Freeze Screen Issues

[5xad0w]

I am having the same issue, though I tend to only freeze up/have a black screen for 5-10 seconds at most.

From checking around the internet this seems to be a problem affecting Nvidida cards and is believed to be the driver crashing and recovering.

Has anyone had this happen on an AMD card?

For reference I have a Gigabyte GTX 780 GHZ edition in a rig that is less than 1 month old. (4770K based, mild OC to 4.0GHz)

The max temp my card hits (using OHM to see peak temps) is right at 50c. So overheating is not an issue.

I can play Dayz, a game that pushes the temps pretty high, for over 5 hours with no problem. I've also ran Valley in a loop for just over 2 hours and used OCCT to error check my card. CPU and mem are both fine as well as I've ran a full memtest (4 passes) and ran Prime95 over night to test the CPU stability.

Even ran Prime and Valley together for an hour to check temps with both under load with zero issues.

PSU is new (80+ gold and well over wattage and 12v rail requirements for system/GPU) and my mobo's BIOS is flashed up to the current version.

Drivers are current with the previous drivers having been scrubbed in safe mode before installing the latest. I also tested the beta (then scrubbed and re-installed the current) but I seemed to get freezes/driver crashes more. (3 in less than 15 mins vs. 1 maybe every hour or so in non-beta)

I play on Maximum, but have gotten the freeze/blackscreen even with lowered settings.

This is the only game I have any sort of issue with.

[Marishi-Ten]

I am having the same issue, though I tend to only freeze up/have a black screen for 5-10 seconds at most.

From checking around the internet this seems to be a problem affecting Nvidida cards and is believed to be the driver crashing and recovering.

Has anyone had this happen on an AMD card?

For reference I have a Gigabyte GTX 780 GHZ edition in a rig that is less than 1 month old. (4770K based, mild OC to 4.0GHz)

The max temp my card hits (using OHM to see peak temps) is right at 50c. So overheating is not an issue.

I can play Dayz, a game that pushes the temps pretty high, for over 5 hours with no problem. I've also ran Valley in a loop for just over 2 hours and used OCCT to error check my card. CPU and mem are both fine as well as I've ran a full memtest (4 passes) and ran Prime95 over night to test the CPU stability.

Even ran Prime and Valley together for an hour to check temps with both under load with zero issues.

PSU is new (80+ gold and well over wattage and 12v rail requirements for system/GPU) and my mobo's BIOS is flashed up to the current version.

Drivers are current with the previous drivers having been scrubbed in safe mode before installing the latest. I also tested the beta (then scrubbed and re-installed the current) but I seemed to get freezes/driver crashes more. (3 in less than 15 mins vs. 1 maybe every hour or so in non-beta)

I play on Maximum, but have gotten the freeze/blackscreen even with lowered settings.

This is the only game I have any sort of issue with.

Do the panic logs show any specific .dll/.exe causing the crash? With a full driver crash, your monitor will flicker out while the driver attempts to reboot and reload the card. It must not be getting that bad for your GPU.

Pull the panic logs and try to find a commonality in them. Ideally, all the separate instances will point to a specific .dll/.exe/.rb/.gz/.sh/.plist/etc. causing the crash.

[5xad0w]

^^

Well, first off, I did a clean install of Win 8 to rule out any other program conflicts causing this. Namely any rogue driver files floating around. (i've had both AMD and Nvidia drivers on this rig as well as remote access apps that install their own display drivers)

With Win 8 fully patched up (but not updated to 8.1) and nothing else aside from Nvidia's latest drivers and FFXIV I still got the "goes black with sound, then comes back about 5 seconds later" issue 4 times in 2 play sessions.

In event viewer it's tagged as Display driver nvlddmkm stopped responding and has successfully recovered. which certainly fits the profile of what I and some others are seeing. At least those of us who aren't straight up crashing or hard locking.

I've also tried (before the fresh install) running the game in windowed mode and borderless windowed mode instead of fullscreen. No change.

As well I tried a registry fix that extends the time before windows tries to recover the video driver (in case of a false positive from the card switching power states) but this resulted in DX errors straight to desktop instead.

Going to follow the nvlddmkm line a bit and see if I find anything.

[5xad0w]

Ok. I've been going at this for a while now and MAY have found a fix. I want to test it a couple times before I jump to conclusions.

Firstly, I found a 100% way for me to reproduce the error by watching a certain movie from the Inn Room.

