CactusCat
Excellent points all. And indeed, Firefox is my default browser. And as of this post, still no black screen or lockup. Nothing more done than the 'prefer maximum performance' mode in the Nvidia control panel that was mentioned earlier. Since everything is plenty fast already, I went ahead and turned off the gpu acceleration settings in Firefox. I'll monitor and post back soon. And yes, the fact that EVGA tech reached out is awesome. I'll work with them. I've only needed EVGA tech one time in the past and they were amazing. Fixed my problem over the phone with one suggestion. Appreciate all the advice and replies.

If selecting the 'prefer maximum performance' option fixed the issue, then it is indeed a problem with the PSU being unable to handle the transient load of the GPU. A lot of modern PSUs suffer from this. Modern GPUs are just more affected by this.

ZoranC
It -can- be sign of a problem with PSU, but it is not necessarily 100% guaranteed sign it is PSU and not something else. For example, I am using 2080 Super with Corsair AX1600i yet I have experienced this problem.

If your 2080 Super uses the reference PCB and your AX1600i is a couple of years old you might be suffering from the same issue actually. The output capacitors on that PSU are quite crowded, the fan turns off at low loads, which means that those capacitors run really hot under the circumstances, massively reducing their lifespan from this and the constant variable current, variable frequency changes caused by the GPU suddenly drawing massively and the topology used by the PSU.

Just because something is expensive, has good numbers and is over-the-top does not mean that it is up to everything you throw at it.

Long story short, if the 12v output 16v caps on the PSU are mostly worn out, you are suffering from the same issue. RMA the PSU, it should be within warranty.

ZoranC
You are making massively stretched assumptions about my case just to avoid admitting your earlier statement doesn't work 100% of the time. Unfortunately for your argument at the time I was experiencing this my card and AX1600i were practically brand new and my card is configured to run fans 100% of the time, vast majority of the time its temperature is in the low 30s. Sorry.

I said "you might", not "you certainly", not "you definitely", not "you do". The exact details can only be guessed with a certain degree of accuracy remotely. Even for well known problems you might be wrong in the end.

One thing that is certain though, is that your PSU is a lemon and needs to be RMAed if you get a system shutdown/reboot consistently with power saving enabled on the GPU as the other user was getting.

ZoranC
Reread carefully what OP and I said. He explicitly said he isn't getting shutdowns/reboots but app/system becoming unresponsive. And I didn't say I was getting shutdowns/reboots but something -similar- to OP, which is Firefox stopping to play video while rest of the system kept working.

Indeed. In that case the GPU protections probably kicked in first.

@CactusCat
sorry for the delay in your post above this one, I had to "release it" from filter
 
 
FYI:
 
listing All background software (non-OS) that is running can help  Things like Corsair's iCUE software are a Known issue
 
add your Hardware & OS information to "your signature" helps & then your not needing to type it in every help post
 
https://forums.evga.com/editprofile.aspx
 
 

As stated, I'm here to report that the settings of "normal" on the Nvidia Control Panel and all performance turned off in Firefox, I just had a black screen/lockup. The computer did not shut off, there simply wasn't any video to see what was going on. There obviously was nothing I could do except press the power button. Doing that, the system then shut down instantly. There was no need for a long press of the power button like you'd do if the OS was totally locked up.  I was using Firefox and browsing Amazon for a bluetooth headset. Not playing any video, simply looking at some prices and reading a few reviews. So while I've had about 4 days of no problems, this just occurred and I'll continue to monitor. I've left the NCP at "Normal" for the moment, but I've enabled "Use recommended performance settings" in Firefox. We'll see how that goes. Will continue to report as issues occur.
CC

CactusCat
Another black screen/lockup this morning. Was doing nothing more than reading an email. This time, after rebooting, I've turned off all performance settings in Firefox. Firefox is my default browser and I'm pretty sure that every time a black screen lockup has occurred, Firefox has been running. There could be a correlation but will continue to monitor/test. I still have not underclocked the GPU so that might be an option very soon. I had no black screen lockups when the NCP was set to "prefer maximum performance" for 4 days. I'll probably turn that on the next time a black screen lockup occurs. Also, every time this has occurred, my system never shuts off. It simply goes black screen, Cntl-Alt-Del has no effect, nothing works whatsoever, but it's still running. A slight touch of the power button shuts the system off instantly (so its definitely a system crash) and then a quick press and it starts up normally.
CC

Can you try Chrome for a few days instead and see how it goes? unless it happens without using browsers i would give it a shot. I personally had problems in the distant past with Firefox that was affecting the PC in a way it shouldn't.

