Leamas.5803

Everything you said is correct, the trick is reliably reproducing the problem so a repair shop can actually troubleshoot it and determine whether a change actually fixed it. Like most others, outside of GW2, I've never managed to find a way to reproduce the crash at all, much less reliably, on Windows 7 or Windows 10 (Skipped 8). Even on Windows 7 it never went to a BSOD...always either a hard freeze, typically with looped sound, requiring a hard reboot (Reset button won't work)...probably the most common type for mean instant reboot as if the reset button was presseda complete shutdown that required the computer to be started twice, first restart start seems to reinitialize hardware, then the PC shuts back down after a few seconds, the second start starts up the PC. On my machine, this was the least common of the 3 types of crashes. The only other time I've seen this type of crash was when I doing tech support/development for a steel company (Small IT shop where I did all the hardware and wrote the software) and the computer case received a large static shock. Metal chair and metal desk and it would arc between them when they touched and cause the computer to immediately shut off. I ended up tethering the chair and desk together with a wire to keep them at the same potential...which also save me from getting wicked static shocks....and nothing ever truly useful in the Windows event logs...just a generic nVidia error. If they can't reproduce it and hardware passes their testing and it only happens in a single application, EVERY reputable shop will tell you it's the software. The last time I took it in, I left it with them for 8 days and they put it though all their testing rigs and could not reproduce the crash or find any hardware faults. Now, I don't think that it's entirely the software, but I think we can agree that ANet is doing "something" that most other developers are not that is triggering a hardware fault most other developers seem to avoid, which makes it particularly problematic to debug. So short of giving the shop your GW2 password and seeing if they can reproduce the error that way, what else can be done? It's really the only option if you can find a trustworthy tech. It's actually against the privacy policy with the shop I deal with. They weren't allowed to take my password. I offered. Even then though, the duration between crashes can vary wildly from several hours to seconds. One guy in this thread said it went away for 2 days before returning. On my machine I've seen as much as 6 consecutive hours (Basically a full evening), but that was not typical. More typical, it was between 20-120 minutes. The other issue, of course, is that if you are not under warranty, paying for the type of support required to diagnose this can get expensive fast since it requires a great deal time and baby sitting of the machine. Between gas (Since it's a 4 hour round trip for me) and paying for tech support I've spent probably $400-$500 unsuccessfully trying to get it resolved, and they're one of the largest PC tech support shops in Edmonton.

Sorry to hear that, but contrary to what some here would have you believe, this issue has existed in GW2 pretty much since launch and has never been nailed down to a single definitive cause. Driver version or GW2 patch sets can sometimes make it better or worse. I started having these crashes at the end of 2012, on a brand new machine. Over the years, what I've seen people report, sometimes a new PSU fixes it, sometimes new memory, sometimes a new video card, sometimes lowering over clocked settings (Which fixed it for me), sometimes changing the power management settings in Windows and/or the nVidia control panel, I've even heard people claim conflicting HD audio drivers causing it (i.e. Realtek vs nVidia). There is no consistent fix and no definitive cause, the only consistent facts most people report, is that it is almost always an nVidia equipped system and it affects GW2 exclusively and no other game or application on their system. While I've since replaced it as my primary rig, I still have that 2012 machine (My kids are now using it) and it will still have those same crashes if I don't down clock before starting up GW2, but works fine for absolutely everything else they play.

That's what system benchmarks are supposed to do. Either way, it still only happens in GW2. The most similar games I've played to it are ESO and FFXIV and neither had any issues at all. In fact, in order to keep playing GW2 while I was playing the other two, I had to find another solution because neither ESO or FFXIV would work on the ancient video driver I was using because GW2 was the most stable with it. For a while I would install/uninstall drivers, depending on what I was playing, but that wasn't a feasible solution. At the time the driver releases were in the 350s I was still using 306.97. While I still had crashing, anything newer than that made GW2 completely unplayable, while everything else was happier.

EXACTLY!! On top of what you listed, I've also tried (Was getting desperate and grasping at straws): Disabling various features in the BIOS that I don't use, such as RAID controllers Tried many different versions of video drivers, mostly minimal installs, to see and one was more stable than another (Which was the case). Various windows security settings that I no longer remember Swapping to a different KB/mouse combination (Cheap test) Changed the monitor (Wanted a bigger one) Changing all the memory (Had some dead stuff, so was a free replacement) Had the PSU load tested (3 times) Brought it to the shop and left it with them and they stress tested various things for an entire week. Found nothing. Reinstalled Windows and tested GW2 on a "clean" install...twice Moved Windows and GW2 from a HDD to a SSD and moved the system drive to a different port on the MB. Removed all USB attached devices, other than the mouse and KB Ran GW2 with the repair option to validate the file cache I'm sure I did other things that neither you or I mentioned as well, but it's been some time. My conclusion ultimately was two forked, 1. "Something" in the machine was just flaky AND 2. GW2 is doing "something" (Be it a bug, or rarely used functionality) most other applications do not and that is triggering a crash. I did NOT try a new video card or replacing any other hardware other than what I've already listed. I could not justify the cost for the sake of a single game since the computer still to this day, now nearly 5 years old, has no issues in any other game or application other than GW2. That has been the case since I bought the computer new in 2012...so it is not hardware degradation.

