Digital_Fuzion

EVGATech_DanielM

Digital_Fuzion
I would recommend an online beta slow ring testing team for NVIDIA. Alpha, beta then a final build to the general user. Make a solid driver if you going to release and back it with WHQL. NVIDIA needs an insider program before dumping it out with issues to users. 

I agree. A lot of these updates with Microsoft and NVIDIA have been a nightmare for some users, me not personally as I use the Insider Previews so I knowwhat to expect and I do a bit of bug testing in my spare time and use the report feature, but NVIDIA should consider something like this and get the community involved. 

I found that Microsoft Windows 10 Pro 64-bit build 1803 is rock solid. Their counter parts with NVIDIA drivers are not....

Could not agree more. It really is a solid update and I am glad they finally added the option to be able to add custom colors to the layout. I honestly haven't ran in to any issues on my 1070 but I am always here so I haven't had too much time to test 

Seems people have their heads in the sand if they  think 1803 is release quality.
 
Issues it launched with:
VTD hang
default apps settings crash
Chrome / Cortana hang   [Fixed]
Incompatible with Intel 600/6000 SSD's
Systems with Toshiba SSD's suffer high battery drain.
Focus Assist causes gaming issues.
Surface Pro BSODS
Boot Loops
Won't install on Alienware products
usb Cambridge bluetooth adapters crash the  blthport.sys when 'turned off'
vsync broken
(Intentionally) broke Asus gsync  turbo buttons.

EyeDeeNo
I did run into a problem with Focus Assist but not with the few games I ran before switching it off.
Focus Assist was causing my screensaver Marine Aquarium 3 to open and close slowly. So for now Focus Assist is a no go but a nice idea though, nothing like being distracted by and covering the screen by a popup notification when in a game.
Yep Microsoft's Settings "app" a fundamental part of Windows SystemSettings.exe froze than closed while I was using it May 3rd. I was like huh okay ha-ha that just happen.
 
But outside of that all is running well.

Ha Ha. Yeah, my thoughts too. Almost the same effect too. Yep, them built in MS sure are crap..no,no, wait typo. Very nice and reliable!! Thanks again MS for your thoughtfulness. We know money is a distant second motivator for you!!

Sajin
My unigine superposition 8k score went up again when I installed these drivers. 

That's interesting. I think you have said before that small differences are meaningless in the real world - sorry if it was someone else. Anyway I fully agree with that, so would be interested to know roughly by
what percentage your improvements were?
 
IMO and experience an increase or decrease of say 1% or perphaps even 2% in a bench like firestrike doesn't mean much. I don't know the cut off point - never heard anyone say that, but I think a repeatable 
increase or decrease of 3% or more certainly warrants investigation, or if just increase, a beer!!
 
Roughly what kind of increase did you get Saijin (as in statistical vs. relevant in the real world PC usage)? Thanks.

Thanks Saijin. So it seems these drivers are indeed slightly better, validated by your repeat testing. Good stuff. If it will be noticable in games is a different matter.
Still a definite, re-tested increase even if only 1% is definitely a step in the right direction! Congrats!!

Looks like another swing and a miss from Nvidia
 
only fixes some issues, performance almost flat
 
thanks for the feedback to all the Brave Beta testers

still causing low gpu usage with SLI + Gsync :(

LOL! Nvidia still dragging and lagging to release some performance beta drivers to us all again.

squall-leonhart
Don't expect to get any more performance out of Pascal, its fully tapped.

Yep, I agree, and drivers won't/can't make any noticable increases.
Still IMO Pascal has been a good gen - if we forget ludicrous price gouging then a second round of ludicrous prices for bit miners. But back to your point, it's maxed out.
 
Maxwell wasn't bad either but the memory debacle with GTX 970 4GB really only 3.5GB full speed leaves a bad taste. My HOF (Galax Hall of fame) card GTX 980 was solid, but it was also the most expensive.
 
Personally didn't like much about Kepler, although many will disagree. My "GHz edition" Gigabyte 780ti never really was fully stable. Biased I guess.
 
