Random Crashes due to OpenGL
By
Addicted,
Posted on: 2010-11-16 18:17 in ALL ACCESS Gameplay Technical Issues Bugs
Much like others in this thread, I was getting the NVidia "driver has encountered an error and will recover" dialog box when I was running with an Nvidia GTX 285. Nothing I did seemed to make any difference... until I swapped video cards.
Yes, it sucks. And it's expensive. But it completely solved the issue for me. Now I'm running an NVidia GTX 480 with the 260.89 drivers and I haven't had a single crash on the same machine that was running the 285 video card. I'm even overclocking my CPU.
I don't know if there are stability issues with the 2xx chipsets or the older card had hardware issues.
No MaelwysAlts, you still multiple by 2 with DDR3 memory to get it's rated speed. You are thinking about maximum memory bandwidth with triple channel memory.
Actually memory seems to be running a tad slow, more like DDR3-1066 than either 1333 or 1600. I don't believe idle CPU multiplier reduction affects reported memory speed, notice your CPU in the first screenshot is only running with a 12x clock multiplier instead of 20x or more if the CPU is busy.
Father Xmas - Level 50 Ice/Ice Tanker - Victory
$725 and $1350 parts lists --- My guide to computer components
Tempus unum hominem manet
Quote:
You lost me here.|
Originally Posted by Father Xmas
Actually memory seems to be running a tad slow, more like DDR3-1066 than either 1333 or 1600. I don't believe idle CPU multiplier reduction affects reported memory speed, notice your CPU in the first screenshot is only running with a 12x clock multiplier instead of 20x or more if the CPU is busy.
|
EDIT: OK, I think I get ya now. Did a little reading, seems the default settings for this MoBo caps RAM to 1066. I manually raised it to 1600. Now I'm getting a DRAM Frequency of 796.
Could this be causing/contributing to the OpenGL crash?
SI Radio has many DJs and listeners whom hold City of Heroes close to their hearts. We will be supporting many efforts to keep CoH ALIVE!!
Sorry, often I over explain things and confuse the listener about their issues.
Yes, the "official" memory speed for an i7-920 is DDR3-1066, anything faster is considered a hack by Intel. But everyone, as in motherboard manufacturers, does it anyways.
I don't see how system memory would affect what seems to be a driver/card issue. Maybe if the memory is glitchy but if that's the case you should see more software crashing. You could download the memtest86+ ISO, burn a CD (not a DVD) with it and then run it overnight.
Father Xmas - Level 50 Ice/Ice Tanker - Victory
$725 and $1350 parts lists --- My guide to computer components
Tempus unum hominem manet
Well, bumping to 1600 did not fix the crash.
SI Radio has many DJs and listeners whom hold City of Heroes close to their hearts. We will be supporting many efforts to keep CoH ALIVE!!
Quote:
I now have the exact same error dialog appearing after upgrading to the 260.99 drivers (which doubled my FPS in the game from 30 to 60) when I open my contacts list sometimes. Even when it doesn't crash it gets twitching (display flickers) until I do a /unloadgfx|
Originally Posted by Suzumebachi
Strangely I had this issue, or something very much like it for the first time tonight.
 So far it's a one-off but the only thing that's changed on my system is an update of the graphics drivers to the current 260.99 |
Very weird.
Quote:
How does this test differ from Prime95?|
Originally Posted by Father Xmas
You could download the memtest86+ ISO, burn a CD (not a DVD) with it and then run it overnight.
|
I don't want to overlook things, but I'm having trouble believing this is a memory issue since nothing else is affected.
Unlike what some have reported, I can find no consistent pattern that triggers the crashes. They've occurred while zoning, mid-fights, and on really simple maps. Really intense maps (ITF, LGTF, Arachnos base maps) haven't affected things at all.
Thinking of reloading Aion, see how the system performs there. I had no issues with Aion, APB, or AoC.
SI Radio has many DJs and listeners whom hold City of Heroes close to their hearts. We will be supporting many efforts to keep CoH ALIVE!!
For what it's worth I've been having crashes since they changed whatever was changed in I19, and i get the same message you get. There are no problems with any other games and I can easily run more memory intensive ones. I'd guess it's an issue on CoH's end.