It's been nearly two weeks since my last post. Glad to report that no issues have occurred. The two setting that were changed in my previous post seems to have pretty much resolved the problem of random black screen lockups. The only downside I've found is a little bit of additional coil whine but that is acceptable as I expected it. I'll continue my testing and report back soon.

The release notes document for NVIDIA GeForce Game Ready Driver version 517.48 (released on September 27), under the "Fixed Issues in Version 517.48 WHQL" section, 
contains the following entry:
 
"On rare occasions, video playback in browser may result in bugcheck code: 0x116 [3508109]"
 
Myself, I've seen bugcheck 0x119 (VIDEO_SCHEDULER_INTERNAL_ERROR) under similar circumstances with a 3090 card that is *not* made by EVGA. Perhaps someone here, who experienced this issue previously, could try this new driver release (at default power management settings) and see if the issue is still reproducible?

Hi, not at all. Here are the details:
 
I am running a GeForce RTX 3090-based card on a system with an AMD Ryzen 9 5950X CPU, 128GB of RAM and a 1200W ASUS ROG Thor PSU. On some not-so-frequent occasions, when engaging in activity that involves video playback (this would most prominently be reproducible shortly after opening the Netflix application or shortly after accessing my cable operator's IPTV service through a browser), the image would appear to be frozen, followed by a black screen. Following this, the driver would usually recover, but the system would then become less responsive and exhibit a BSOD shortly thereafter. Having extensive software engineering background, I analyzed the minidumps obtained. The crash *always* happens in the same driver module (dxgmm2.sys) and in the same function (VidSchiSendToExecuteQueue()). The bugcheck is always the same, too – 0x119 (VIDEO_SCHEDULER_INTERNAL_ERROR). This would happen on both Windows 10 and Windows 11.
 
Said driver is made by Microsoft, and its name stands for DirectX Graphics MMS (Microsoft Media Server, a proprietary streaming protocol that serves to implement transfer of unicast data in Windows Media Services). In turn, that driver is interfacing with the installed display miniport driver (which is the kernel component of a graphics driver, in my case NVIDIA's). When said bugcheck happens, the display miniport driver appears to be faulting during the execution of the DXGKDDI_SUBMITCOMMAND callback function. Every time this would happen, without exception (no pun intended), the violation, the arguments and location thereof would always be the same. I could go into more technical detail if needed, but to summarize I can say that while I understand *what* happens, I cannot understand *why* – that is, without sufficient knowledge of both DirectX and the display miniport driver in question.

I did try changing the power management mode to "Prefer Maximum Performance" and it appears to help (it is hard to say conclusively though, since the BSOD, in itself, is very rare). While I realize that the bugcheck that I experienced is different from the one discussed here, the circumstances under which it manifests seem to be somewhat similar. I therefore hoped that this new driver release could help me, too – mostly at reducing my electricity bill by not having the run the card at maximum speed during low workload.



 

Hi,
 
No, I am not a fan of overclocking and the system runs strictly at manufacturer recommended settings, for all components. I've tested the RAM and could see no reliability issues.
 

Thanks a lot for sharing your findings with the rest of us, that was (and is) very helpful!

pavel.pru
Hi @ThomasAsmussen!

Do you happen to have any findings with regards to this new driver release?


Thanks!

Yes i do!
 
Ive tried two things since last time, swapped back in my 32gb kit of memory, turned them down from 3600mhz to 3200mhz, and then ontop of that installed the latest nvidia driver.
 
