Guild Wars 2

BDave.9386:

goodbeerhunting.8942:

It would be more helpful if they’d turn on the memory debug overlay so they could see when the client was almost of VM address space, and switch zones to free up some space and delay the inevitable crash.

You can do that in the config file. I’m not quite sure how to interpret the data, though, since it’ll say that there’s 2.1GB worth of allocations in 1.8GB of RAM (not a typo) as soon as GW2 launches. Seems like it’d insta-crash if that was the case… <insert joke about lag here>

If you turn on the memory overlay, the only line you need to keep an eye on is Virtual Address Space Left. You’ll notice this value drops as you play and new textures and models are mapped. In high-traffic areas (world events, wvw zergs, pvp lobby, LA, etc.) you’ll blow through this quickly. Using generic models (lowest model quality) will slow this, but eventually you will run out of VM space and crash.

When this gets down below 100 MB free, you might want to think about switching zones. This will free up ~500 MB of space (assuming it’s to a different zone.)

When it gets down below 60 MB free you could crash on a zone or character change. Bringing up the map will chew through this as well. It’s also possible to crash on quit here. You might notice graphical glitches here as well, like the background of the character selection screen goes all black.

One tip that might help after a crash:

If you can’t get rid of the modal dialog box that comes up after a crash to close the client, open up the terminal app (use spotlight) and use this command: killall cider

That should force cider to quit and close the game window and dialog box. You can also ssh / telnet in if you’ve opened that access and kill cider from a remote shell (from your phone, etc.) Saves a hard reboot in some cases.

It’s a little disheartening, but I guess I understand why people still think changing config settings for an X11 driver that isn’t used has an effect (look at the “GraphicsDriver” setting. Cider uses a quartz-based driver, not X11.)

It would be more helpful if they’d turn on the memory debug overlay so they could see when the client was almost of VM address space, and switch zones to free up some space and delay the inevitable crash.

Since 64-bit cider isn’t likely to drop from the sky anytime soon, If we could have a setting to use generic weapon and backpack textures like the current armour setting, that would help a lot for OOM crashes in areas with a large number of players.

Also, if you considered bootcamp but thought you needed an optical drive or couldn’t get it to install with a USB stick, el Cap let’s you install straight from an ISO by creating a partition to hold the installer files. Might be worth your time to check it out.

If the client crashes on the mac, there’s a chance that two processes didn’t exit: Wineserver, and Cider. If you can’t restart the client, check for these two processes in the activity monitor and force quit them.

Also, if the client crashes to a black screen which you can’t tab out of / escape from, ssh into your Mac (from your phone, etc.) and ‘killall cider’ from the shell. That’ll usually get you free without a hard reset.

BLTC issues back for me now too. Not only do I get the spinning wheel waiting to load the home TP page along with the gamma dimming, when I close the window the client freezes and can’t be option-tabbed out. Needless to say when you’re in fullscreen, it sucks.

I have to ssh in and ‘killall cider’, but why should I be forced to do that? Is it another expired certificate? Why bother finding out, too much time wasted.

Since it happens every time I close that window, I’ll just sit this game out for a while.

I doubt these crashes will be truly resolved unless there’s a 64-bit version of Cider, but that’s unlikely given an ABI incompatibility / reserved register issue.

I’d like to see a switch to load generic back pieces and weapons in addition to the basic armour for a better temporary fix.

I’ve suffered Out-of-Memory / null pointer crashes for a while.

It was clear what was causing the crashes (lldb not needed,) but for fun I went through the support checklist. Repair. Login Items. New user. Reinstall. I drew the line at zapping PRAM. I wanted to play, and I’m no longer running OS 9.

If you stay long enough in any decently competitive wvw map, you’ll crash too. Turning character graphics to lowest (generic armour) isn’t a fix; the client will eventually run out of VM space and crash, albeit at a slower rate.

Since the cider port isn’t going to magically get 64-bit address space, I’ve take a proactive approach. I turned on the memory debugging overlay as per umms’ post (link below) so I can watch the virtual address space available. When it gets down around 120MB, I change maps. Changing maps frees up ~500MB of address space, so I can return to the fight without worrying too much about crashing… at least until that next big zerg starts loading in.

https://forum-en.gw2archive.eu/forum/support/mac/Client-Crash/first#post4360855

If you’re suffering TCMalloc, lockedRect.pBits or other pointer crashes, this might be an approach that works for you too. You still might crash, but you’ll see it coming.

