rootwyrm.6408

That behavior sounds a lot more in line with packet loss rather than DX11 related. I'd get out pingplotter or the like, run /ip, and check packet loss during the fight. Please note that the final point (the address from /ip) will always be 100% because they are configured to not respond to ICMP, and it's likely one or more other hops will be configured similarly.

Hooboy, they're finally paying attention, so now I'm gonna have to do the bug dump. For the love of all that is holy, fix character creation and cutscenes! Seriously. I can live with heads and legs being chopped off on character selection. I cannot live with literally being unable to even read the text during cut scenes or character creation. Which is exactly what you get at 5760x1440. White text on a near white background with no outline, no shadow, and no text background box at all is decidedly not readable. "MY BANNERS ARE EPIC oh wait now they aren't." I don't know any other way to describe it. Sometimes dropping my banners I get HUGE GLOWY SINE-WAVE HIGHLIGHTS! BLOOM! SHIMMER! SHINE! RAINBOWS! IT'S LIKE A GRATEFUL DEAD CONCERT! IT'S AN UNDULATING CIRCLE THAT SAYS 'GLORY AND VIOLENCE FOUND HERE!' YAAAAH, I WILL DESTROY YOU! And now it's a flat, thin, blue circle that's only there if you hold your head at just the correct angle and completely invisible under the 65534 red circles and explosions going off. I guess there's banners there. It's kinda hard to see them with the textures flickering constantly even when they aren't stacked. Which will it be? Roll 1d6 and your guess is as good as mine. I can't find any specific trigger or condition - it happens both ways literally anywhere and everywhere. (For the record, I would much prefer the Grateful Dead concert if we have to choose one.) Inefficient/Insufficient VRAM Utilization Leads To Poor Performance! We both know full well that DX11 doesn't tell you how much VRAM the card has, while DX9 did (up to 4GB.) Because you aren't using the full - or even half - of the 16GB on a Radeon 6000-series? Massive delays and severe texture pop-in on zoning and even moving around zones, especially New Kaineng and Echovald. Especially for those of us with really big monitors so we can pretty much see behind us. (No, GeForce cards aren't much better.) The smallest current gen card is 8GB, and even going back three generations we were already commonly at 8GB+. Just leveraging what's there would improve performance significantly. Promise. I call this one "The Audience is Now Deaf" Audio volume sliders were broken in the last DX11 update, though this one may be Sound Blaster AE-9 specific, in which case... uh... can you expense a $350 sound card? Anyway. If you set Master Volume down to zero but have Environment Volume below 100, take Master to ~10%, oh hey it just took the Environment Volume value instead and ow my ears. I'll just, uh, slide the master over here to make it... just a lot louder across the board. But restarting fixes it. Insert shrug emoji here? Wishlist Item: FXAA/SMAA Applied to Shadows Seriously. You have no idea how utterly aggravating it is when I've got all the smooth I can wring out of the engine, and all the shadows on Ultra look like they were rendered on an SNES. Especially charr tail shadows. Pretty please. (A simple "we'll add it to the roadmap" or "are you insane? Shadow performance is already a nightmare" would be cool. I am 100% onboard with it not getting fixed at any point between now and next expansion. I'd just like to know if it's something I can look forward to.) Wishlist Item: can we please not occlude things that are not ambient like the UI? I know what a pain that'll be. But seriously. I like to run ReShade for improved MXAO differentiation. It makes things truly breathtaking. And renders the UI completely unusable because it's occluding even the hero panel. Better still, can we just.. completely separate the UI render so that we can apply things like DPX, FXAA, and FidelityFX CAS without completely blowing out the UI and making text unreadable?

