Guild Wars 2

Before you continue reading, you should know that this fix only pertains to those users who receive (quite often randomly) a CTD with the error…

“Exception: c0000005
Memory at address XXXXXXX cannot be read/written”

…and are users of programs such as MSI AfterBurner, EVGA Precision X, or any other video card overclocking utility that relies on the RivaTuner core program/statistics server. These programs are written by the same author of the once-popular RivaTuner, so they are very similar/based upon the same code as far as I know.

The following blurbs reference my use of ATI cards while diagnosing this issue, however should be relevant to NVidia users as well.

Since I dislike using Catalyst Control Center for my ATI video card, I opt to use MSI AfterBurner instead, as it gives me complete control of clock speeds, fan speeds, and voltages, offers performance monitoring, an in-game OSD, and many other advantages.

After suffering through these “memory CTDs” for a couple weeks, and watching the game get updated almost nightly but in no way fixing my CTDs, I went into fix-it mode. I tried nearly everything – forcing Catalyst A.I. completely off (requires registry editing for newer cards such as 7xxx series), raising voltages in the system (GPU, CPU, CPU PLL, FSB, NB, RAM), clean driver re-installation, getting the new CAP for the 12.8 driver, updating Windows 7, uninstalling .NET 4.0, etc. I was nearly stumped.

Then it occurred to me – since I recently upgraded to the 7870 GHz Edition up from an aging 4890 (which I had BIOS flashed to overclock, meaning I didn’t use CCC -or- AfterBurner), I had started receiving these memory CTDs in GW2. I didn’t have those with the old 4890. So I asked myself: what did I change? Besides the actual hardware upgrade, I installed the latest version of MSI AfterBurner, in order to make separate 2D and 3D profiles – not only for overclocking, but for fan speed control.

So the other day, I really sat down with AfterBurner and went over ALL the settings available. Now, I had used AfterBurner before in the early days of my 4890, long before I flashed it’s BIOS. I figured that -maybe- I ought to re-review all of the individual settings, as perhaps something had changed or I overlooked something important. It should be noted that I have AfterBurner set up to use it’s Automatic Profiles Management system, simply meaning it will detect (through the use of the RTSS server that runs in the background) application calls to the Direct3D and OpenGL runtime libraries, and automatically change profiles to what you have selected for 2D and 3D. In essence, when it detects application calls to D3D or OpenGL, AfterBurner automatically loads my custom 3D profile – upon exiting the application/game, will automically re-load my 2D profile. This is convenient for those that do not want to use hotkeys and the like for changing profiles.

So, upon perusing the many options in the AfterBurner settings page, I came back to one single thing that caught my eye. There is a setting called ‘Enable low-level hardware access interface’ in the main tab of the settings page, with the options of ‘User mode’ (this is the default setting) or ‘Kernel mode.’ The tooltip is lengthy, but basically states that ‘User mode’ gives the best monitoring performance and doesn’t consume kernel address space, while ‘Kernel mode’ can protect hardware-mapped memory ranges from unauthorized access by badly-designed applications* and in this way can protect the system from instability. That got me thinking – I had been receiving, seemingly at random (could be 2 minutes or 2 hours), these memory CTDs in GW2 and apparently the default setting of ‘User mode’ does not provide protected kernel address mapping for AfterBurner/RTSS. I decided to change that setting to ‘Kernel mode’ and hoped for the best, as I could not find any other setting that may cause system instability that I had not already changed…

(continued in post #2)

Well lo and behold…I have now gone several days, with upwards of 7+ hours continuous uptime in-game, with zero crashes. While I will not claim this to be the end-all, be-all fix for the related memory CTDs (feedback required), I think we have found our culprit – or at least one of them.

////////////// SEE ATTACHED IMAGE //////////////
If you suffer these memory CTDs and are using MSI AfterBurner:

1. Change the ‘Low level hardware access interface’ setting from ‘User mode’ to ‘Kernel mode.’ Restart AfterBurner. If you are not using the automatic profiler option, this should be the only option needing to be changed, as the OSD server .exe should not be running in the background. This can be verified by checking Task Manager for RTSS.exe; it should not be there. If it is running, proceed to step 2.