Active (Freedom): Setna (Ice/Psi Dom), Arram (WP/KM Tank), Tesmiel (Elec/SS Tank), Astredax (Robot/Dark Mastermind), Operative Vidali (melee fortunata)
Retired (Virtue): Gaav (Inv/EM Tank), Baqra (Fire/SS Tank)
Spent a few days playing X-Blades and Tomb Raider: Underworld. None of these game crashed and both are considerably more graphic intensive than CoH.
SI Radio has many DJs and listeners whom hold City of Heroes close to their hearts. We will be supporting many efforts to keep CoH ALIVE!!
Ok, CoHSupport just sent me back to driver version 197...
This has disabled SLI, cut my FPS from 60 to 30, and oh yeah the game looks like crap now.
I'll be playing to with this driver over the weekend, to see if the crashes continue or not.
Needless to say I'm extremely annoyed by this.
SI Radio has many DJs and listeners whom hold City of Heroes close to their hearts. We will be supporting many efforts to keep CoH ALIVE!!
I have a GTX 260,
should i try to go back in drivers?? I just want to stay logged in!
Well, the game hasn't crashed yet, but I haven't run an Apex TF. That TF seems to trigger multiple crashes during the sewer portion.
I was able (with EVGA's assistance) to get SLI working with the 197 drivers, but game still looks bad, and the frame rates still down. As an added benefit, there is now an annoying flickering when ever I click on a contact or contact window...
SI Radio has many DJs and listeners whom hold City of Heroes close to their hearts. We will be supporting many efforts to keep CoH ALIVE!!
Sorry to necro post, but figured this was the best place to update this.
After running every single test the EVGA support team could come up with, we could find no problems with either my 295GTX or the any of the drivers. In a last ditched effort, I removed the 295GTX and swapped in an older 7900GTX.
Much to everyones surprise, the crashes ceased. I don't know if this means there is a problem with the 295GTX that the tests are missing or if there is some incompatibly with CoH and the card. The folks at EVGA could not explain this, but have agreed to RMA the card.
I've been stuck with the 7900GTX all week without a single crash. The card is on it's last legs, multiple graphical glitches, which is why I replaced it with the 295GTX in the first place. Presently I am waiting for FedEx to deliver the 295GTX to EVGA, at which point they'll send me a replacement.
SI Radio has many DJs and listeners whom hold City of Heroes close to their hearts. We will be supporting many efforts to keep CoH ALIVE!!
On March 18, I received my GTX 295 replacement. I installed it immediately and logged into game. I ramped up the graphics to max and CoH crashed immediately. Called EVGA, made several tweaks and a few driver roll backs. Nothing prevented the crashes. EVGA asked me to RMA this card so they could send me yet another card.
...
SI Radio has many DJs and listeners whom hold City of Heroes close to their hearts. We will be supporting many efforts to keep CoH ALIVE!!
Des, i'm having the exact same thing with a GTX 275. I get the exact same error dialog that Suzumebachi posted.
I really hope you figure this out because I don't have the energy to do all the work you're doing and I hope I don't have to replace this card.
Good luck man
Level 50 is a journey, not a destination.
▲Scrapper Issues List - Going Rogue Edition▲
As far as I can tell, the only common factor here is CoH itself.
I'm on the very latest nVidia drivers with my GTX 560 Ti and I random-crash out with the OGL issue all the time. Sometimes I can just be standing there and it'll go POOF.
Strangely enough, I didn't have this problem with my GTX260 or my GTX460.
So the fact that it's doing so now is immensely frustrating.
I've even tried contacting nVidia and can only go so far with them since they want me to test it with other games. And CoH is the only game I actually play.
I'm probably going to have EVGA RMA the card for me. But I'm seeing enough of these posts on the board that that I'm uncertain it'll help.
At least this time, unlike the StepUp program, I paid for cross-shipping on RMAs.
Clicking on the linked image above will take you off the City of Heroes site. However, the guides will be linked back here.
Quote:
That's just disturbing... This is the first I hear of 400 or 500 series cards having the same issue. |
Originally Posted by Hyperstrike
As far as I can tell, the only common factor here is CoH itself.