"Way Down the Hole" Cutscene #3 will always result in the "Display driver nvlddmkm stopped responding and has successfully recovered." error for me at the end of the scene as it transitions back into the room. It may work with other cutscenes, but I wanted to stick with what I knew would work 100%.

I've tried various fixes, even ones that seemed absurd or highly unlikely, and have always had the error crop up every single time. Of note was a PCI-E state fix from power management that actually resulted in heavy artifacting with my card. Had to stop testing to run OCCT for 2 hours and then run a burn test with FurMark to make sure I didn't damage my card. (others have reported artifacts along with lock ups, so it still seems to be driver based)

Anyway, I've since ran the cutscene I mentioned twice in a row, with a reboot between each, with no problem. This is the first time, including when it originally played in game, that it has not triggered an error.

This is the (mind numbingly simply) fix for it:

http://www.tinkertry.com/display-dri...ed-responding/

It does seem that the core problem is windows thinking that a change in the power state of the GPU is actually the card becoming unresponsive. The end result being that windows creates a very real error (in the form of prematurely resetting the graphics driver) in response to a problem that doesn't exist.

I'm going to go run the cutscene a couple more time then go out and actually run around the world and see if I can stay error free for at least an hour but wanted to post what has been the most promising fix so far for others to try.

[Gummy]

I had those problems when I had a program called remote admin installed. It had installed a 'Mirror Driver' under display in order to operate remotely. As soon as I removed that driver I was able to play 100% fine no driver crashes etc. Obviously I wasn't able to remotely control the PC after that but it was worth it, lol.

[Kestreel]

I never get the black screen, my screen just freeze for say a few a minute or two. Mouse pointer still works and ppl I play with tells me im getting the red bar of DCing. After they stop seeing the red bar on my toon , my screen would resume again.( im still on teamspeak talking while this is happening). I wish I can pin point the issue or know if its there serever side causing the lag/freezing so I can just stop trying to trouble shot this.

[5xad0w]

I never get the black screen, my screen just freeze for say a few a minute or two. Mouse pointer still works and ppl I play with tells me im getting the red bar of DCing. After they stop seeing the red bar on my toon , my screen would resume again.( im still on teamspeak talking while this is happening). I wish I can pin point the issue or know if its there serever side causing the lag/freezing so I can just stop trying to trouble shot this.

What is your GPU?

I have had a couple freezes where my screen didn't turn black and I still heard sound and had control of my mouse. Though never for more than 10 seconds I would say. It also does't black out when I would get the crash and restore error watching a cutscene, it would just visibly freeze for a few seconds. Almost always near the end of the cutscene as it tried to go back to the game world.

You may want to try to TDR registry fix I posted.

After setting the timeout to 10 I finally got to play for about an hour and a half with no errors. Even went and checked event viewer just to make sure there where no other reported problems and all was good.

[Kestreel]

The rig Im having problems runs window 8.1 ,i7 860,8g rams and a pny gtx 670 water cooled.No heat issue of any kind. Stess test out the a$$ b/c of this issue. No software/program other then FFXIV and gpu drivers. ( I have try all beta and old drivers).The pc I have a work is smiliar just not water cooled and runs a gtx 570. Funny enough the pc at work has ZERO problems,runs the game 8 hours a day and nothing. both using save drivers.

[5xad0w]

As I said before, i'd try the registry fix.

If that doesn't work, check even viewer and see if there are any error or warning messages related to display.

Also, does this only happen on FFXIV?

Nvlddmkm errors seem to be the source of woe for many people across a narrow selection of games.

[5xad0w]

Welp, after a couple hours of playing I got a DX error (100000d or something to that affect) using just the RegEdit fix I posted.

Locked up and crashed out to the desktop with the DX message. Was hung for maybe 5-6 seconds before I got dumped out.

I have since got it to replicate (though not 100%) using cutscene transitions again.

I've noticed on a great many of the error threads floating around regarding Nvidia cards the vast amount of people who report similar issues are on Win 8 or (much more often) Win 8.1.

Going to clone my current HDD to a back up and install a copy of Win 7 and patch it up and try FFXIV and see if I get similar errors.

TDR started with Vista, so 7 has it as well, but I'm going to start with it set to default and no preventive measures applied and see what I get.

I kind of hope reverting my OS isn't the only reliable fix since others won't have previous copies of Windows lying around. :/

On a side note, I did find a thread where Nvidia acknowledged the error with TDR and their driver, but it seems to have died off in terms of their input.

Some people have said the problem disappears on older drivers, but which driver version exactly I haven't gotten a consensus on.