Guild Wars 2

Azure Prower.8701:

If your ring is infused before heart of thorns, it won’t appear in the mystic forge.

Also. I infused a Ring of Light today. Then I joyfully discovered I couldn’t attune that either after finding out the previous fact minutes before hand.

Let’s seriously hope these are bugs that are going to be fixed.

A workaround for this is probably to use one of the Infusion Extractors; I was about to say that I didn’t have this problem with my Infused Rings, then remembered that I used those to extract the infusions first…

Yup. I’ve posted it as well, and sent reports.

Win7 x64, Gilded Hallow Guild Hall.

Enter, walk over to the War Room Proprietor in the tavern, select “what’s required to build a war room”, also open the guild panel while that’s open, the game will shortly crash with one of several errors – I’ve seen an OOM and an access violation from a pointer deref to 0×00000008.

If it seems to work fine, click any of the tab options in the guild panel, and it will load them and immediately die – so badly that the “GW2 crashed” dialog will close within 5 seconds.

I submitted a crash report for the first (access violation), but the OOM: Heap error results in the crash report dialog closing before it sends successfully.

Neat.

This has become slightly less painful with the HoT patch drop, since dungeon progress and which paths have been completed is now displayed in the individual dungeon achievements, so the information is available, even if not in convenient API format.

I’d slightly care about the existence of an API query for “has X character on Y account done dungeon Z path # today”, but what I’d really like is to query against the bitmask for “what dungeon paths for dungeon Z have I done, ever”?

Because there’s no user-exposed way to get that information, and short of keeping our own spreadsheet or trial and error, you just can’t see it…and when you’re talking about one missing path in Arah, trial and error is a time(or gold)-intensive task…

So, I turned the texture levels down, and did Teq again, and got through it without the game dying.

Shortly thereafter, though, it crashed, so I screencapped procexp and copied the crash text.

CPU/Private Bytes/WS/PID/Desc/Company/GPU/GPU System Bytes/GPU Dedicated/GPU Committed
http://i.imgur.com/rxGMpRj.png
http://pastebin.com/J0r9eqP5

It looks like it was indeed failing to allocate memory for graphics effects and bailing out…

e: For my next trick, I’ll see if an ETW trace has anything more to say.

mrstealth.6701:

I understand that, and would want the same. I wasn’t suggesting it as a proper solution, just a test and potential work-around.

Unfortunately, you’re likely right about being out of luck unless Anet actually makes a 64bit version of the game(or implements limits on the 32bit client). If you’re seeing 3.3GB RAM usage, not including GPU memory, you’re probably just hitting the 4GB memory addressing limit on 32bit applications. I don’t think there’s anything you can do about that yourself.

Are you running the game at an extremely high resolution, or across multiple displays? I’d expect to be hearing more reports of this if it was happening at 1920×1080 with those settings turned up.

I’m only at 1920×1080, no 1440p or multi-monitor setup here.

Yeah, I’d guess it’s just trying to allocate a chunk and failing to allocate a contiguous chunk in its address space that size, and bailing, but don’t know. Womp womp.

mrstealth.6701:

You could try a different driver version. Either an older one, or a beta driver if one is available.

And have you tried lowering graphics settings to see if the issue persists with your new GPU? Since this is happening at larger events, character model limit/quality would probably have the biggest impact.

I did lower both of those, yes, on the same logic, and it hasn’t crashed since.

But I’d prefer to not have to lower it when I seem to have sufficient resources that aren’t being utilized.

mrstealth.6701:

Are you using DSR (Dynamic Super resolution) on the GPU, perhaps in combination with the game’s render sampling set to supersampling? I don’t have a GPU that supports the feature, so I’m not sure how it interacts with the game’s own over-rendering/downsampling setup.

I have it set at “native” for texturing, and no DSR configured in the NVIDIA driver, so I don’t think that’s biting me.

So, I recently upgraded my GPU to a 4GB GTX 980, and upped the graphics settings in GW2.

Since then, I’ve started having issues with GW2 OOMing in very large group content (world bosses like Tequatl) – as far as I can tell, this is not an issue with actual RAM in my system (16 GB at the moment), nor video RAM usage on the GPU itself, but just an issue with GW2 on higher settings hitting somewhere around 3.2-3.3 GB heap and dying, badly, on allocation failure.

I’ve tried a repair just to be sure nothing is awry in my installation itself, but to no avail.

I’m running on Windows 7 64-bit, so this isn’t “just” an issue with only ~2GB of per-process address space.

Any plans on the horizon for either a native 64-bit build or possibly leveraging AWE to get more than a few GB of memory from a 32-bit build?

(Or just a method of more robustly handling this than crashing out to desktop, I suppose…)

Exactly as it says in the topic – I was in Ruined Shore, chopping down Cypress and Orrian Saplings without a problem – I got attacked while chopping down a Cypress Sapling, and two of the three logs came out as Ruined Logs.

I later was chopping an Orrian Sapling and was attacked, and one of the three came out Ruined.

This does not, as far as I know, happen with normal axes – is it somehow expected that this happens with this axe, or is this a bug?