So, update as of 126,047: CHANGED: EoD Chapter 8 breaking and leaving you in black and white is still present, but no longer consistently reproducible. Argh. FIXED: Colors no longer shift inappropriately in Chapter 8 area, unable to reproduce FIXED: Night light mode no longer breaks all the textures. Definitely fixed. UNCHANGED: New Kaineng still has several flickering areas in Old Kaineng, Tengu Square, and Bori Ward UNKNOWN: Haven't seen enough New Kaineng metas to verify if texture corruption is still present NEW: New texture corruption (rainbow rug effect) issues particularly in high transparency areas in Echovald and final map NEW: Ley Energy textures with transparency in final meta always turns into a black waffle of sadness. Doing some experimentation, certain dyes DEFINITELY trigger corruption. Abyssal Sea will turn your whole world blue, especially if it's on your mount. So no more excuses. That is a solid reproduction scenario - and I would bet the other Abyss dyes behave similarly.

Sigh. Ultrawide and Megawide is still completely broken, but I knew that was faint hope only because it's a nightmare to convert DX11. (The engine uses DX9 best-practices for viewports and layers, which is exactly the opposite of how it should be done in DX11.) But holy hell, the bugs somehow got orders of magnitude worse. Note, this is written very specifically to avoid spoilers as much as possible. So the vagueness is intentional. Texture Issue: textures completely break when the color shift occurs in EoD Chapter 8, resulting in trying to fight in black and white. This is 100% reproducible. Lighting/Texture Issue: colors shifted to the point of being unplayable in the area unlocked by EoD Chapter 8, appears to be associated to lighting adaptation shifts. 100% reproducible. Texture/Color Issue: if the system enters "Night light" mode, numerous textures become static blue and will not render correctly under any circumstances. This will persist with client restart. Particularly observed with minipets and Tengu clothing. 100% reproducible. Lighting/Occlusion Issue: certain surfaces in New Kaineng have severe flickering that could be seizure inducing. Observed three during the quest to open the apartment, generally in corners. Texture Issue: during New Kaineng meta events, the flashing status panels will often trigger texture corruption, including on themselves.

Welp, I thought the vertex issue had been fixed because I couldn't reproduce reliably or even remotely consistently. And all of the texture issues were no longer reproducing aside from dye channel problems. NOPE. And I found two near to 100% reproduction test case. Finally. Steve and Drakkar. But when it only happens when fighting Steve, and it happens 9 out of 10 Steve fights? It's Steve. Which also obviously makes it a difficult variable to control. But thanks to the power of obsessive ToT bag farming, I was able to isolate it. Fighting Steve and Hollowed Gourdbinder at the same time for over 5 minutes, absolutely no problems. Gourdy dies, the corpse despawns, near INSTANT crash. Viscount and Steve, switch focus to Steve, and crash. Group composition, group size, character I'm playing, none of that changes things. But sometimes, sometimes Steve doesn't crash. Drakkar though? Drakkar is a 100% always consistent you will die crash rate. Within 30-45 seconds of a wall emergency or 10 seconds of the floor stomp, I crash. Every time. I can reproduce more or less at will. So what we thought was fixed, is not, and needs bumped to the absolute top of the stack. It's breaking two world bosses already, and those are just the two I've tested. How many more are broken the same way that nobody's noticed?

