World Boss Crashes

#1 - Sept. 18, 2014, 11:56 p.m.
Blizzard Post

These world boss crashes are getting real tiring. How many times do I need to submit these similar callstacks?
I crash almost daily at Karka Queen, or Shatterer. Today, I crashed at both.

I don’t crash anywhere else. I think in the years I’ve been playing, I’ve had one (1) crash that wasn’t at a world boss.

I have been sending in callstacks for over a year now for crashes at world bosses, and while I’ve not been looking that closely at them, they almost always look the same. These crashes showed up around the time the culling feature was removed (when was that, a year and a half ago?)

And no, I’m not interested in playing at “low quality” settings. That is a BS solution, and by doing that stupid workaround, Anet won’t get the volume of callstacks they should be getting over this ridiculous issue.

Further, I play 99% of my gameplay with everything maxed, and there is no preset for that. High only sets about half the settings to max. I still have to go in and manually set half the settings higher after I choose the high preset. I’m not interested in fiddling with a dozen settings multiple times a day just to workaround some crashes that should have been fixed a long time ago.

Please, for the love of {deity} put some priority on fixing these crashes. I have sent you countless callstacks now. The issue is easily reproducible:
1) Go to a world boss where a zerg collects
2) Auto attack boss and wait
3) Crash

#25 - Oct. 17, 2014, 11:05 a.m.
Blizzard Post

These world boss crashes are getting real tiring. How many times do I need to submit these similar callstacks?
I crash almost daily at Karka Queen, or Shatterer. Today, I crashed at both.

I don’t crash anywhere else. I think in the years I’ve been playing, I’ve had one (1) crash that wasn’t at a world boss.

I have been sending in callstacks for over a year now for crashes at world bosses, and while I’ve not been looking that closely at them, they almost always look the same. These crashes showed up around the time the culling feature was removed (when was that, a year and a half ago?)

And no, I’m not interested in playing at “low quality” settings. That is a BS solution, and by doing that stupid workaround, Anet won’t get the volume of callstacks they should be getting over this ridiculous issue.

Further, I play 99% of my gameplay with everything maxed, and there is no preset for that. High only sets about half the settings to max. I still have to go in and manually set half the settings higher after I choose the high preset. I’m not interested in fiddling with a dozen settings multiple times a day just to workaround some crashes that should have been fixed a long time ago.

Please, for the love of {deity} put some priority on fixing these crashes. I have sent you countless callstacks now. The issue is easily reproducible:
1) Go to a world boss where a zerg collects
2) Auto attack boss and wait
3) Crash

Thanks for submitting your log files. I’ve found your corresponding crash reports and we will look into the issue immediately.

The skinny is that it looks like we’ve allowed you to shoot yourselves in the foot with maxed out graphics settings combined with crazy boss zergs. The GW2 client is a 32bit process which has access to just 4gb of RAM, even less in practice due to fragmentation in high churn environments like a zerg fest. I can see from your log files that you are hitting an upper limit of about 3.5 gb of allocated memory before crashing on a failed allocation(i.e. running out of memory).

In the meantime here are some things you can turn down that will free up some memory in zergs without compromising graphics too much:

-Disable “High-Res character textures”
-Ease back on “Character Model Limit” and “Character Model Quality”
-Ease back on “Lod Distance”

The memory use for characters can roughly be described as = character count * character quality * 4 (if high-res is enabled).

Report back if those changes don’t help and I will do the same when we have a fix in-bound.