2. If you’re using the automatic profiler to detect 3D apps and change profiles on-the-fly, attempting to use other features of the OSD server (not recommended in GW2), etc., I would recommend the following steps:

2a. Open the OSD server (the other icon in the system tray). Change the following options to: Show On-Screen Display – OFF, Application detection level – LOW, Stealth mode – ON, On-Screen Display support – OFF.

2b. Click the wrench (options) icon in the OSD server. Ensure ‘Enable compatibility with modified Direct3D libraries’ is disabled – unchecked. Click OK to accept. This is not pictured in the attached screenshot (forgot, sorry!) but should still be verified.

3. Minimize the OSD server back to the system tray and restart AfterBurner.
////////////// SEE ATTACHED IMAGE //////////////

That should do the trick – at least it did wonderful, magical things for my gaming experience. No longer do I suffer memory CTDs at random intervals. What a beautiful thing.

Well that doesn’t help me at all. I don’t use Afterburner or any type of overclock software. I do use Catalyst Control Center. Is there a setting in there doing the same thing? My vid card is Radeon HD 7700. I didn’t have these CTDs until the last week, more so after the September 14th patch. Although, I did update the Catalyst profile that was just released before that, the one that specifically makes playing GW2 better or something, not that I had any issues before anyway.

I do not have a solution for those receiving this type of error that are not users of MSI AfterBurner or other programs based on RivaTuner/Statistics Server. Sorry.

Though it may be a related issue – I mean, after all it does say that “Badly designed applications” can sometimes reference hardware-mapped memory ranges, resulting in system instability. One can only wonder exactly what GW2 is trying to access, but from what I can gather, I’m fairly certain it shouldn’t be trying in the first place…

I was having this issue and managed to fix it by disabling overlay in mumble. Forgot I had that on and it definitely doesn’t work with GW2 for me. With mumble and MSI Afterburner open, GW2 would not start without crashing. I didn’t get constant CTDs though, hardly ever really…

Overlays (Mumble, TeamSpeak, MSI AB, etc.) can cause CTDs as well, and not just in GW2 but many games – this is pretty well documented. Also the reason my post instructs to disable the OSD from within AB.

I was getting CTDs with the overlay disabled (I wouldn’t exactly call them “constant,” more like “random” – could be 5 mins, 2 hours, and occasionally I could play without it occurring at all), so I knew it had to be something else. Turns out it has something to do with AB’s kernel address space and GW2 trying to access hardware-mapped memory ranges. Again, I’m fairly certain GW2 or any game should not be trying to access this memory space in the first place, hence the MSI AB tooltip regarding “Badly designed applications” (and the option for ‘kernel mode’)…haha!

I have been experiencing these exact errors since launch/first few new patches. I made my RivaTuner stats match yours, but I use EVGA Precision X and there is not a compatibility option that I see.

Well my issues are gone. Not sure what I did as I changed a few things around. Namely turning off AMD Fuel and AMD external error utility via the msconfig start up options. I also uninstalled Google Chrome, as that was crashing a lot lately. Also updated Adobe Flash. Not sure if any of those affected the CTDs, but I haven’t gotten one in 3 days, knock on wood. So strange. Of course, now I have a wierd issue of my mouse not fully bringing my display back online when the monitor goes to sleep. I have to hit ctrl alt del and then escape but that’s a minor issue and doesn’t cause anything else but inconvenience of 2 seconds. Still strange though.

Loyram.1829:

I have been experiencing these exact errors since launch/first few new patches. I made my RivaTuner stats match yours, but I use EVGA Precision X and there is not a compatibility option that I see.

Right. I couldn’t find it in Precision X either. It’s one of the options that’s missing from the main settings page vs. MSI AfterBurner.

You should give MSI AfterBurner a shot if you’re still having CTD issues with Precision X. It’s literally the same program, written by the same author, with a different skin (and more options). Just use the settings I recommended and you should be good to go with AfterBurner. You don’t have to uninstall Precision X, either. Just disable it from startup etc., as you don’t want both AB and PX running at the same time.

Glad to say that I have still not experienced any more of these ‘memory’ CTD – 5 days later, and still going strong!