So interesting fact: The character select screen and cutscene issues were not a consistent problem on DX9. In fact, it was very difficult to reproduce after I fixed it once. But I can also tell you exactly why as well. These were restricted to the UI viewport. Which on a triple head setup, is the maximum resolution of a single monitor, and automatically centered (or on Eyefinity will go to the priority display.) Which is why the UI is a good two feet wider - physically - for me. The UI layer is now equal to the full viewport, taking it from 1920x1200 to 5760x1200. It definitely was a locked layer previously. And cinematics appear to run in the UI layer and do not have a bounding box. That makes cinematics in particular virtually impossible to fix easily, if at all. One way ANet could fix cinematics would be to re-render them at MUCH higher resolutions (which also would make them much larger on disk.) This is, frankly, not going to happen nor should it. Dredging up assets that date to launch and reworking all the textures for a tiny percentage of players that would benefit? I'd be the first one saying "hell no." The second way would be to create a 'cinematic' layer with a strict bounding box. But that comes with it's own issues; transparency outside the box ruins them one way (imagine the regular game on either side of the cinematic. Yuck.) And putting a black box layer over the entire screen plus a UI cinematic layer presents it's own problems. Both cost a lot of money to implement. And again, this would only benefit a very small percentage of players - ANet might have stats, but I'd be absolutely shocked if more than 5% over the past year are playing on anything other than 16:9. So I am more than willing to give ANet a complete pass on cinematics. It sucks that we miss out on them, but it's just a cost that can't easily be justified. (But who knows. Maybe they'll read this and go "hey, we can do that layer thing in 10 minutes. Can you test it for us please?" And if so: yes, I will gladly help test anything.) However, super not giving them a pass on character creation, selection, and UI element dragging. These are very much game breaking defects. Not simply "quality of life." I'll concede that UI elements are the lowest priority of the three, but still, they're a must fix not a "nice to fix." That inability to drag UI elements means that you literally have to take your eyes completely off the action to check the party for any reason. That's a real show stopper for support. Character creation and selection being completely broken is just glaringly obvious.

I've tested with both (yes, I have multiple GPUs.) Certain things can't be 1-1'd with AMD due to addon limitations. The resolution you're at is simply not enough screen to even get near high VRAM utilization. Especially with textures as they currently sit. And I've already confirmed three times that the frame limiter is broken. Period. Patch notes are wrong, it's just broken. Even my machine can't do 5760x1200 full screen, has to be full screen windowed.

5760x1200 is hitting OOM conditions on >8GB cards. The engine does not appear to be handling higher VRAM cards correctly. More in a moment here. Camera shake is WAY broken. It is detecting the full view (5760) and there seems to be a missing bound check. This translates directly to a drastic 'wobble' that quickly ramps to literally spinning. Reproduces very reliably in The Crystal Dragon. So, OOM. "IBgfxTextureTryCopyRectToDeviceFull" - and I don't see any other reports of it here. Every one of those failures has been with VRAM below 5GB allocated. But also indicating a VRAM allocation failed. And not always popping the crash reporter . The game memory utilization is also WAY lower than I would expect. Even getting it to use a 4GB working set on a 64GB or 128GB system seems to be impossible now - it just fights like hell. So I went for complete broke - Teq at reset. Absolutely fought exceeding 4500MB on VRAM. Turned on supersampling, and that only pushed it to an absolute peak of about 4800MB even as visible players kept going up, and the engine is clawing back every single kilobyte it can. So I went further. And it hits an absolute wall at 4900MB VRAM allocated. Just goes straight to freeing up VRAM at all costs. And repeated at LLA. WvW zergs? Yep. Same wall. Managed to push Gw2-64.exe to 4,230,000ish working but it's just an absolute brick wall at 4900MB VRAM with a very significant performance hit and failed renders (zero filled textures) as it actively fights pushing past that.

Help get attention on this thread, please. There are numerous widescreen issues right now that we really need ANet to fix before they even think about releasing.