I'm on the very latest nVidia drivers with my GTX 560 Ti and I random-crash out with the OGL issue all the time. Sometimes I can just be standing there and it'll go POOF. Strangely enough, I didn't have this problem with my GTX260 or my GTX460. So the fact that it's doing so now is immensely frustrating. I've even tried contacting nVidia and can only go so far with them since they want me to test it with other games. And CoH is the only game I actually play. I'm probably going to have EVGA RMA the card for me. But I'm seeing enough of these posts on the board that that I'm uncertain it'll help. At least this time, unlike the StepUp program, I paid for cross-shipping on RMAs. |
Hyperstrike, do you have the EVGA OC Scanner? According to the folks I talked to at EVGA, it is OpenGL based. If you run it, it should be able to tell you if the card is functioning properly. I ran it on my original GTX 295 and it passed, but when I RMA'd the card, EVGA told me it failed the OC Scanner in their labs.
Curious to know if you have one of those older cards you can swap in, to see if the problem continues. I'm currently stuck with an old GTX 7900 while I wait for EVGA to receive the 2nd RMA'd card. I didn't pay for cross shipping, starting to wish I had.
FYI, the issue does not seem unique to CoH. I've read several posts in various tech forums regarding this issue popping up with graphical programs. I presume these programs are OpenGL based? CoH is the only game I am aware of that is affected. That kind of makes sense since most games seem to be more Direct X based?
SI Radio has many DJs and listeners whom hold City of Heroes close to their hearts. We will be supporting many efforts to keep CoH ALIVE!!
One thing that seems to make the crashes more frequent for me is to have youtube going in a browser as I'm playing.
Have you guys noticed anything like that?
Level 50 is a journey, not a destination.
▲Scrapper Issues List - Going Rogue Edition▲
Quote:
Yup.|
Originally Posted by Desmodos
That's just disturbing... This is the first I hear of 400 or 500 series cards having the same issue.
|
Quote:
| Hyperstrike, do you have the EVGA OC Scanner? |
Heck, with the GTX 590 right now, it's fricking DANGEROUS! (267.52 series).
Quote:
| According to the folks I talked to at EVGA, it is OpenGL based. If you run it, it should be able to tell you if the card is functioning properly. I ran it on my original GTX 295 and it passed, but when I RMA'd the card, EVGA told me it failed the OC Scanner in their labs. |
Quote:
| Curious to know if you have one of those older cards you can swap in, to see if the problem continues. I'm currently stuck with an old GTX 7900 while I wait for EVGA to receive the 2nd RMA'd card. I didn't pay for cross shipping, starting to wish I had. |
Quote:
| FYI, the issue does not seem unique to CoH. I've read several posts in various tech forums regarding this issue popping up with graphical programs. I presume these programs are OpenGL based? CoH is the only game I am aware of that is affected. That kind of makes sense since most games seem to be more Direct X based? |
Clicking on the linked image above will take you off the City of Heroes site. However, the guides will be linked back here.
You could try testing with the Unigine Heaven Benchmark demo. I believe it has an OpenGL mode.
http://unigine.com/download/
Quote:
Just found this: (for ddr3)|
Originally Posted by Desmodos
I have 6 GB (3 x 2 GB sticks) of RAM; Corsair XMS 3 DDR3.
Timings were manually set to 9-9-9-24. Voltage manually set to 1.65 V. These were the recommended settings. Since the problem persists after a fresh install, it seems unlikely to me that some extraneous background software could be a factor, but I'll see about posting a HijackThis anyways. |
>>>
According to JEDEC[2] the maximum recommended voltage is 1.575 volts and should be considered the absolute maximum when memory stability is the foremost consideration, such as in servers or other mission critical devices. In addition, JEDEC states that memory modules must withstand up to 1.975 volts before incurring permanent damage, although they are not required to function correctly at that level.
<<<
Maybe drop the voltage to 1.575?
I was having a similar problem with WOW a while back, everything else worked, system was fine etc, wow just kept crashing, increasing my ddr2 voltage to 2.05 fixed it. May be another part of the system and not your vid card/cpu that isnt keeping up.