There is a trial version of HWMonitor Pro That will allow you to monitor statistics in real time (Well, update once a second) via network. I've never used it for this purpose, so don't know how to set it up off the top of my head, but it might tell you if you're getting sudden spikes or drops in your voltages right before the crash.

Since down clocking didn't work. As Player indicated, how are your temps looking? You can use HWMonitor to check this. If it were the video card/OC I would typically expect to see an error in the event log related to nvlddmkm. Where you have a kernel power error, it may ultimately be a faulty power supply. I would take it in and have the PSU tested.

Since GW2 doesn't require a super powerful GPU, personally I would go more heavy handed on the clock and reduce all possible clocks by a full 10%. If it works, then you know this is the problem and you can start increasing the clock incrementally.

Morfedel, also as a summary, can you let us know what you have tried in order to resolve it? It'll help us so we don't keep rehashing things you've already tried. Maybe take a screen shot of GPUz or nVidia Inspector and post it so we can see all the numbers. Both look the same, with the exception that the latter allows clock changes. As DeWolfe says, the clock speed sounds low.

Geek Squad is Best Buy's tech support. The problem with this issue is that even professional techs can't reproduce the problem with any of their testing tools. If they can't reproduce it, it makes it very hard for them to debug it. When I had this problem with my PC I debugged it with the folks on the nVidia and eVGA forums and brought it in to tech support 3 times...and no one could figure it out and nothing anyone had me try could reproduce the crash outside of GW2. The conclusion on ALL fronts was that it was most likely a software issue and that it should be resolved at that level.

This can help get us started. Download and run it.https://www.techpowerup.com/gpuz/ so, the GPU Core Clock is running at 300.0 MHz, and the GPU Memory Clock is running at 150.0 MHz. anything else I need to check?Which card is it?

This can help get us started. Download and run it.https://www.techpowerup.com/gpuz/ nVidia Inspector is really good and easy too.http://www.guru3d.com/files-details/nvidia-inspector-download.html

