GTX 660 Graphic Driver Issues - Looking for suggestions

I've been having some issues with FF14 and my new Geforce GTX 660. Pretty regularly I get graphic tears and "disco strobe effects" as the game starts drawing multicolored vectors randomly. Shortly after, my computer freezes and requires a hard boot.

Oddly enough, I have less problems with crashes in the open world. Dungeons, on the other hand, are almost certain to crash the game. Sastasha, Thousand Maws, and Hatali are particularly bad for it. Dungeon runs only last about 20 - 30 minutes before it is disco time. The open world is usually good for several hours.

FF14 is the only game I have problems with. All others (Left 4 Dead 2, Shogun 2, etc) run at maximum settings without a hitch.

I have tried Nvidia drivers 326.8, 320.49, and 326.41 without any luck.

I have played with all sorts of settings in the game. My current set up (windowless border, 60fps lock, 1280x, culling on, and standard settings is the most stable. Stable, that is, if you count dungeon crashes as normal.

Does anyone have suggestions?

Further info:

SYSTEM_OS Windows 7 Home Premium 64-bit (6.1, Build 7601) Service Pack 1
SYSTEM_CPU Intel(R) Core(TM)2 Quad CPU Q9450 @ 2.66GHz (4 CPUs), ~2.7GHz
SYSTEM_MEMORY 6142.539MB
SYSTEM_GRAPHICS_CHIP GeForce GTX 660
SYSTEM_GRAPHICS_MAKER NVIDIA
SYSTEM_GRAPHICS_VRAM 1990.313 MB
SYSTEM_GRAPHICS_SHARED_VRAM 2048.309 MB

I'm currently running 326.19 with no issues. It's a beta I believe. I have a 660 Ti, playing (when I can), Full screen, 1680x1050, at max settings. I really haven't had any issues with this driver or settings, however I've never been in a dungeon! I'd say look up issues with the drivers with the 660, as there have been problems in the past that have forced me to roll back drivers. In addition, make sure you do clean installs, I would think. That being said I don't know much, I'm more a trial and error to find things that work type of person!

I occasionally experience vector tearing while in game. I am hoping that once the DX11 support comes in the update, it helps address a lot of the GPU issues.
I am running a Radeon 7990.

This may sound super retarded but:

You do have the most current drives for this card right?
http://www.geforce.com/drivers/results/65047

It's a beta driver, but some times those beta versions work better then the "current live" version. Try that out and see if it helps.

Quote:

---

Originally Posted by Design

This may sound super retarded but:

You do have the most current drives for this card right?
http://www.geforce.com/drivers/results/65047

It's a beta driver, but some times those beta versions work better then the "current live" version. Try that out and see if it helps.

---

It's in the OP's list of drivers that have been tried.

I've got a different card, but I'm also having graphical issues in a similar manner.
Mind posting images of these glitches if you can get them?
I'm curious to see if what you're experiencing is similar to mine.

saw weird tearing on my PC due to Cascading Shadows---disabled them and it cleared up most of it. Could give that a try.

My laptop (670 series) has been fine with the older 307.07 drivers. Makes me wonder if XIV isn't going to fall prey to the same problems as XI where latest and greatest isn't always the way to go.

I will have to try 326.19. The Casting Shadows option also makes sense. If the 326.19 doesn't fix the problem, I will try that. And I will see about getting a photo of the vector tearing next it happens.

Almost sounds like an overclock gone wrong. If you keep your card overclocked, try disabling. If you don't have it overclocked but still have Afterburner running, turn it off.

It could also be a heat issue,keep an eye on your GPU temps next time you're in a dungeon.

Make sure your video card isn't overheating. Also if you've overclocked it, set it back to stock. Does the video get "sparkly" right before the game crashes?

I have a GTX660Ti, was getting some pretty intense tearing for a while until I capped my frame rate at 60fps. Occasionally I'll get a weird flash of color, but mostly I think that's been due to scenery. Also try turning on ambient occlusion if you haven't, really helped my performance out.

almost definitely sounds like an unstable overclock, overheating, or needing to do a clean install of drivers. Clean install is very important when going from driver to driver looking for better performance

Yeah, sounds like overheating GPU.
Had some serious glitches on my Laptop, and it was because GPU was overheating.

My gpu is at stock and it only gets up to 55 C when under a load (using Nvidia's System Monitor). I always clean install drivers. Learned that a long time ago! Tried installing 326.19 but still experienced crashes in dungeons. This time, I finally saw a blue screen! It seems nvlddmkm.sys is throwing the error (DRIVER_IRQL_NOT_LESS_OR_EQUAL). Locking the framerate to 60 fps has helped my stability but dungeons are still a problem. Again, the open world rarely crashes, which is odd considering there is more to draw.