How many dimm sockets do you have and how are they slotted? Does your Bios support various ram timing and interleaving options? If so try backing them down a bit, maybe to the next slower memory speed and try turning off interleaving or at least way down.
On another note, since it is being a pain to find the problem, have you tried clearing your Cmos and setting up bios to defaults?
Quote:
I was reading up on Open Gl and Nvidia a few weeks back and saw a recommendation to use "compatibility performance mode" rather then single display mode for open gl problems.
|
Originally Posted by Xaphan
I've been having this problem too. One thing I tried today was going into the NVIDIA control panel and disabling Multi-GPU mode. Not sure if it's fixed it for sure, but so far I've been logged in several hours (running two instances) without a problem. But that could just be blind luck.
But still, Disabling Multi-GPU mode has worked for me so far. I'll post here if it turns out I'm wrong and it crashes again. But hopefully it won't.  |
Quote:
My gtx 280 worked just fine till I upgraded. |
Originally Posted by FryGuy
Much like others in this thread, I was getting the NVidia "driver has encountered an error and will recover" dialog box when I was running with an Nvidia GTX 285. Nothing I did seemed to make any difference... until I swapped video cards.
Yes, it sucks. And it's expensive. But it completely solved the issue for me. Now I'm running an NVidia GTX 480 with the 260.89 drivers and I haven't had a single crash on the same machine that was running the 285 video card. I'm even overclocking my CPU. I don't know if there are stability issues with the 2xx chipsets or the older card had hardware issues. |
If it is the vid card many can be fixed easily (imo) and cheaply by baking them. I can get you a thread if you need more info. It has to do with the solder process used sometimes makes joints that will crack internally after enough heat up/cool down cycles. I've read a lot of posts about dead vid cards fixed this way.
-- Edit --
Are you running the Evga Precision, Riva Tuner or any other utilities for tweaking your vid card? There have been reports of incompatibilities with COx, particularly when opening dialog windows in game.
What is your power supply? Watts and brand. These new vid cards suck a whole lot of juice a stressed PS could be causing problems as well. Do you hear a high pitched whine from the case or anything like that?
Quote:
The EVGA X58 X3SLI mo bo has 6 dimm slots. Currently, I am using 3 slots with 6 gigs of ram. Corsair XMS3 were the recommended RAM for his mo bo. Per specifications and at EVGAs recommendation, the voltage was set to 1.65 v. RAM timings were set to 9-9-9-24, per specifications. RAM sticks are slotted into slots 1,3,and 5, as per mo bo specifications. RAM speed was initially left at default. Per Fatherxmas suggestion, I did raise them to 1600 Mhz, but that did help the issue.|
Originally Posted by _Zep_
Just found this: (for ddr3)
>>> According to JEDEC[2] the maximum recommended voltage is 1.575 volts and should be considered the absolute maximum when memory stability is the foremost consideration, such as in servers or other mission critical devices. In addition, JEDEC states that memory modules must withstand up to 1.975 volts before incurring permanent damage, although they are not required to function correctly at that level. <<< Maybe drop the voltage to 1.575? I was having a similar problem with WOW a while back, everything else worked, system was fine etc, wow just kept crashing, increasing my ddr2 voltage to 2.05 fixed it. May be another part of the system and not your vid card/cpu that isnt keeping up. How many dimm sockets do you have and how are they slotted? Does your Bios support various ram timing and interleaving options? If so try backing them down a bit, maybe to the next slower memory speed and try turning off interleaving or at least way down. On another note, since it is being a pain to find the problem, have you tried clearing your Cmos and setting up bios to defaults? |
Memory has been tested, with no issues found, but I'll bring up your comments when I next speak to EVGA. The CMOS was reset for a BIOS flash, at EVGAs recommendation. Currently waiting for the them to send me a third card. Thing is, I am running CoH now with no crashes on EVGA GTX 7900 (though be it at much lower graphics settings). All signs seem to point to the card, either by defect or design.
SI Radio has many DJs and listeners whom hold City of Heroes close to their hearts. We will be supporting many efforts to keep CoH ALIVE!!
Not sure what else there is for me to try.