My fondest memories are of Fermi however. Loved those cards. The GTX 580 was a true jewel for the time. (I had MSI Frozr 3 GTX 580 lovely.)
 
But, in terms of performance increase between gens, what you guys reckon?? I don't recall it being discussed before?? But the opinions of guys here would be really interesting.
 
Tough call, but I feel while Maxwell was good, Pascal totally overshadowed it. A large  jump (ave over 30% for equiv. card.).
 
It's partly subjective, and should be mostly objective - e.g Maxwell 980 vs Pascal 1080. It's something that could be argued for fun because there is more to it than just that. Still. I like Pascal!!
 
Errmmm sorry, I really did get off topic kind of, at the end there, but didn't want to start a new thread. Mods, please do as you wish. (which you always do anyway! Heh Heh. )

I can now say with confidence this driver is inflicted with a kernel mode access violation when using anything that reads from the graphics cards.
 
I can only suggest that if you need this particular driver, to not keep these utilities open.
 
...i will need to put aside some time to turn kernel dumps back on, the minidumps aren't going back far enough to find where the pagefault occurs.

Digital_Fuzion
I would love to see a developer group like the one from Omega Drivers which was an unofficial, third-party NVIDIA graphics driver team. They differed from the official drivers in that they offered more customization and extra features. To bad that's not the case with poor buggy drivers on aging 1080 hardware. A driver shouldn't be so bad at this point of the hardware being in the market place for sometime now. I have not had any issues other then dealing with NVIDIA driver problems.  

OMEGA, yeah that rings an old rusty bell from years ago....Maybe??
The only thing I know which is in some way related is NVidia Inspector which simply reveals, and allows us to alter a lot of paramaters. Paramaters that exist in the base driver but are hidden in NVCP. Not good.
Of course they also provide profiles for many games.
 
So no new code a such, just adds a lot of accessabiltiy, for extra functionality as long as your know what sparse grid  SS AA  and more new options.
Do OMEGA go one step further and actually add/take/change the driver code itself? Complex stuff if so!!
 
P.S. I used to use NVInspector a lot for that functionality and got some beautiful screens, but for some reason, it doesn't have such an impact with Pascal? I also have 1080ftw and 1080ti ftw3 in other PC. Could just simply be as you said " poor buggy drivers on aging 1080 hardware." Basically MS + NV = shoddy, premature software that causes headaches. I think they should include a free box of aspirin with every new W10 and New NVidia GPU. I didn't say GPU drivers, I know the issues, but I have had very very few GPU driver issues in my single GPU system.

Gawg36
Dam, that's not what I call acceptable. What % of NV owners know this? 2%?? Am I safe in saying that most users will never trigger this?

 
any application using nvapi will eventually trigger a bsod in the nvidia driver, presently as i don't have kernel dumps enabled i only have a symbol from the nt exception handler.
 
If you use multiple together it seems to halve the time it takes to hit the issue, but uptime is a factor.
 

Kernel access mode/violations I have to admit on kernel mode access violations I'm not up to speed. I understand Kernel of course. Next.....

 
Kernel Mode access violation  is the term used when a driver does something illegal with memory, like attempting to read from a region that was already freed, contains unexpected data or freeing data that already was.
 

....S!!!, I found some mini dumps but after running diagnostic 1 minute test SFC and DISM all was fine, PC running fine too, so I got rid of them. Game OK too. 

 
SFC and DISM are for file system integrity errors.
 