There isn't a thread for it yet, so hopefully this will help get some actual attention to these issues. Not sure if your display qualifies? Allow me to be your guide: * Ultrawide is aspect ratios of 21:9 only (so no 21:10) * Megawide is an aspect ratio of 21:10 or with a total resolution width greater than 3440 pixels or but lower than Megaultrawide * Megaultrawide is multi-panel or multi-display with a MINIMUM width of 5760 (3x 1920) and a vertical resolution that is 66% of the horizontal or less An example of a Megaultrawide would be 3x 1920x1200 (16:10) panels in landscape orientation run in AMD Eyefinity or NV Surround, or 3x 3840x2160 (16:9.) It'd be nice if for once the developers didn't just completely ignore us and leave us with a broken game or having to give up the whole point of big screens. Especially when things were working fine previously. HINT, HINT, ANET. Seriously. D3D11_VIEWPORT is probably a good chunk of what you are looking for to fix the showstoppers. So far already hit several outright game-breaking issues and smaller QoL issues myself at 5760x1200. - Character creation? Yeah. No. This is completely broken. You literally can only see one third of the selection at most. - Character selection is also horribly broken. I can rub my charr's belly and get slapped. My asura is a head and that's about it. My norn is a torso with arms. - FPS limiter is fully broken for DX11 windowed fullscreen. Alt-tabbed away from the game, FPS uncaps (I can see 140+ in many areas.) Tabbed into the game, FPS instantly caps to 60. That's clearly backwards of the intended behavior. - Ambient Occlusion + SMAA High + Supersample introduces extremely prominent and visible strobing bands on characters regardless of ambient occlusion and shadow settings. Literally epileptic warning bad strobing. Doesn't reproduce at 1920x1200. - Depth Blur is completely non-functional (I'm guessing NYI for DX11 beta there though.) - Chat, party, and map windows need to be made arbitrarily draggable. How has this never been addressed as a QoL issue? - World map render needs some love. I love being able to see all of CT at once. Except now it looks absolutely terrible. (Art issue, I'm sorry! But it really is atrocious looking.) - The top of the target window is rendered about 20 pixels lower than before. Doesn't reproduce with 1920x1200 DX11 either. - Internal FPS reporting is not working correctly. Switching from Unlimited to 60 will instantly drop reported to 60. Switching back to unlimited has a 2-5 second delay before it updates. Did not reproduce at 1920x1200.

As a very (very very) senior platform engineer and architect, this outage raises a number of serious questions from an architectural and process standpoint. Why weren't outstanding queries and disk throughput/latency/queue already being monitored? Those are all very much core metrics, and often the only warning sign you get as AWS has another disk outage due to depending on us-east or the host dying with them being completely unaware of it. (Oh yes, this has happened. Many times.) Why is the database still architected as single A/P? (We'll avoid the digression on live service issues for the moment, might come back to it.) This leaves you vulnerable either to single region outages or subjects you to a self-inflicted Denial of Cash attack. That also means in the event of future corruption or failures, you still have the same exact problem. Which brings me to three: why wasn't there a cold or warm standby? If there were known good backups, what prevented operations from spinning up a new instance while troubleshooting was going on? (Yes, I know the caveats. But it should have been able to bring things back in a lot less time or at least provide an A/B set.) While I know it's completely unthinkable and we totally don't want to admit it could ever happen, why wasn't there a predefined "EPO" (Emergency Power Off) process outside of change control? Believe me, I can tell you exactly how bad it can get if bad data is being written to a live service and you can't perform an immediate shutdown. Especially a shared tenant live service. (Buy me a beer and I'll tell you the whole story, Robert. It's a good one.) Stepping away from the live service portion, why is ANet self-managing the guest OS? (I can make some educated guesses. Neither of us want me to. Trust me. They come with a lecture series.) Unless you've got a full time junior me doing nothing but maintaining that, that's a recipe for exactly this kind of root cause. What changes has ANet made to the dev/stage/live test process to better align it to the real world loads? Yes I know you can't fire up 5,000+ desktops running bot clients. But have you implemented things like Hell Queries and Keyboard Chewers? Okay, I said I might come back to it. So let's. Why is the game still using a single database architecture? Yes, certain data must be present always (account, inventory, guilds, etc) and synchronized across all instances. But as GW2 is not a seamless world, I would have thought backend would have switched to zone shard and flush. (Also would have assumed RDS, but good on you for avoiding that Hotel California. Seriously.) Especially as backend changes like that require fewer or no client changes. That's all that's coming to mind for the moment though. On the outside looking in, there's only so many questions you can ask without making bad assumptions like what database type (I mean, I'm guessing MySQL but hoping it's PostgreSQL, though MS SQL is also fine) or operating system ("is.. is that Mandriva?!") or the like.