Here are screens of what is happening:
http://media.japanpowered.com/images/vector-tearing.jpg

http://media.japanpowered.com/images/tearing-2.jpg

I thought it was a heat problem too, but no other game causes this problem.

I have the same issue playing only in windowed mode, no matter what graphic resolution. My fix was to play in fullscreen. It's a little more tedius, but it works (just a hard flash when alt tabbing). I had no issues in windowed mode with my gtx570. But with my 660 and my 760, both have tearing in windowed mode. Again, fullscreen fixed it without any issue, and actually I have smoother playback.


This was of course, during beta and early access only. As I have not been able to play since Aug 24 because of service account/customer support congestion :(

edit: after reviewing your screenshots, my tearing was not this significant, not by any means. 55c IS hot, but falls in the category "slightly above average" i would try to, if your expertise allows, remove the heatsink, apply some new thermal compound (preferably silver), and remount the heatsink/fan (you may even consider a new fan). I would try to push it to 45c. If that amount of tearing still persists, you are most likely looking at a developed problem due to gpu damage from sustained heat.

hmm.... wonder what kind of state the power supply is in too. Is it an 80+ certified model (may not have 80+ logo, but maybe a Xfire or SLI certified logo)?

Another thing you might want to look for is the power profile for the card. Forge texactly what and where it was located in the nVidia settings, but there is a section somewhere in there about Adaptive or Performance mode for it's power consumption. If it's on Adaptive, it might not be kicking in/out of high power/performance state properly.

irq not less then equal usually means ram related issue or theirs a problem with the north-bridge controller on your motherboard.

IRQL_NOT_LESS_OR_EQUAL error in Windows 8 is displayed in Blue Screen of Death (BSOD). Just like other Blue Screen of Death errors, IRQL_NOT_LESS_OR_EQUAL error also arises due to corrupt software application files, corrupt system files, incompatible or inappropriate device drivers, corrupt device drivers, faulty hardware devices, etc.

Typically IRQs are the interrupt requests that are sent to the microprocessor via south and north bridges. If IRQL_NOT_LESS_OR_EQUAL error occurs, chances are that an issue with the application or driver is related to the memory or interrupt request.

For Driver irql not less then equal it could be hardware overheat... It could be the card somehow got dislodged from the slot (reseating cards in the case sometimes fixes issues) I have personally had a video card some how come up out of the pci e slot just enough to throw windows into a shit fit.

It could be corrupt system files.. Or a in comparable program or a conflict between a virus scanner and the program itself..

you can also try a manged start-up and disable all the apps on start-up and try it if that works re enable apps tell you find one that causes it..

clean install of windows and reload can also help.

314.22's are stable too

Crazy artifacts! You may try using the videocard on another computer (if you have one). If you don't, try playing another game and see if it causes the same issue to make sure it's just not the game itself.

A more possible error could be your videocard VRAM itself that is defective.

Hope you find a solution!

55 is actually really cool under full load for a 660. Anything under 70 is considered fine, 70 - 80 is pushing it, and 80+ is too hot.
I play at1080p high quality on my 660 and it hovers around 65-70. never have any problems though.
It certainly looks like a classic heat problem. As said above could be faulty ram.

What are the rest of your system specs? (specifically motherboard, cpu and powersupply)
I see you mentioned playing LFD2 and shogun 2, are you able to try somemore graphically intensive games, max them out and see if you get the same problems. Something like one of the crysis games/battlefield 3 etc would do the trick.

Reseat your gfx card intoyour motherboard.
Try dial back your core and memory clocks on the card slightly.
Drivers you've obviously tried.
If possible run the game on a different system with your gfx swapped into it (alot of trouble I know)
Don't run in windowed.
Manually set your GPU fan to max before playing (75 I think)
Check your CPU temps
Disable all system monitoring software (precision x, afterburner etc) if you have any.

Good luck

Quote:

---

Originally Posted by JamieLee

irq not less then equal usually means ram related issue or theirs a problem with the north-bridge controller on your motherboard.

IRQL_NOT_LESS_OR_EQUAL error in Windows 8 is displayed in Blue Screen of Death (BSOD). Just like other Blue Screen of Death errors, IRQL_NOT_LESS_OR_EQUAL error also arises due to corrupt software application files, corrupt system files, incompatible or inappropriate device drivers, corrupt device drivers, faulty hardware devices, etc.

Typically IRQs are the interrupt requests that are sent to the microprocessor via south and north bridges. If IRQL_NOT_LESS_OR_EQUAL error occurs, chances are that an issue with the application or driver is related to the memory or interrupt request.

For Driver irql not less then equal it could be hardware overheat... It could be the card somehow got dislodged from the slot (reseating cards in the case sometimes fixes issues) I have personally had a video card some how come up out of the pci e slot just enough to throw windows into a shit fit.