Thanks for sticking with this Squall-leonhart,  you are the only person or site I've heard about this from.

 
https://forums.guru3d.com/threads/geforce-397-64-whql-game-ready-download-discussion.420903/reply?quote=5548213
https://forums.guru3d.com/threads/geforce-397-64-whql-game-ready-download-discussion.420903/page-10#post-5548565
https://www.reddit.com/r/nvidia/comments/8i6baz/driver_39764_faqdiscussion_thread/dytg8dw/
https://www.reddit.com/r/nvidia/comments/8i6baz/driver_39764_faqdiscussion_thread/dyw52wq/
https://www.reddit.com/r/nvidia/comments/8i6baz/driver_39764_faqdiscussion_thread/dywgz0z/
https://www.reddit.com/r/nvidia/comments/8i6baz/driver_39764_faqdiscussion_thread/dz229l3/
https://forums.aida64.com/topic/4220-aida64-sensor-panel-freezes-the-machine
 https://forums.aida64.com/topic/4210-aida64-make-my-pc-bsod-kmode-execption-not-handled-asus-x99-deluxe-i7-5930k-gtx980ti-win10-b17134/
 

@Squall-leonhart: Thanks mate. Very useful answers. I've checked out the links too, they were clear and explanatory too. The amount of reference to this, the amount of people who have experienced it. The the overall ideas and conclusions all come to similar conclusion - Also exactly what you have said too. Seems it's by no means isolated with this latest driver the worst of the lot.
 
I don't want to say details too much about this particular issue, as I've just learnt' about it. (although I feel I have fundamental and useful understanding of it now, thanks to you.)
 
So, in summary, my opinion is that this is beyond the pale!! I feel pretty disgusted with NV releasing so called WHQL (I guess MS cut corners too with certifying?) with these drivers, and to lesser degree the few before it.
 
Perhaps 390.01 is better? That was mentioned somewhere. It's pre- 1803 however. Not sure if that really matters?
I feel it shouldn't but cannot be sure.
Bottom line: This is disgraceful. I may go back to 390.01 after a good search that it's OK. Or I will wait for the next driver. I mean this should be fixed by next driver right?!? NV sure know about it. The guy I mentioned earlier, Sora, said that usually they tidy up and streamline drivers before new architecture. Seeing as some Volta cards, low level ones, will be released soon, this may happen? That's what Sora says without directly saying it. Or maybe that's what he hopes. Makes sense either way.
 
One thing. I won't run Aida, HWinfo and the like until this big fail is fixed.
 
Thanks again Squall-leonhart. Your posted info and the links you provided have been excellent! 

Sajin
@Gawg36
 
squall-leonhart is sora.

Spoil sport, i was going for a surprise reveal. lol.
 
 I panicked for a moment because i have Process Explorer running which reads vram and gpu usage,  forgetting it was doing so via windows api's :p
 

Gawg36

Digital_Fuzion
I would love to see a developer group like the one from Omega Drivers which was an unofficial, third-party NVIDIA graphics driver team. They differed from the official drivers in that they offered more customization and extra features. To bad that's not the case with poor buggy drivers on aging 1080 hardware. A driver shouldn't be so bad at this point of the hardware being in the market place for sometime now. I have not had any issues other then dealing with NVIDIA driver problems.  

OMEGA, yeah that rings an old rusty bell from years ago....Maybe??
The only thing I know which is in some way related is NVidia Inspector which simply reveals, and allows us to alter a lot of paramaters. Paramaters that exist in the base driver but are hidden in NVCP. Not good.
Of course they also provide profiles for many games.
 
So no new code a such, just adds a lot of accessabiltiy, for extra functionality as long as your know what sparse grid  SS AA  and more new options.
Do OMEGA go one step further and actually add/take/change the driver code itself? Complex stuff if so!!
 
P.S. I used to use NVInspector a lot for that functionality and got some beautiful screens, but for some reason, it doesn't have such an impact with Pascal? I also have 1080ftw and 1080ti ftw3 in other PC. Could just simply be as you said " poor buggy drivers on aging 1080 hardware." Basically MS + NV = shoddy, premature software that causes headaches. I think they should include a free box of aspirin with every new W10 and New NVidia GPU. I didn't say GPU drivers, I know the issues, but I have had very very few GPU driver issues in my single GPU system.

Omega tweaked the drivers and optimized them for performance. This forum just has tech talk thinkers not developers. I think companies like EVGA should tweak their own drivers with Nvidia's source code.  Point made these drivers suck and EVGA can't do anything about it now.