Advanced search

Message boards : Number crunching : Nvidia Driver 416.34 and above

Author Message
PappaLitto
Send message
Joined: 21 Mar 16
Posts: 504
Credit: 4,282,538,251
RAC: 372,854
Level
Arg
Scientific publications
watwatwat
Message 50944 - Posted: 21 Nov 2018 | 20:23:43 UTC

So I just came home to a frozen computer and after restart I found my 1080ti computing at 95C with the fan at 0%, completely off. AVOID THESE NEW DRIVERS LIKE THE PLAGUE.

[CSF] Aleksey Belkov
Avatar
Send message
Joined: 26 Dec 13
Posts: 40
Credit: 740,453,409
RAC: 787,188
Level
Lys
Scientific publications
watwatwatwatwatwatwatwatwat
Message 50946 - Posted: 21 Nov 2018 | 23:52:58 UTC - in response to Message 50944.
Last modified: 21 Nov 2018 | 23:55:42 UTC

Speed control of the fans is done by an embedded controller, and default fan speed configuration is located in the "BIOS" of video card.
For example, 10xx series graphics cards at GPU temperature <= 40 C stop the fans. This works without loading OS/drivers.
Therefore, your statement is highly doubtful. It is not necessary without any evidence to throw accusations and throw a tantrum.
The problem could arise for example, because of the appeal to the controller of fans of the video card in parallel from several programs(MSI Afterburner(can control fan speed curve), HWinfo, GPU-Z & etc) or because of some hardware failure.

PappaLitto
Send message
Joined: 21 Mar 16
Posts: 504
Credit: 4,282,538,251
RAC: 372,854
Level
Arg
Scientific publications
watwatwat
Message 50947 - Posted: 22 Nov 2018 | 0:40:48 UTC - in response to Message 50946.

The above is a warning not a "tantrum." I have had my entire windows operating system fail to boot due to this driver and was forced to completely wipe it. I can assure you it is not "some other software." After the new windows installation I used driver 399.07 with no problems and tried reinstalling 416.34, with the exact same issues.

Just because something hasn't happened to you doesn't mean it can't happen.

I think I have tacked down the cause of the event. The incompatibility of my previous windows build, 1803 and the new driver, 416.34. I have changed to "Semi-Annual Channel (Targeted)" rather than regular "Semi-Annual Channel" which updated me to 1809 which, at least so far, has solved the problem.

[CSF] Aleksey Belkov
Avatar
Send message
Joined: 26 Dec 13
Posts: 40
Credit: 740,453,409
RAC: 787,188
Level
Lys
Scientific publications
watwatwatwatwatwatwatwatwat
Message 50949 - Posted: 22 Nov 2018 | 8:08:13 UTC - in response to Message 50947.

You did not even bother to indicate that the problem occurred on a computer with Win 10, famous for unstable updates.
I worked enough time with the drivers 416.34 and 416.81 on Win 8.1 to say that there are no fundamental problems with these drivers.
Your same statement urged without any exception not to use the drivers 416.34 and above. Notice the difference now?
And yes, when you write LIKE THIS, and without any evidence - that it is exactly the tantrum.

Before you write something, you should:
1) Calm down
2) Think twice over the semantic content of their words.

PappaLitto
Send message
Joined: 21 Mar 16
Posts: 504
Credit: 4,282,538,251
RAC: 372,854
Level
Arg
Scientific publications
watwatwat
Message 50960 - Posted: 25 Nov 2018 | 16:35:11 UTC

Please think of this as a warning, something that people can chose to heed or ignore, you act like I am forcing people to not install this driver. You come off as extremely rude and should really think about the difference between a suggestion and treating people like children.

[CSF] Aleksey Belkov
Avatar
Send message
Joined: 26 Dec 13
Posts: 40
Credit: 740,453,409
RAC: 787,188
Level
Lys
Scientific publications
watwatwatwatwatwatwatwatwat
Message 50961 - Posted: 25 Nov 2018 | 21:29:12 UTC - in response to Message 50960.

Is that the way(" AVOID THESE NEW DRIVERS LIKE THE PLAGUE.") you're trying to "just" warn other people? Serious?
Your "warning" doesn't sound like that you're trying to explain it now.

Instead of admitting that your "warning" came out at least very "loud" and at the same time completely uninformative, you're just trying to justify yourself, and my attempts to point out that it was not worth writing your "warning" in such manner, you're trying to expose for rudeness.