It could be corrupt system files.. Or a in comparable program or a conflict between a virus scanner and the program itself..

you can also try a manged start-up and disable all the apps on start-up and try it if that works re enable apps tell you find one that causes it..

clean install of windows and reload can also help.

---

Actually, it's more an issue with invalid memory access. Typically, it's caused by a driver or program trying to access memory already in use by something else. But, I have seen this arise erroneously within the swap file because of corruption. As a sort of quick and easy test/fix for that, you can change the advance system settings to run without a swap file (remember to SET the change, then apply), reboot so that it deletes the swapfile, and then go back and turn the swapfile on again so that it builds a new one. It can sometimes be caused by RAM that isn't seated well or has dirty contacts also, but that should actually trip a different error. But, it's worth it to re-seat them just in case it is that rare occasion (especially if your area gets a lot f humidity...oxidation is a pain and it has to be done routinely in some areas).

Otherwise, it is often an issue of drivers not getting along. The trick is finding out which driver is causing the problem. You may see errors in your logs (eventvwr.msc) pointing to issues with a particular device. That's a good place to start. But, more often than not I've found it to be tied into what once was commonly installed as "chipset drivers" (drive controllers, usb controllers, etc.) not getting along with what was commonly drop-in devices (vid/sound and network controllers). Make sure your critical Windows Updates are up to date, then get updated motherboard chipset drivers from your manufacturer--it's usually a single package that installs quickly, knocking out a bunch of devices in one fell swoop. I once had a laptop that was getting knackered up by the Trackpoint software... had to actually roll back to an older driver to fix that one. IBM's power management software for laptops would get flaky sometimes too. So, if you see some feature specific updates (or previous releases) available, might want to grab them too, just in case you need to test them.

After that, start disabling devices you don't use. First in the BIOS, then in Device Manager. Some obvious things are the old COM, Parallel and IrDA interfaces (most everything is USB nowadays), but also try things like cameras and such too that you won't need when playing the game. This will significantly narrow down your list of suspects. One common peripheral culprit (unfortunately) are the HID's. Usually plugged into the USB port. So, if your controller or something is installed as a Human Interface Device (HID), I would look into updating those drivers if they are vendor specific (like Logitech or APC) before updating Video and Sound (changing Audio/Video can sometimes lead to other issues). One common product that has been at the source of issues for me over the years are the HDMI crossbridges. It's sometimes a tricky one to find a specific update for though. Sometimes ATI/nVidia will release updates specifically for the HDMI Audio interface, or just updated WDM descriptors for Windows. These simple updates have sometimes fixed issues, especially with ATI's.

After that, it's time to look into things that load in the background that are basically idle and *shouldn't* be impacting the system when their related devices/software is not in use. iTunes and all the associated Apple services (Bonjour, Apple device detection services, etc.)---yep, might have to update that i-product's software again, or maybe set the services to Manual so they don't auto load. Cyberlink products (PowerDVD, CD/DVD burning Software). Security Software... any and all of it... make sure it's current. Sometimes even the stupid stuff you'd never think about at all like Adobe, Java, or Quicktime can be at fault (they all load quick start programs that sit in memory, waiting to be called upon).

Yeah... I know... it's a long laundry list. There's even more examples that can be put on the list...just trying to give hints on what to look for. Basically, anything running in the background could be causing a conflict. But, I've found by the time I've dumped the swapfile, reseated RAM, and updated the important drivers... the problem typically disappears.

win7 64bit
i5 2500k
gtx 660 ti running 320.49
8gigs corsair ddr4

no problems

I have a GTX 660 too and I was actually having problems with the 314 driver while playing. The driver would stop responding and recover frequently so I went ahead and did a clean install of the 326.80 driver and its been running fine so far.

Another update:

I reseated my RAM and the GTX. Cleaned up various old programs and killed some services that snuck up on me. Darn Apple software anyway. - Crash

Turned off and cleared the Swap file. Restarted. Turned it back on. Restarted - Crash.

I downloaded the Crysis demo and cranked the settings to max. Ran just fine with multiple play throughs. I am generally not a FPS guy, but the game is pretty. I like the vegetation. Rebooted and loaded up FF14. Played for about 20 minutes and crashed. Now I am trapped in 3102. However, it seems isolated to FF14. Nothing else causes any problems. Windows runs all day without a hitch; all games I've ran so far work fine at max settings for hours a time. This has me stumped.

Next on the agenda, uninstalling FF14 and re-installing it.

I have to say, I am getting a little frustrated because the log in issue prevents me from properly testing.

Quote:

---

Originally Posted by Segraine

Another update:

I reseated my RAM and the GTX. Cleaned up various old programs and killed some services that snuck up on me. Darn Apple software anyway. - Crash

Turned off and cleared the Swap file. Restarted. Turned it back on. Restarted - Crash.