No chest either. GoM since launch, meta done early and 1st place achievement unlocked. I’ve had problems receiving the world bonuses in the past, maybe there’s some connection.

Either way, I kind of expected it.

It’s the option key. Hold down option, drag stack. No fn key needed.

Go with a bootcamp partition. The iMac is a good machine; You’d enjoy it a lot, especially the screen resolution when you’re doing real work. You’d also get your work done faster (and don’t skimp on a fusion drive, or the upgraded graphics.) Of course running GW2 under Windows will give you the performance of a comparable PC, so you’re set there. It doesn’t make sense to buy a PC given that you’re on OSX for work. Get the iMac. If you’re self-employed, I’d imagine you could claim it as an expense for taxes as well.

Some suggestions on iMac hunting – I’m not sure which european market you’re in (Eire?) but I’d keep an eye on the refurbished section, and look at the 27" model. You’ll save some cash and have the same warranty as new. The 27" would also let you upgrade ram later yourself (much cheaper,) and be more useful for work. If it came down to choosing between 16GB of ram or a fusion drive, I’d take the drive. You’ll get by fine with 8GB of ram, but being stuck with a 5400RPM system disk hurts.

(I’m speaking from my own experience as a retired graphic designer, with a 2012 27" iMac – 3.4 GHz i7, 8 GB RAM, 1TB Fusion drive, GTX 680MX.)

There’s a small free utility called BootChamp which is helpful when using bootcamp, turning it into a simple one-click reboot, and you don’t have to set the startup disk back on the windows side – when you shut down, it’ll reboot back into OSX.

1. Native client / OpenGL rendering. Turbine pulled it off quite well, as have other developers of note. I hope this would address the majority of issues out there related to the wrapper and cider (performance, OOM texture crashes, gamma bug, TP) and come close to parity with the Windows client. Effort put into improving the current client without addressing the bigger picture is, frankly, a waste of resources.

2. Communication. If, for example, you know that there’s been yet another certificate revocation & reissue that will break the TP for the mac client, it’s not enough to post a paragraph in the Mac sub-forum and call it good. This is a major bug that affects the entire mac client-base. The info page that appears when the client is launched should contain a note about the bug and a link to the forum post. Surely the server that generates that page can tell if the request is coming from the Mac client, and send a tailored response.

3. Testing. We’ve recently had updates with game-breaking bugs that have affected a huge number of players, and these bugs have become apparent after only a few minutes of play. I have to believe they would have be caught with the barest minimum of unit testing on the mac client, but since they weren’t, I wonder what kind of testing, if any, is carried out, or even if it’s possible to test in a traditional developmental manner with cider (is it a black box?)

Somebody on the team needs to understand how the wrapper and cider work at a low, technical level, and flag code changes that might affect the mac client. A QA process needs to be put in place to specifically test these flagged changes.

There’s no need for a 4th and 5th ‘wish’ if the 1st one is granted.

Mahkno.7593:

this terrifying thing was in my console log:

3/18/14 9:33:43.000 PM kernel⁰: process cider⁴⁵³ thread 4147 caught burning CPU!; EXC_RESOURCE supressed due to audio playback

Don’t worry – your CPU isn’t literally burning.

It’s one of the new system-level features of Mavericks, part of the push to get developers to write more energy efficient code. When a thread passes a certain percentage of CPU usage over a certain amount of time it’s flagged and logged, with a spin dump (check under DiagnosticReports if you want to see the gory details.)

It’s also useful if you’re an inquisitive user and want to know which apps are draining your battery fastest.

It’s disappointing that such an obviously widespread (and quick to experience – 10 seconds after spawn in my case) error wasn’t caught before release, but it highlights how the Mac client was more about PR than priority. There didn’t seem to be any intention to support other platforms early in the design process when an exclusive DirectX path was chosen. It doesn’t even feel like a second-class client. Didn’t I read in another thread that there isn’t even a dedicated developer on it?

What’s the end game here? If the Mac client isn’t going to be supported with a level of commitment that at least approaches the Windows client, it should be scrapped. I’d hate to see it, but I suspect it would be for the best. It certainly doesn’t feel like they’re busy adding an OpenGL path for a native Mac client, and the current situation is far from ideal.

I’ve never received the bonuses either, and I’ve never transferred servers. I know a few of my guild mates get them – they never miss a chance to point out a bonus node harvest. I’d like to think this bug has priority given its major effect on play, but seeing how old this thread is…

Sigh.