It's ridiculous.

Jacob Klein
Send message
Joined: 11 Oct 08
Posts: 1111
Credit: 1,781,999,639
RAC: 1,130,985
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 50977 - Posted: 29 Nov 2018 | 4:23:10 UTC

I once had a problem where my custom GPU fan curve would get set to "off completely" despite tasks running the temps up to 100*C.

When that problem happened, it turned out to be a buggy version of eVGA PrecisionX, that didn't work right when suspending or shutting down and then loading Windows - It stored the very bottom of the curve as a constant setting when resuming from Fast Startup! Ugh. They fixed it eventually, but that was a nasty nasty bug!

I have since ditched that software, and use MSI Afterburner, and highly recommend it. I also disabled Fast Startup in Power Options, so my PC won't be susceptible to hibernation/resume problems.

I'd advise you to double check your fan monitoring software, as it seems possible that the problem you are having is with that software, instead of the NVIDIA drivers.

Make sure to report any results of any more testing you do! Thanks!

Kind regards,
Jacob Klein

PappaLitto
Send message
Joined: 21 Mar 16
Posts: 504
Credit: 4,282,538,251
RAC: 372,854
Level
Arg
Scientific publications
watwatwat
Message 50978 - Posted: 29 Nov 2018 | 12:06:42 UTC
Last modified: 29 Nov 2018 | 12:10:46 UTC

While I am also using MSI Afterburner, I have never had a problem with it in the past. The problem occurred directly after updating my drivers from 399.07 to 416.34, 416.81 and 416.94. All three new drivers had several issues. The 0% fan speed only occurred once, so I will call this a fluke.

The main problem is when the screen wakes up the primary screen (I have 3 monitors) it has scan-line errors. If I let the screens go back to sleep again and try to wake it up, it looks like it tried to wake up, but it stays black forever. I would then restart the computer and it would prevent booting entirely. The first time I installed this driver I tried to recover it with windows recovery and everything failed. I was forced to reinstall windows.

With a fresh install I tried driver 399.07 with nothing but success. Now I try 416.34 and above on the fresh install with the exact same errors, this time without MSI afterburner installed.

Jacob Klein
Send message
Joined: 11 Oct 08
Posts: 1111
Credit: 1,781,999,639
RAC: 1,130,985
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 50979 - Posted: 29 Nov 2018 | 12:37:54 UTC
Last modified: 29 Nov 2018 | 12:45:12 UTC

Have you been active in NVIDIA's GeForce User Forum, in the GeForce Driver Feedback threads?
https://forums.geforce.com/default/board/33/

There is a thread for each driver version, and I read them all, and yes I believe that I have heard of some people having similar problems with Release 415 drivers. But I haven't had any problems, personally.

Make sure to visit there to post your feedback, and also see if there are any workarounds, tips, or solutions. And if you want to report an issue more thoroughly so NVIDIA has a better chance of fixing it, then you should follow the steps in the first post of a driver feedback thread, to fill out the NVIDIA driver feedback form.

Profile Retvari Zoltan
Avatar
Send message
Joined: 20 Jan 09
Posts: 2036
Credit: 14,766,164,919
RAC: 2,105,858
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 50981 - Posted: 29 Nov 2018 | 18:05:36 UTC - in response to Message 50977.

... I also disabled Fast Startup in Power Options, so my PC won't be susceptible to hibernation/resume problems.
I recommend it on every system which runs from an SSD, as Fast Startup works by exiting apps during shutdown, then hibernating the "empty" OS. At the next power on it resumes the empty OS from hibernation. This method cause the monthly Windows updates to be delayed up to 30 days, it also reduces SSD lifetime (while it don't result in much faster bootup time on systems runs from an SSD).
On systems which run 24/7, it is advised to turn off hibernation completely as the unneeded (hidden+system) file named hiberfil.sys takes up large disk space (equal to the 60-100% of the RAM size).
You can do it by the following steps:
1. right click on start
2. click on Windows PowerShell as administrator
3. if UAC is on, you should click yes to enable PowerShell to act as administrator
4. type
powercfg -h off
and press <enter>
5. close the PowerShell window

There is no output of the command, but you can check the increase of the free disk space to verify the effect of the command.
If you want a more sophisticated check, you can type
dir \hiberfil.sys -ah
(and press <enter>) before and after step 4. to see if this file is really gone.
Turning off hibernation also turns off fast startup.