The results are fairly simple, no crashes since then. I dont know if its the ram change or the driver thats fixed things, but im leaning towards the driver currently, despite having a memory crash a few months back where i remember the error code was specifically pointing towards the ram(With that said, it could be the case that something with the nvidia driver made it look like the ram failed or something along those lines)
 
So yeah if you can, try the new driver, it sucks quite a bit performance wise honestly, but it makes things stable, im seeing more stutters and the likes, but im willing to live with that for now till a better performing driver with the fix comes around.
 
Do throw me a reply if theres anything youd want more details on.

pavel.pru
Thanks so much for writing back, @ThomasAsmussen!
 
I already have the latest driver installed, albeit the power management setting is set to "Prefer Maximum Performance". I see no perceptible difference performance-wise, however, between the previous driver release (version 516.94) and the current one.
 
Just to clarify, your system has been stable on the latest (version 517.48) driver release using the "Normal" power management setting, not "Prefer Maximum Performance" – whereas before, you had to use the "Prefer Maximum Performance" mode in order to achieve stability, right?

With regards to RAM, the most prominent scenarios that are indicative of a memory reliability issue are data corruption (for instance, where a game storefront client re-downloads game files after performing the post-download verification phase) and access violations (this typically manifests in user-mode application crashes [like a STATUS_ACCESS_VIOLATION error in Chrome] or bugchecks like PAGE_FAULT_IN_NONPAGED_AREA (0x50)). Moreover, where memory reliability issues are concerned, the location of the crash is mostly probabilistic and is variable between BSODs.

What I'm seeing here though, in my particular case, is a *somewhat* reproducible issue, occurring under a specific set of circumstances, and always in the same module and function. The fault would also be consistent (that is, the display miniport driver would be faulting during the execution of a specific callback function). The system *appears* to be stable on the "Prefer Maximum Performance" power management setting (although I cannot say for sure, as the BSOD is rare). This does not resemble a memory reliability issue in any way.

I had been affected by this issue for about one year and a half, before switching to the "prefer maximum performance" power management mode. I would, for instance, avoid using my IPTV service on my PC, out of concern for losing unsaved work due to a BSOD. Troubleshooting this problem had been extremely frustrating, but I guess I don't have to tell you that. :)

Yes you are correct, its stable on normal power management settings, before i could limit the amount of crashes/lockups by setting it to prefer maximum performance, but it wouldnt eliminate them entirely(once a day vs once a week, roughly)
 
As for ram, ive today turned them back up to their rated speed of 3600mhz to either lock the issue to my ram or the driver, im expecting to see its the driver that fixed things as the ram were fine paired with a 1080ti for over a year prior to the new gpu being introduced to the system.
 
Driver performance has been odd, it gives a stutter or two at first but i suspect thats due to the normal power profile, basically it stepping up its clock speed while loading in things causes these hitches/stutters. So its not terrible, but the fps i get does also seem to fly around a bit more. As an example lets say id normally run at 120fps in sea of thieves (Locked to that on purpose) before id see a drop here and there, nothing youd really notice, with this new driver its more common and the drops are bigger, 20-30 fps drop for a second or two, compared to maybe 5-10 before. Also on the desktop im seeing some weird behaviour with facebook videos and videos on reddit, not sure if they use the same player or something along those lines, they will artifact a bit, and force the page to refresh itself eventually, and if i leave it to do that over and over the page eventually becomes unreadable till i refresh manually, this happens very rarely, but enough to notice it. Also should be said this never happened before, not on the 3080ti or the 1080ti or on any other driver. I suppose it could also be a browser problem, everytime it happens i fire up my two backup browsers and try to replicate the issue in them, but with no luck, another problem thats hard to replicate.
 
But im hoping my driver concerns are fixed in the upcoming launch driver of the 4000 series.
 
Thanks for you feedback and i hope youll throw me another message if anything is unclear to you, or if you need more details. That issue is for sure the most frustrating thing ive ever encountered in my 15 years of pc building.
 
Ill update if the ram fails or not in about a week if nothing else happens here.

By chance are you using an EVGA PSU? I was having that issue with ECO Mode enabled. As soon as I turned that off it solved all my problems.