I’d still like to hear some feedback from other affected users.

Didn’t work for me. I really want to play this game. =/

Loyram, I’m not sure what that crash relates to (I’ve never even seen it – and I’ve seen a whole lot of different CTDs with GW2), so more than likely you have some settings wrong somewhere [display driver, AfterBurner, Windows] or your overclock isn’t stable – I assume you are oc’ing since you’re using AB/PX? One thing I noted with my specific card (7870) with using AB is that it will set the voltage a bit lower than what you manually set it to, so you may want to verify your unloaded/loaded voltage using GPU-Z and FurMark. That may be a specific issue with my card – I’m not sure, but it’s worth checking. If you are not over-volting the card, then you can completely DISABLE low-level hardware access, as it seems to only be needed for voltage control, as far as I can tell. Also, which version of AB are you running? They just released 2.2.4, but I did all my testing with the previous version (I think it’s 2.2.3). I recall them making some Nvidia-related changes in the newer version, so if you get to your wits end, maybe try the previous version.

Thisisdell – I need more info than just “didn’t work” – such as the error you are receiving (full crash report plz), AfterBurner settings, video card make/model (full specs would be better), etc.

I have still been able to play just fine with no CTDs since making the changes in my OP.

Idolicious.6091:

Loyram, I’m not sure what that crash relates to (I’ve never even seen it – and I’ve seen a whole lot of different CTDs with GW2), so more than likely you have some settings wrong somewhere [display driver, AfterBurner, Windows] or your overclock isn’t stable – I assume you are oc’ing since you’re using AB/PX? One thing I noted with my specific card (7870) with using AB is that it will set the voltage a bit lower than what you manually set it to, so you may want to verify your unloaded/loaded voltage using GPU-Z and FurMark. That may be a specific issue with my card – I’m not sure, but it’s worth checking. If you are not over-volting the card, then you can completely DISABLE low-level hardware access, as it seems to only be needed for voltage control, as far as I can tell. Also, which version of AB are you running? They just released 2.2.4, but I did all my testing with the previous version (I think it’s 2.2.3). I recall them making some Nvidia-related changes in the newer version, so if you get to your wits end, maybe try the previous version.

Thisisdell – I need more info than just “didn’t work” – such as the error you are receiving (full crash report plz), AfterBurner settings, video card make/model (full specs would be better), etc.

I have still been able to play just fine with no CTDs since making the changes in my OP.

09:07 CST
The patch notes from today say it fixes various crash bugs.

We shall see. Thing is I don’t overclock because I thought that might have something to do with it. So staying default still yields problems.

I’m working with support right now, giving game advisor logs with startup utilities running normal and with all disabled and still crashes. However, I have still yet to try today. I’m on it

Edit: 9:48 – game crashed.

Well, unless you are running AfterBurner for separate fan profiles, I don’t really see the need for you to run it at all if you aren’t overclocking. Or do you get the same crashes running CCC instead of AB? If not, I suggest you don’t run either AB or PX, just stick with CCC and see what happens – if you haven’t already.

edit: I assumed you had an ATI card. Even if you have NVidia you can just try running their display driver software in place of AB if you aren’t overclocking.

With only CCC I had a crash every 5 minutes (c0000005). After installing MSI AB (without unistalling CCC) and setting it according to the picture posted by Idolicious, I am able of playing for an half hour/an hour. Still it CTD, but the error is now different:
(1st CTD)Assertion: link.m_linkToFull.IsLinked()
(2nd CTD)Assertion: activeBlock == next->prevPhysicalBlock

That is yet another strange CTD. I can only assume you’ve fixed the original CTD issue (c0000005) – but by doing so you may have uncovered a separate issue. I don’t think your [new] CTDs are related to the c0000005 errors you were getting…have you ran MemTest86+ over night to check your system RAM? That’s probably where I’d start if all your system temperatures check out okay.

Ran MemTest, great results. Still not sure why I’m experiencing this. I figured I could try and roll back to 295.73 cause my gtx 570 buddy is using those, but no dice.

Thinking I might have to wait until my next gaming PC is built before I can play again.