Profile Beyond
Avatar
Send message
Joined: 23 Nov 08
Posts: 1104
Credit: 6,101,732,079
RAC: 46
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 51191 - Posted: 6 Jan 2019 | 6:52:38 UTC - in response to Message 50981.
Last modified: 6 Jan 2019 | 6:53:53 UTC

type [pre]powercfg -h off
There is no output of the command, but you can check the increase of the free disk space to verify the effect of the command.
Turning off hibernation also turns off fast startup.

I agree, turning off hibernation is always a good idea.

PappaLitto, are you using an ASUS MB? I ask because there's a known issue between ASUS motherboards and monitoring programs such as HWiNFO that randomly causes the fans to go to zero. Bad, bad bug. AFAIK this issue only affects ASUS.

PappaLitto
Send message
Joined: 21 Mar 16
Posts: 504
Credit: 4,282,538,251
RAC: 372,854
Level
Arg
Scientific publications
watwatwat
Message 51198 - Posted: 7 Jan 2019 | 0:12:33 UTC

Unfortunately the solution is not that simple. I have an ASRock mobo and I am still restricted from using any 400 series driver.

Azmodes
Send message
Joined: 7 Jan 17
Posts: 10
Credit: 585,165,015
RAC: 1,308,793
Level
Lys
Scientific publications
wat
Message 51205 - Posted: 7 Jan 2019 | 11:57:25 UTC - in response to Message 51191.
Last modified: 7 Jan 2019 | 11:59:04 UTC

PappaLitto, are you using an ASUS MB? I ask because there's a known issue between ASUS motherboards and monitoring programs such as HWiNFO that randomly causes the fans to go to zero. Bad, bad bug. AFAIK this issue only affects ASUS.

Can confirm. When I got my Threadripper/ASUS combo back in 2017, I had to uninstall HWiNFO because the CPU fan kept shutting off for no apparent reason and since I sometimes wasn't around to notice, the computer would eventually end up going into a sort of sleep mode to prevent overheating. Not great. I tried a bunch of things; enforcing a constant fan speed, disabling the ASUS sensors in HWiNFO, etc., but nothing worked. So I removed HWiNFO and all was fine. I tried again a year later or so, but the exact same issue returned. Some conflict between that kind of software and the integrated monitoring systems. It's too bad, since that's a really handy program. None of my other (non-ASUS) hosts running it experience any such problem.

As for the driver, I'm using 416.94 on my main Windows 10 machine and haven't encountered any issues.

Jacob Klein
Send message
Joined: 11 Oct 08
Posts: 1111
Credit: 1,781,999,639
RAC: 1,130,985
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 51207 - Posted: 7 Jan 2019 | 13:34:37 UTC - in response to Message 51205.
Last modified: 7 Jan 2019 | 13:37:06 UTC

Did you try HWiNFO 6.00?
Also, might read this thread and reply there if problem persists:
https://www.hwinfo.com/forum/Thread-ASUS-Prime-X470-Pro-Fan-headers-shutting-down-after-a-while-when-running-HWiNFO?highlight=Asus

Profile Beyond
Avatar
Send message
Joined: 23 Nov 08
Posts: 1104
Credit: 6,101,732,079
RAC: 46
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 51210 - Posted: 7 Jan 2019 | 20:09:10 UTC - in response to Message 51207.

Did you try HWiNFO 6.00?
Also, might read this thread and reply there if problem persists:
https://www.hwinfo.com/forum/Thread-ASUS-Prime-X470-Pro-Fan-headers-shutting-down-after-a-while-when-running-HWiNFO?highlight=Asus