I first had this issue on my eVGA nVidia 670gtx SC back in late December/early January of 2012/2013, first reported it to ANet in late January 2013 and have been following it ever since. While the creator's update may have made it worse in recent months, it has been an issue with GW2 for much longer. I can say that 99% of the issues I've seen over the years have been with nVidia, but far more people use nVidia, so that number is a bit skewed. With nVidia, there is also typically a driver crash in the Windows event log that accompanies it (nvlddmkm.sys) and searching the vNidia and eVGA, as the vendor in my case, forums typically garners little help. Driver version can affect it. Back when I first started having issues I was using driver version 306.97, but updating to 310.90 made the game unplayable as it would crash as soon I as panned the camera. Newer versions of the driver also have the same problem. Last one I had crashing on (Because I forgot to down clock) would have been about a year ago, when I last played with the kids, so probably 37x.xx. Unlocked frame rate makes it worse, shaders make it worse, heavily populated maps makes it worse. Updates to GW2 can sometimes make it a little better or worse. Typical behavior is that the problem is ONLY present in GW2 and most often presents itself in two ways, a complete system shutdown as if the power was pulled or a freeze to a blank screen with looped sound, requiring a hard reset (Holding the power in for 5 seconds). The color of the freeze screen can vary. The time to crash varies wildly where crashes can happen within seconds of opening the game or occasionally one can play hours before a crash, but crashes typically happen in less than an hour, making group content such as dungeons/raids/map metas/etc. not very viable. It happens with both the 32-bit (When they still had one) and 64-bit clients. All other games, stress test utilities, and benchmark utilities work fine. ANet acknowledged there is a problem back some years ago, with some apparently unfortunate wording that some are "sensitive" to :p, but they've never fixed it (The problem OR the wording).https://forum-en.guildwars2.com/forum/support/support/Crashing-Issues-1/first#post4146866 Over the years I've done a massive amount of debugging with ANet, nVidia and eVGA as well as brought the machine in to tech support 3 times, but never resolved it and nothing has been able to replicate the crash. I've tried heavily abusing the card with things like the Furmark burn in test, pretty much setting the card on fire, and setting the view distance way out with HD textures on Skyrim to make it grind in a more random way. Ultimately, ANet's recommendation was to try down clocking overclocked hardware, which is what I eventually did to fix it and have never had a crash since. (And didn't have to buy a new card) I did the entirety of HoT and most LW Season 3 down-clocked without a single crash (Then took the summer off from GW2, so just getting back to finishing up S3). Exactly... I haven't been looking for answers in this thread, it WAS answered...by ANet...~3 years ago, according to the forum. The "obvious" answer is that there is a problem with the GW2 software, since ANet openly says there is. They have long acknowledged that their software can have crashing problems with overclocked hardware. The problem does not present in any way in any other application or game. If it was a truly unstable overclock, other applications/games should have issues and something should be able to reproduce the problem in a controlled fashion (Something you NEED to be able to do if you want to get the card RMA'd), neither of which are true here. At the end of my battle with this card (Which I did fight for WAY too long), downclocking was ultimately also the suggestion put forth by both the nVidia and eVGA technical forums. You're arguing about wording semantics only technical users would understand, and most of those could care less about since we understand what ANet is getting at. You're arguing against an issue and potential fix the software developer has openly stated is there, which is just ponderous to me. Let's be honest, not everyone here is technically savvy. Could it be possible ANet might have been simply trying to explain it in layman's terms that most nontechnical users could understand? Have you ever done technical helpdesk support for nontechnical users or developed software or technical designs for nontechnical users? You really need to dumb down the language, AND you need to be able to do it without sounding condescending. Often we use words or comparisons that are aren't ideal in our world, but that they'll understand. Here at work it's an ability we commonly call "speaking user" and not every technical person has the ability to speak user. In fact, there are many you should keep far away from end users. You don't like the word sensitive, OK, we get that. How else would you word it in nontechnical layman terms. They could say GW2 "has a bug that will not be addressed", but that sounds far more ominous. They could say "GW2 may not work well with all overclocked hardware", but that is a lot more wordy and basically says the same thing as sensitive. In layman terms, in my mind, sensitive is a good word to explain it to nontechnical users, since GW2 is happy most of the time, but sometimes not. The reality is most nontechnical users could care less about "why", they just want a fix and preferably one that doesn't require new hardware, which I haven't seen you propose anything short of the possibility of failed hardware...you're mostly arguing semantics and technical definitions. At the end of the day, while not technically accurate, "sensitive" describes the problem well for most people...so what's the big deal? If it bothers you that much, perhaps you should put a support ticket and ask them to change the wording on the sticky to something more technically accurate. I've post the URL a couple times. I don't need it changed since I speak user very well, and understand "sensitive". After fighting this issue for over 3 years myself and following it for almost 5, I've never guaranteed this will work for anyone, but I know it has worked for many, and at least I'm trying to help since I understand how frustrating this issue is. I also understand computer hardware is expensive and not everyone has the option to freely swap hardware at will to test things, so I tend to start with suggestions that do not require new hardware. Unless there's something obviously wrong that points to hardware, like crashing in multiple applications or random BSODs, start with software since a software fix is always preferable when possible (Unless I run a shop and am trying to upsell you), that's basic tech support 101. What are you doing to help, other than bitching about technical terminology?

https://forum-en.guildwars2.com/forum/support/support/Crashing-Issues-1/first#post4146866 You can push that the card is failing or that it's buggy all you want, but at the end of 5 years of owning the card the facts do not support that view. "Either the hardware works properly or it doesn't." - Absolutely, the hardware works within its design limitations...it does what the software tells it to do and, with a lot of modern hardware, if the software tells it to do something that's potentially damaging, very often, it immediately shuts down (aka crashes), which is what it's supposed to do to protect itself, and at this point is still working properly. Not like back in the day when my buddy wrote an assembly TSR to burn hard-drive motors out. The HDD was happy to do what it was told, until it fried itself. All I'm trying to do is save people money and frustration. Changing the clock speed is a simple test to see if the game stabilizes. After that, if it works, they can decided whether to keep doing this or whether they want to try a new card. That said, I have seen people replace their card only to have the same issues...and replacing the card seems A LOT of money to lay out for the sake of a single game. If changing the clock doesn't help, change it back, then what are they out but a few minutes of time? I've come across others who are adamantly opposed to changing the clock speed, and I used to be one of them, now I just don't understand why I resisted for so long. I spoke with a guy last month who had this issue on a 1070/7700k and did the clock speed change. Not only did it fix his crashing problem, he claimed he gained a substantial amount of frame rate while also being able to use higher graphics settings, which he kept low to make the game more stable. I'm also a software developer, Oracle DBA, and have done my share of electronics over the years, but most of my experience is on the software side, near on 30 years (Much longer than that if I go back as far as when I was a kid programming Basic games on the Vic-20). I too am old and have been quite stubborn about not buying a new card since in this case I don't think the core issue lies in the hardware at all.