I downloaded the Crysis demo and cranked the settings to max. Ran just fine with multiple play throughs. I am generally not a FPS guy, but the game is pretty. I like the vegetation. Rebooted and loaded up FF14. Played for about 20 minutes and crashed. Now I am trapped in 3102. However, it seems isolated to FF14. Nothing else causes any problems. Windows runs all day without a hitch; all games I've ran so far work fine at max settings for hours a time. This has me stumped.

Next on the agenda, uninstalling FF14 and re-installing it.

I have to say, I am getting a little frustrated because the log in issue prevents me from properly testing.

---

I was doing some googling this morning and it seems there are a lot of other people out there with 600-700 series cards with problems, myself included. I was fine during beta 3 and 4 and even early access. For me, the crashes started off seldom at first but have become frequent, and I can easily cause a crash in under 2 minutes just by standing in Limsa looking out to sea. I can also reproduce it while streaming video on my second monitor, something that has never been a problem with any other game. My entire system freezes up and the screen either goes white or display output stops completely. It seems unlikely to be anything other than a GPU problem, and one I can't reproduce in any other game or stress test I've tried.

I've already tried various drivers, but I have a couple more to try as well as a couple other things (turn off Afterburner, unplug second monitor, etc.). I uninstalled and started downloading the game again earlier this morning, but I won't know if that worked until I get home later today. I'll report back if I can get anything to work, but I'm not hopeful. There seems to be a driver problem, but I'm not going to try anything older than 314.22.

try turning off CUDA

I was having issues with lag in the game but realized it was not my video card. Turn off anything like malewarebytes or advanced system care that run auto checks daily. Sometimes when that kicks in, if you have it scheduled, it won't warn you but will use system resources to do its job. Only thing I leave running while playing is Comodo. The same can be said for antivirus scans scheduled during your play time. I changed all my settings to run when I'm asleep even if I stay up later than normal.

did you turn your frame limiter off?


i had similar issues with SLI 670s because i turned frame limited / vsync off and when i would get 120+ fps on the login screen or dungeons the visuals would go fuckin crazy

@Dremic

I haven't turned off the frame limiter. Last time I tried it, the game was very unstable and crashed more often. I would get 300+ fps on the login screen.

Quote:

---

Originally Posted by Segraine

@Dremic

I haven't turned off the frame limiter. Last time I tried it, the game was very unstable and crashed more often. I would get 300+ fps on the login screen.

---

turn off the frame limiter in the game, force adaptive vsync and triple buffering through nvidia control panel.

also make sure you have power managment set to prefer maximum performance over adaptive.

Time for another update:

I reinstalled the game and it (mostly) seemed to fix the problem. I haven't been in a dungeon yet so the verdict is still out. However, it is stable enough that I swapped the graphics settings back to Desktop (High) presets without any issues. I have only seen two crashes since I reinstalled.

Quote:

---

Originally Posted by Dremic

turn off the frame limiter in the game, force adaptive vsync and triple buffering through nvidia control panel.

also make sure you have power managment set to prefer maximum performance over adaptive.

---

I was advised against this when I spoke to Nvidia.

Don't feel bad. I was getting tearing with a EVGA GTX 780 Classified. . . It would bench Heaven and play BF3 like a boss, but FFXIV, nope. I had to lower the memory freq back to 6k.

Quote:

---

Originally Posted by animapain

I was advised against this when I spoke to Nvidia.

---

what was their reason?

sounds to me your card is making artifacts. run a benchmart tool like furmark to see if it does the same thing. try using cpuid hwmonitor to see if your card is overheating. what is the brand on the video card? if it is evga then they have a better tool for testing artifacts with on there site.

Quote:

---

Originally Posted by dstyle

what was their reason?

---

They just said it's not advisable unless absolutely necessary. I tried to ask why but they were pretty vague about it.

Okay, reinstalling has helped a little. I am still having problems time to time but not as often. After running a few temp checks the hottest the card gets is 62 C, but it crashes independently of the temperature. It will crash as readily at 50 C as at 62 C. My processor is barely above the minimum recommendations (Intel Core 2 Quad) so I suspect that isn't helping matters. My rig is about 5 years old; I am considering building a newer updated version this year. I tend to build one every 5 years or so.

Removing the frame limiter made the game terribly unstable.

EVGA's Artifact scanner had no problems. However, the core is blistering at over 70 C during the tests. I need to get some better air flow going it seems!

I will monitor the card for a while using EVGA's tool and tweak the fan speeds to see if it helps. It is odd that Crysis and other games I have tried do not produce the same problems.

removing he frame limiter for me works fine...haven't crashed (usually do at around 15-30 minutes of gameplay). but have yet to play longer than 1.5 hrs so far