I was about to post that link but you beat me to it. In fact the CanoeBeyond post on the thread is mine. No answer yet, but it's been less than 2 days. There is indeed some attempt attempt to resolve this in v6.00, as when installing it on my ASUS (Prime X470 Pro) MB it asked if I wanted to disable certain MB sensors. No such message on any other MBs (MSI, Gigabyte, Foxconn, ASRock) and also no problem on any of them with the old version. ASUS really needs to fix this deadly bug. HWiNFO is pretty much required for CPU temp monitoring when using BoincTasks/TThrottle at this point and this bug can be triggered by ANY sensor monitoring software. In fact I quit buying ASUS years ago after running into longevity problems wth their MBs and GPUs (in fact posted about it years ago on this very forum). Thought I'd give them another chance, but now regret it. Probably will exchange this ASUS MB for an MSI unless there is a VERY quick solution. It's not clear on the HWiNFO thread whether the problem is completely fixed. If not there's no way I want to chance frying a machine due to ASUS bugs. Even if HWiNFO has solved this problem, it could also be triggered by any number of other programs. The good news is that TThrottle appears to have saved the machine when ALL the fans stopped. I discovered it a few hours later with the temps north of 75C and the CPU and 3 1060 GPUs throttled way back by TThrottle. Much thanks to Fred at eFMer, think I'll send him another contribution for his great work on BoincTasks and TThrottle. Here's a link to his great programs:

https://efmer.com/

BTW, also contacted ASUS. No response. Also ridiculous trying to get access to post on their forum.

I digress, wrote more than I intended...

Azmodes
Send message
Joined: 7 Jan 17
Posts: 10
Credit: 585,165,015
RAC: 1,308,793
Level
Lys
Scientific publications
wat
Message 51211 - Posted: 8 Jan 2019 | 0:14:10 UTC
Last modified: 8 Jan 2019 | 0:15:00 UTC

Oh yeah, BOINCTasks and TThrottle rule. I couldn't imagine BOINCing without them anymore. Unfortunately, the first time the ASUS shenanigans happened, I wasn't yet using them (and the second time a year later I was present to notice and take care of it immediately):

Thanks for the link, Jacob. That is interesting and no, I haven't tried it with 6 or above I don't think. However, from reading the thread, it seems like there's still no definitive fix? It also mentions disabling the ASUS sensors in HWiNFO, but I did exactly that and was in fact asked whether I wanted to disable them just as Beyond described. It did however not help in the end.

On this one computer I'm using CoreTemp and GPU-Z plus some Windows monitoring Gadgets instead.

[CSF] Aleksey Belkov
Avatar
Send message
Joined: 26 Dec 13
Posts: 40
Credit: 740,453,409
RAC: 787,188
Level
Lys
Scientific publications
watwatwatwatwatwatwatwatwat
Message 51216 - Posted: 8 Jan 2019 | 20:03:43 UTC - in response to Message 51207.

Did you try HWiNFO 6.00?
Also, might read this thread and reply there if problem persists:
https://www.hwinfo.com/forum/Thread-ASUS-Prime-X470-Pro-Fan-headers-shutting-down-after-a-while-when-running-HWiNFO?highlight=Asus

I have the same all-fans-stop-problem on ASUS PRIME X399-A. This issue usually occurs 4-10 days after the last restart of the computer.
Since using a computer without the risk of overheating and subsequent damage to its components is impossible without constant human control, to solve this problem, I purchased a pair of external fan controllers Noctua NF 1.
In this case, all fans are connected and powered from the controller itself, and the controllers are set to a static rotational speed(in %, all fans must be PWM controled). The motherboard is completely excluded from this scheme.
The solution is not perfect, but these controllers are inexpensive and I'm not worried about their reliability.
I want to believe that ASUS still listen to the owners of the problematic motherboards and release the necessary update of "BIOS" to solve this problem.

Profile Beyond
Avatar
Send message
Joined: 23 Nov 08
Posts: 1104
Credit: 6,101,732,079
RAC: 46
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 51358 - Posted: 19 Jan 2019 | 0:54:34 UTC - in response to Message 51207.

Did you try HWiNFO 6.00?
Also, might read this thread and reply there if problem persists:
https://www.hwinfo.com/forum/Thread-ASUS-Prime-X470-Pro-Fan-headers-shutting-down-after-a-while-when-running-HWiNFO?highlight=Asus

An update on the "all fans stopping" (including CPU fan) problem. Martin at HWiNFO answered my post in the above referenced forum, and I quote:

"This problem can be solved only by ASUS with a BIOS update. See this recent discussion for more details:
https://www.overclock.net/forum/21-benchmarking-software-discussion/1235672-official-hwinfo-32-64-thread-198.html#post27799512"

So it seems (sadly) that v6.00 does not solve the ASUS bug.

Post to thread

Message boards : Number crunching : Nvidia Driver 416.34 and above