolts) are involved on any aspect of the system - XMP included.
Theres a reason why Custhelp requires you to try the game with debug mode checked before continuing with the lodgement of a driver, and i feel bad for the custhelp team every time they have to deal with a clueless bro gamer that didn't put any detailed research into their build before slapping it together and turning on the xmp memory overclocking features.
Because of the nature of what a 0xC00005 error actually is, if the software was the direct cause everyone would experience it when the same software mix is replicated because software side memory violations are not especially random in nature, you would hit the same crash in the same function, at the same point of the game every single start..., and your buddy next door on the same driver with the same graphics card would experience it too even on his 10 year older cpu, excluding some rare cases where a windows kernel driver has corrupted the stack way back at boot. (Case in point, the Access violations in Batman AK on pascal till recently, the functions were the same every time, minus the memory addresses which will never be the same from start to start on windows).
On the other hand, since the case here as usual involves the stability of the data between cpu and system memory, the timing of the illegal memory op will be random as heck, have no discernibe trigger, occur more or less with variable loads and could even sporadically take down the windows kernel itself in the process.
Every time someone vomits out "my system is fine" hurr durr, I find myself questioning what they are doing on the one Tech forum on the internet where the users are supposed to have the experience to know better. This is the one community on the internet where i expect users to remove their fricking undervolt, or return to stock.
There is not, and never has been certainty when it comes to Computer electronics and their behavior when run out of spec, your system is only TESTED stable, until it comes up against the first case where its not - Hell, this is the case even if you're running completely stock, because silicon defects are not the rarity they are made out to be.
DDR can literally pass a week worth of Memtext, started from a cold boot, but stop and start the test again and the test spits out errors because of gate leakage.
People like to come out with "well, the driver has to support blah blah blah", no, the driver only has to support the components that nvidia reference tested them for, when the system is running at the reference specification that the cpu vendor confirmed as supported with the memory speed that they advertised it was rated for.
You the computer owner are responsible for maintaining the stability of your out of spec PC, not nvidia, nor AMD, not Intel - if that means staying with old drivers because your hardware can't run newer ones. Too bad, so sad.
When you want to provide feedback for something, you have to perform the due dilience and own your responsibility for the device enough to do it properly, thoroughly and not waste customer supports time.
Ryzen with ram at 3600mhz? not stock, go back to 3200 and test again.
Nvidia's Video drivers have been exposing poorly tuned out of spec system configs since Windows Vista, i had to go back and retweak my nahalem after the multithreaded optimisations hit in R347 and that was on a very solid W7SP1 - There was the same people for 2 years after crying about crashes not being fixed because those crashes were in their hardware.
This driver is randomly crashing for me atleast in Quake Live which is an opengl title.
Running win10 with latest updates and I have everything on stock with 6850k and measly 2133mhz ddr4.
GPU is msi gaming x 1080ti with the small default factory oc.
No idea if that could cause the crashing..?
Anyone know if there is compatible stock 1080ti firmware I can safely flash on this GPU?
100mhz is indeed slight, but all is possible.
the driver or the game? is there any visual issues when it happens?
it keeps shifting down to low power 3d clocks, you might get some stability by locking the quake live profile to prefer max.
image freezes and sounds start looping. it recovers when i alt-tab out and back. i have set prefer max perf in the game profile
ITS the driver as far as black ops goes. Game runs perfect with 472.12 drivers. Install these drivers and get 5 cashes a day.Ive been doing this for over 20 years.
TBH you can do stuff for 100 yrs and not know what youre talking about, time spent at doing something is not a proof of proficiency. I know people who are in programming for over 20 yrs and they still work at junior lvl because they never really applied themselves over that. Go 4 posts up and read what Astyanax wrote.
I personally never got over 1700 mmr rating in HoN which i play on and off for 10 years, by your logic i should be capping the mmr by now, stuff doesnt work that way.
You can do this for 100 years and still learn new things every day, and hey, i wouldn't be surprised if theres an access violation specific to this driver that occurs because you have the wallpaper change periodically by itself (hell wallpaper engine was the apparent cause of the nvcp crashing recently).
anyway, all im saying is check everything before reporting because atleast then you know you've ticked off the checklist, and you'll also find that in doing so you can make note of things you turn on and settings of your hardware that the vast majority of folks entirely omit.
That sounds awefully like an issue that several emulator authors have encountered, and also reportedly affecting DooM BFG, and even being on reportedly affected hardware i cannot replicate it with a clean driver install and having purged the GL_Caches, there may be a factor that i'm not meeting.
DirectX 12 is an api not an error and there are many different forms a D3D12 error can take, the API is fragile as all heck and 90% of crashes with it are a result of a client fault that an overlay tool provokes or the driver has unwittingly stepped into by trying to behave more accurately to Reference/WARP but the game developer was playing loose and fast with memory "safety"
post an error code specific to the crash, or the dump file.
there are still people expecting drivers to fix Fifa d3d12 instead of revising their hardware or user environment.
I think you're going to discover its the game, the new season introduced crashes even before the new driver came out, unless you undervolt and believe drivers are required to work fine with you starving the gpu (>.>, yeah, those people exist)
like a told you its dx error (its basic the error every on is talking) since the game released i have it never had a single problem until this driver, have no problem with other games even instaled other games that ppl are saying this drivers has crashes no problems here, the only one That crashes sometimes is cod bo cw.
Its on cod bo cw not fifa, but its already been recognized.
but yes the combination of the this driver and the latest patch can be the problem, thing os on the previous driver nobody has the problem…
Well, in my case the crashes were caused by newest beta version of Afterburner.
EDIT: Nope, they're back. Was fine yesterday, not fine today. There is something wrong with these drivers.
Nope, it's always everything other not driver, the local paid monkey astyanax said lol
I'm just curious how many dozens of people must write about broken driver and problems with different systems to shut up craps like astyanax lol