A more likely scenario is that the overclocking on the card is more or less fine since it works for everything else, but GW2 has some code that is causing some runaway or fault condition with certain hardware configurations that it's not handling and down clocking prevents the condition. You don't need to down clock much. In most cases, if it's going to work, a reduction of only 5% will fix the issue and in many cases I've seen over the years, people actually see an increase in frame rate for GW2 after the down clock. Typically I use 10% on that machine since GW2 does not need a super powerful card and the 670 was a beast of a card when GW2 was released, though I'm hearing PoF is more of a pig, but that machine is only running the FTP version now. I've been researching and following this issue since December of 2012, the first time I posted about it on the old forum was January 2013. I've seen on the eVGA and nVidia forums other people complaining of similar issues with other games, but none of which I've owned to test with. Apparently one of the GTA games is notorious for this type of crashing. FYI, for memory, if you can easily get at it, it's A LOT easier to take it all out and test each stick individually. Doing it this way can also reveal issues not obvious with all the memory in the machine. If Arenanet says GW2 is sensitive to over clocking, then it is, but that doesn't mean it sees the over clocking or that it's bad hardware or that the over clocking is bad, the way ANet states it means there is some bug in their code they don't know how, or are unwilling/unable, to fix.

Well, I keep hearing this and in theory it should be true. In practice that may or may not be the case and may be more a matter of opinion than anything. SLi, for instance, again, should be completely transparent to the application/game layer, but back when I bought GW2 in Sept 2012, I had a machine with two gtx275 cards running SLi, it did not work at all for GW2 and would get abysmal frame rates until you switched it to single card mode, or maybe I had to take a card out...I don't remember (And, again, it was ONLY GW2 that disliked the configuration, everything else worked fine). That may be better now, with newer hardware/drivers, but I don't run SLi anymore. That machine ultimately received a 560 and worked fine after that, but has since been retired (Still not a bad machine with an i7-2600k CPU). The machine that followed, the one that has issues with GW2 has a 670 running on a 3930k CPU. It's now the kids' machine and does everything they want it to do...primarily Minecraft, Terraria and Garry's Mod. It never has any issues, ever...unless we try to play GW2 together, which WILL crash the computer within minutes unless the video card is down clocked. My current machine has a 1080 running on a 7700k CPU and does everything I need at the moment and requires no special treatment. What I can tell you for 100% sure, in the ~5 years I've had that machine (Bought December 2012), ONLY GW2 has had consistent issues and I've played many games far more graphically intensive games such as Bioshock Infinite, HD Skyrim (Modded to death), ESO, Wolffenstein: The New Order, Alien: Isolation, among many others, as well as bench marked and stress tested every part that one can get software to test (Furmark, eVGA stress test, Nexuiz Benchmark test, Intel CPU stability test, Memtest86 memory test, etc, etc, etc.). I have also had the PSU (1200W) load tested 3 times by the shop where I bought the computer (Since you need special shop hardware for that). Again, ONLY GW2 has issues with that machine and dropping the clock speed on the video card fixes it. You're right of course, it could be a flaky overclock straight from eVGA or slightly faulty card, or some combination of hardware GW2 does not like but, again, as I said, in my case, no other game or application has issues with the factory clock...so where does the big finger point? Of course, it points at GW2 as being the source of the problem, which is what the hardware manufacturer will say. Now the problem with this is that if the issue cannot be reproduced in a controlled environment no one will warranty or RMA it. This I discovered after 3 years of fighting with the shop and eVGA (Since I had an extended warranty). The short of it is that they need to be able to reproduce the fault before they'll replace the hardware.

Yup, I've been researching this issue and it's been around for a decade. I just dropped my OC by less than 5% and it's been running solid. What's hilarious is that I run the frame limiter and GW2 never maxes out my card anyway. If people are having issues, set the limiter too. It's error protection more than anything. Frame limiter reduced the amount of crashes, but never fully fixed it for me. Down clocked, I have the frame limiter set to unlimited and my graphics options set much higher than it was before down clocking, when I was reducing graphic setting just trying to stabilize it. Back when I first started having issues, certain versions of the video driver were also more stable than others. For instance, 306.97 was vastly more stable with GW2 than 310.90.

Crashing like this has been a problem with GW2 since at least the December after launch (2012), when I first started experiencing it. Downclocking my video card fixed it for me. Arenanet has stated that GW2 is sensitive to overclocked hardware. If I forget to reduce the clock speed on that machine (My kids use it now) before playing, it WILL crash within minutes. I have NEVER had a crash while the video card was downclocked...not once. No one wants to downclock a fancy video card, but give it a try. Reduce it by maybe 10% and see how it goes.